Dessart, Laurence (2015) Consumer engagement in online brand

communities. PhD thesis.

https://theses.gla.ac.uk/6638/

Copyright and moral rights for this work are retained by the author
A copy can be downloaded for personal non-commercial research or study,
without prior permission or charge
This work cannot be reproduced or quoted extensively from without first
obtaining permission in writing from the author
The content must not be changed in any way or sold commercially in any
format or medium without the formal permission of the author
When referring to this work, full bibliographic details including the author,
title, awarding institution and date of the thesis must be given

Enlighten: Theses
https://theses.gla.ac.uk/
[email protected]
 Consumer Engagement in

Online Brand Communities

by

Laurence Dessart

Submitted in fulfilment of the requirements for the Degree of Doctor in Philosophy

School of Social Sciences

Adam Smith Business School

University of Glasgow

May 2015
 II

Abstract
This thesis advances the concept of consumer engagement as a valid approach to the
conceptualisation and measurement of Online Brand Community (OBC) participation.
Against the background of rapid technological advances affecting the way consumers
interact online, this thesis posits that past representations of OBC participation fail to
adequately capture OBC participation. It further argues that consumer engagement offers a
new analytical lens, which is more responsive to the interactive, social and
multidimensional nature of OBCs.

The thesis conceptualises consumer engagement in OBC as an affective, cognitive and

behavioural phenomenon whereby a consumer is engaged both with the other members of
the OBC and with the focal brand. It then identifies antecedents and outcomes of consumer
engagement in OBC. The measurement and conceptual model are tested using data from
721 OBC participants. In particular, two original scales of consumer engagement are
developed, reflecting 7 sub-dimensions: enthusiasm, enjoyment, attention, absorption,
learning, sharing and endorsing. The scales are built using interview data from 25
consumers and social media experts, then calibrated and validated using quantitative data
collected from 448 English-speaking members of official Facebook pages spanning
different brand categories.

The conceptual model is tested using structural equation modelling techniques, and the
results largely support the research hypotheses. The results show that online interaction
propensity, attitude toward OBC participation and product involvement positively relate to
OBC engagement, and that online brand engagement is positively related to product
involvement and OBC engagement. Online brand engagement shows positive correlations
with brand trust, commitment and loyalty. Group invariance is largely achieved using data
from another linguistic context – a sample of 273 French-speaking Facebook page
members, which contribute to validating the English sample results.

Overall, the thesis conceptually and empirically contributes to the burgeoning literature on
consumer engagement in OBC and enhances our understanding of OBC participation. The
study provides an improved, more online-relevant conceptualisation and measurement of
consumer engagement and identifies its key individual drivers and relational outcomes.
These findings also provide strategic implications for the community of OBC practitioners.
 III

Author’s Declaration
I declare that, except where explicit reference is made to the contribution of others, this
dissertation is the result of my own work and has not been submitted for any other degree
at the University of Glasgow or any other institution.

Laurence Dessart
 IV

Acknowledgement
If I can call myself the brand manager of this thesis, I must also give credit to the
community of people that has surrounded me along the way and contributed to this end
product.

First of all, I want to thank my supervisors, Cleopatra Veloutsou and Anna Morgan-
Thomas who have been the best pair of mentors I could have wished for. Thank you
Cleopatra for your constant support and generosity. Thank you Anna for your immense
drive and optimism. Thank you Cleopatra for your attention to detail, and Anna for your
holistic view. Thank you both for teaching me so much along these four years and for
always believing in me.

Thank you to all the colleagues I have met and had the chance to work with. Thank you
David for being a great office mate and partner in crime. Thank you to Thomas Anker,
John Finch, Deirdre Shaw, Kalliopi Chatzipanagiotou, Luiz Moutinho and everyone from
the marketing cluster for your friendly fires. Your friendship and insight have helped me
grow. Thank you to all my students for their energy and hard work.

Thank you Joachin Aldas-Manzano for your precious help analysing my data. Thank you
to all my respondents and interviewees for providing said data.

Last but not least, I want to thank my parents for always believing in me, for carrying me
and letting me fly. Thank you David for embarking with me on this journey and being
there unconditionally. Thank you to all my family and friends. Your constant marks of
support have been my anchor and my fuel.
 V

Related publications
The list below includes all material issued from this study’s conceptual and empirical
developments, which has been published and presented at conferences.

Journal articles

Dessart, L., Veloutsou, C. and Morgan-Thomas, A. 2015. Consumer engagement in online

brand communities: A social media perspective. Journal of Product and Brand
Management, 24(1), pp. 28-42.

Book chapters

Dessart, L., Veloutsou, C. and Morgan-Thomas, A. 2014. Brand communities and anti-
brand communities: Similarities, differences and implications for practitioners, In
Grigoriou, N. and Veloutsou, C. (eds) Theoretical and Empirical Reflections in Marketing,
Athens Institute for Education and Research (ATINER), Athens, pp. 63-78. ISBN:
9786185065584.

Conferences papers

Dessart, L. 2015. Consumer engagement in online brand communities. In: Scottish

Doctoral Colloquium, 9-10 April, King’s College Aberdeen, UK.

Dessart, L., Morgan-Thomas, A. and Veloutsou, C. 2014. Materiality of online brand

community. In: 17th AMS World Marketing Congress, 5-8 Aug 2014, Lima, Peru.

Dessart, L., Veloutsou, C. and Morgan-Thomas A. 2014. Customer engagement in online

brand-related communities: The perspective of social media users. In: 9th Global Brand
Conference, AM SIG, 9-11 Apr 2014, Hertfordshire, UK.

Dessart, L. 2014. Developing a scale of online customer engagement. In: Scottish Doctoral
Colloquium, 28-29 April, Stirling, UK.

Dessart, L. 2013. Customer engagement in online brand-related communities: The

perspective of social media users. In: Reflecting on the Past, Celebrating the Present and
Shaping the Future in Marketing Research, 19-20 Sept, Edinburgh, UK.
 VI

Contents
Chapter 1: Introduction ...................................................................................................... 1
1.1. Focus of the study ....................................................................................................... 1
1.2. Research aim and objectives ....................................................................................... 7
1.3. Expected contributions ................................................................................................ 8
1.4. Context of the study .................................................................................................. 10
1.5. Structure and content................................................................................................. 11

Chapter 2: Literature review ............................................................................................ 13

2.1. Introduction .................................................................................................................. 13
2.2. Online Brand Community ............................................................................................ 14
2.2.1. Definition of OBC ............................................................................................ 14
2.2.2. Similarities between OBC and brand community .......................................... 15
2.2.3. Online specificities of OBCs .......................................................................... 25
2.2.4. Classification of OBCs ................................................................................... 28
2.3. Consumer participation in Online Brand Community .............................................. 30
2.3.1. Defining OBC participation ............................................................................ 30
2.3.2. Key issues in the OBC participation literature ............................................... 42
2.4. Consumer engagement as an approach to OBC participation................................... 44
2.4.1. Defining consumer engagement ..................................................................... 44
2.4.2. The subject of engagement ............................................................................. 50
2.4.3. The partners of engagement ............................................................................ 50
2.4.4. The dimensions of engagement ...................................................................... 54
2.4.6. Engagement in the online brand community literature ................................... 55
2.5. Antecedents and outcomes of OBC participation/ consumer engagement ............... 59
2.5.1. Antecedents of participation in the OBC literature ........................................ 60
2.5.2. Outcomes of participation in the OBC literature ............................................ 66
2.5.3. Antecedents and outcomes of consumer engagement .................................... 69
2.6. Gaps and research questions ..................................................................................... 75
2.7. Summary ................................................................................................................... 80

Chapter 3: Conceptual approach ..................................................................................... 82

3.1. Introduction ............................................................................................................... 82
3.2. Consumer engagement framework ........................................................................... 83
3.2.1. Defining consumer engagement in OBC ........................................................... 84
 VII

3.2.2. The subject of engagement in OBC ................................................................ 85

3.2.3. The partners of engagement in Online Brand Community ............................. 87
3.2.4. The dimensions and sub-dimensions of consumer engagement ..................... 88
3.3. Proposed model of consumer engagement in OBC .................................................. 97
3.3.1. Drivers of engagement .................................................................................... 98
3.3.2. Relationship between consumer engagements with different partners ........... 99
3.3.3. Outcomes of engagement................................................................................ 99
3.4. Hypothesis development ......................................................................................... 100
3.4.1. Drivers of consumer engagement ................................................................. 100
3.4.1. Relationship between online brand engagement and OBC engagement ...... 106
3.4.2. Outcomes of engagement.............................................................................. 107
3.5. Summary of the hypotheses .................................................................................... 111
3.6. Validation of the study ............................................................................................ 112
3.7. Summary ................................................................................................................. 118

Chapter 4: Methodology .................................................................................................. 119

4.1. Introduction ............................................................................................................. 119
4.2. The research paradigm ............................................................................................ 121
4.3. Study settings .......................................................................................................... 123
4.3.1. Social media.................................................................................................. 123
4.3.2. Facebook pages ............................................................................................. 124
4.4. Data collection ........................................................................................................ 126
4.4.1. Instrument design .......................................................................................... 126
4.4.2. Administration .............................................................................................. 131
4.4.3. Pre-test and pilot ........................................................................................... 133
4.5. Sample design ......................................................................................................... 135
4.5.1. Sampling approach ....................................................................................... 135
4.5.2. Stage 1: Sampling of the Facebook pages .................................................... 136
4.5.3. Stage 2: Sampling of the consumers ............................................................. 140
4.6. Sample treatment and characteristics ...................................................................... 141
4.7. Data analysis ........................................................................................................... 146
4.8. Summary ................................................................................................................. 148

Chapter 5: Measurement................................................................................................. 149

5.1. Introduction ............................................................................................................. 149
5.2. Measurement of consumer engagement .................................................................. 149
 VIII

5.2.1. Scale development process ........................................................................... 150

5.2.2. Domain specification and item generation ...................................................... 151
5.2.3. Data screening............................................................................................... 170
5.2.4. Item purification ........................................................................................... 172
5.2.5. Confirmatory Factor Analysis ...................................................................... 173
5.3. Scales for existing constructs .................................................................................. 188
5.3.1. Online interaction propensity........................................................................ 188
5.3.2. Attitude toward OBC participation ............................................................... 188
5.3.3. Product involvement ..................................................................................... 188
5.3.4. Brand trust..................................................................................................... 190
5.3.5. Brand commitment ....................................................................................... 190
5.3.6. Brand loyalty................................................................................................. 192
5.4. Summary ................................................................................................................. 193

Chapter 6: Hypothesis testing ......................................................................................... 194

6.1. Introduction ............................................................................................................. 194
6.2. Approach to hypothesis testing ............................................................................... 194
6.2.1. The measurement model ............................................................................... 196
6.2.2. The structural model ..................................................................................... 201
6.2.3. Testing of alternative model ......................................................................... 204
6.3. Summary ................................................................................................................. 207

Chapter 7: Discussion ...................................................................................................... 207

7.1. Introduction ............................................................................................................. 207
7.2. Consumer engagement concept and measurement ................................................. 208
7.2.1. Dimensionalisation of engagement ............................................................... 208
7.2.2. Locus of engagement .................................................................................... 214
7.3. Discussion of the hypotheses .................................................................................. 215
7.3.1. Antecedents of consumer engagement ......................................................... 217
7.3.2. Antecedents of online brand engagement ..................................................... 221
7.3.3. Outcomes of online brand engagement ........................................................ 228
7.4. Other findings: validation procedure ...................................................................... 232
7.5. Summary ................................................................................................................. 236

Chapter 8: Conclusion ..................................................................................................... 238

8.1. Introduction ............................................................................................................. 238
8.1. Theoretical contributions ........................................................................................ 239
 IX

8.3. Methodological contributions ................................................................................. 242

8.4. Managerial implications .......................................................................................... 243
8.5. Limitations and directions for future research ........................................................ 246
8.6. Summary ................................................................................................................. 250

Appendixes ........................................................................................................................ 251

Appendix 1: Cover letter to Facebook page administrators .............................................. 252
Appendix 2: The online consumer survey (English version) ............................................ 253
Appendix 3: The online consumer survey (French version) ............................................. 260
Appendix 4: Sample of Facebook pages ........................................................................... 267
Appendix 5: Examples of Facebook pages ....................................................................... 269
Appendix 6: Data screening .............................................................................................. 270
Appendix 7: Sample characteristics .................................................................................. 273
Appendix 8: Profile of consumer interviewees ................................................................. 276
Appendix 9: Profile of industry expert interviewees ........................................................ 277
Appendix 10: The interview guide .................................................................................... 278
Appendix 11: Qualitative data summary .......................................................................... 280
Appendix 12: Email to academic experts.......................................................................... 282

References ......................................................................................................................... 283

 X

List of Tables
Table 1: Brand community classification criteria 21
Table 2: Overview of selected obc studies 31
Table 3: Definitions and dimensionality of engagement in the social science disciplines 46
Table 4: Key consumer engagement studies in marketing 52
Table 5: Link between engagement and other variables 70
Table 6: Affective dimension and sub-dimensions 90
Table 7: Cognitive dimension and sub-dimensions 92
Table 8: Behavioural dimension and sub-dimensions 96
Table 9: The research hypotheses 112
Table 10: Research contexts in key obc and consumer engagement studies 116
Table 11: Facebook pages characteristics 137
Table 12: Consumer engagement items evolution 168
Table 13: Cronbach's alpha 172
Table 14: First order CFA results – calibration sample 176
Table 15: First order CFA results – validation sample 178
Table 16: Validity- validation sample 181
Table 17: Chi-square test sub-dimensions – validation sample 182
Table 18: CFA second order – Validation sample 183
Table 19: CFA second order validity – validation sample 184
Table 20: Chi Square test – Validation sample 185
Table 21: Item means by product category 187
Table 22: SEM measurement model – Standard loadings and t-values 198
Table 23: Measurement model – reliability and validity 200
Table 24: Structural model – Results 201
Table 25: Hypotheses results 202
Table 26: Alternative model SEM results 206
Table 27: Comparison of this study's dimensions with two other studies 209
Table 28: Summary of the research hypotheses 217
 XI

List of Figures
Figure 1: OBC similarities with brand communities and online specificities 28
Figure 2: Schau et al. (2009) framework of brand community practices 35
Figure 3: Model of relationships in brand communities by McAlexander et al. (2002) 37
Figure 4: Brand community model in Algesheimer et al. (2005) 57
Figure 5: Framework of consumer engagement by Brodie et al. (2013) 58
Figure 6: Process of engagement in Bowden (2009) 71
Figure 7: Process of engagement in Van Doorn et al. (2010) 72
Figure 8: Conceptual model of engagement in Hollebeek (2011a) 73
Figure 9: Process of engagement in Brodie et al. (2013) 74
Figure 10: Conceptual model of engagement in Wirtz et al. (2013) 75
Figure 11: Consumer engagement in an Online Brand Community 97
Figure 12: Conceptual framework of consumer engagement in an OBC 100
Figure 13: The causal model 111
Figure 14: The research process 120
Figure 15: The scale development process 150
Figure 16: Evolution of the number of items 166
Figure 17: The two-level CFA 175
Figure 18: The causal model (2) 195
 XII

Abbreviations
AVE Average Variance Extracted
C2C Consumer-to-Consumer
CCT Consumer Culture Theory
CFA Confirmatory Factor Analysis
CFI Comparative Fit Index
CR Composite Reliability
EFA Exploratory Factor Analysis
EM Expectation Minimisation
EN English-speaking
FR French-speaking
ICT Information and Communication Technologies
MCAR Missing Completely At Random
MAR Missing At Random
NMAR Not Missing At Random
OBC Online Brand Community
OIP Online Interaction Propensity
ROI Return On Investment
RMSEA Root Mean Square Error of Approximate
R2 Squared Multiple Correlation
SEM Structural Equation Modelling
SIT Social Identification Theory
TAM Technology Acceptance Model
TLI Tucker Lewis Index
TPB Theory of Planned Behaviour
TRA Theory of Reasoned Action
UGC User Generated Content
UK United Kingdom
 1

Chapter 1: Introduction

1.1. Focus of the study

The aim of this thesis is to advance the concept of consumer engagement as a valid
approach to the conceptualisation and measurement of online brand community
participation (hereafter OBC). An OBC is defined in this study as ‘specialised, non-
geographically bound community, based upon social relationships among admirers of a
brand in cyberspace’ (Jang et al. 2008, p.57). The study therefore focuses on two key
concepts and the related streams of marketing literature: online brand community and
consumer engagement.

OBCs have recently been identified as one of the most topical and relevant area of study in
the field of consumer-brand relationships in the last ten years (Fetscherin and Heinrich,
2015). Two reasons explain why OBCs have gained such traction as a key research area in
marketing. Firstly, the interest was spurred by a strong stream of studies on brand
communities (e.g. Muñiz and O’Guinn, 2001), from which OBCs derive directly. Due to
technological developments, brand community research was faced with the increasing need
to consider the virtual environment as a game changer in the field of consumer-brand
relationships (Hennig-Thurau et al., 2010; Yadav and Pavlou, 2014) and consumer
behaviour (Mathwick, 2002). This literature provides a foundation for existing studies in
OBCs, driven by the same broad theoretical and managerial impetus leading this study’s
agenda.

Secondly, for the last two decades, OBCs have increasingly been recognised by marketers
and scholars alike as powerful strategies for consumer-brand relationships. Academic
studies have focused on different product categories including travel services (Casaló et
al., 2010), technology software (Faraj and Johnson, 2011), luxury goods (Kim and Ko,
2012) or automotive products (Ewing, et al., 2013), among others. OBCs have sprouted all
over Internet platforms, and can now be found on social media, like Facebook or LinkedIn
(Ewing et al., 2013), on dedicated websites (Healy and McDonagh, 2013) or on discussion
 2

forums (Fielder and Sarstedt, 2014). Scholarship also evidences the advantages of OBCs
for both multinationals (e.g. Kim et al., 2008) and SMEs (e.g. O’Sullivan et al., 2011).
Clearly, OBCs are relevant to a large number of industries and company types, and can be
found on an array of different virtual platforms.

Ample evidence of OBC relevance can be found in practice. Far from being a fad
associated with transient technology enthusiasm, OBCs are still expected to be in the
forefront of the consumer relationship efforts of brands in 2015, as reported by Forrester
Research (Forrester Research, 2014). OBCs are estimated to become marketers’ preferred
approach to social media marketing because of the enormous benefits they bring in terms
of stickiness, consumer reach and loyalty, and relationship building (Forrester Research,
2014). Facebook alone registers over 40 million active small business pages in 2015
(Facebook, 2015a). Furthermore, community-based consumer interactions are considered
to be the ultimate level of conceptualisation of individuals in social contexts for any
attempt to understand online social phenomena (Murphy et al., 2014).

This study focuses on members’ participation in OBC, positing that existing approaches
and measures of this phenomenon are subject to limitations and that the concept of
consumer engagement can provide effective remedies to these limitations. In the context of
growing OBC presence, active member participation is the key to sustaining community
growth and survival. Put simply, an OBC where members do not participate is a ghost
town. Examples of OBC that died due to a lack of participation include butter brand
Lurpak’s ‘BakeClub’ (launched in 2011 in the UK) or Dell’s ‘Digital Nomads’ (launched
in 2008). When exhibiting active participations, however, OBCs are powerful consumer-
brand relationship tools in electronic environments (Kozinets, 2002). Understanding
member participation has been one of the core concerns of OBC scholars (e.g. Casaló et
al., 2008; Brodie et al., 2013).

However, information technologies bring about challenges in terms of OBC management

and managers are now faced with technology-driven consumer empowerment, user-
generated content and social consumer-to-consumer interactions (Christodoulides et al.,
2011). These new technology-embedded trends and market dynamics make OBC
participation particularly challenging to create and maintain. Determining the adequate
approach to fostering OBC participation has proven to be a complex task (Brodie et al.,
2013).
 3

Advantages of OBC participation for brands have been of particular interest for scholars.
From brand perspective, when the community is participative, brands can reap immense
benefits (Algesheimer et al., 2005). These benefits range from brand commitment (Lin,
2010), brand trust (Laroche et al., 2012), brand identification (Bagozzi and Dholakia,
2006) and brand loyalty (Algesheimer et al., 2005). Scholars also show that high
participation in OBCs is a source of innovation ideas for companies (Füller et al., 2008),
and that it increases the probability of products being purchased from the focal brand as
well as the rejection of products from competing brands (Thompson and Sinha, 2008). The
benefits of maintaining active OBC participation are thus non-negligible for brands.

Additionally, participating in OBCs brings valuable benefits to consumers. These range

from social benefits like community inclusion and social identification (Dholakia et al.,
2004), getting informational value through branded content (Wiertz and de Ruyter, 2007),
building identity (Muñiz and O’Guinn, 2001) and getting financial rewards (Garnefeld et
al., 2012). In other words, when managed adequately, OBCs have the ability to foster high
levels of participation and enthusiasm, which results in members’ benefits and in turn,
increased participation and brand benefits.

Despite significant advancements, existing research on OBC participation suffers from

several shortcomings. Firstly, by relying heavily on approaches used to study offline brand
communities, the studies do not fully acknowledge the interactive nature of online contexts
(Quinton, 2013). For example, prior studies in OBC have tended to replicate models from
the offline brand community literature in the online contexts (e.g. Jang et al., 2008). This
transposition of theories can work to a certain extent; however OBCs inherently bear
characteristics due to their virtual nature that are not accommodated in offline studies
(Quinton, 2013). Most notably, there is a lack of focus on the interactive nature of OBC
participation, which inevitably involves two parties, and is a key foundation of virtual
community participation (Colliander et al., 2015).

Secondly, the current approaches to participation show fragmentation whereby each study
focuses on selective aspect of participation from a unidimensional perspective, failing to
capture the complexity of the phenomenon (Brodie et al., 2013). For example, consumer
participation in OBCs has been theorised, explored and modelled using a number of
fragmented, unidimensional approaches. In an extensive literature review, this study shows
that OBC participation has been represented in an number of ways, including but not
 4

limited to a set of participative behaviours (e.g. Jang et al., 2008), attitudinal commitment
and affect (e.g. Kim et al., 2008) or co-creative practices (e.g. Schau et al., 2009).

Thirdly, the social and interactive aspects of consumer relationships tend to be seen as a
backdrop element contributing to participation, rather that one of their inherent and un-
dissociable characteristics (Ashley and Tuten, 2015). Research on OBC participation is
characterised by a lack of inclusion of the social aspects of OBC participation as inherent
to it, which leads to an individual, one-way, rather than conversational approach to
participation. This evidences that one-way relationship marketing concepts are simply
transposed to online environments, without taking into account the interactive specificities
of the online medium, as suggested by Colliander, et al. (2015).

Although past approaches to capturing participation have clear benefits and all exhibit
strong conceptual foundations, they collectively present a scattered, non-comprehensive
and unclear view of OBC participation. Each perspective is anchored in a different
worldview, using specific sets of theories, methods and designs, which diverge
fundamentally. Some studies have attempted to mix and match different theories (see
Casaló et al., 2010) and combine constructs from different theories in the same model
(Bagozzi and Dholakia, 2006), but there are no real attempts to provide a holistic view of
OBC participation. The very multiplicity of approaches to date is a clear indication that
OBC participation is an extremely rich phenomenon and that it should be considered as a
multidimensional rather than one-faceted occurrence (Brodie et al., 2013).

To summarise, the OBC literature focusing on OBC participation seems to suffer from
three key shortcomings, being (1) an inadequate representation of consumer behaviour in
light of the interactive nature of online environments, (2) a multiplicity of separate,
scattered and largely unidimensional treatments of OBC participation and (3) an over-
reliance on one-way models of relationship marketing, overlooking the essential social
nature of OBC settings.

In an attempt to address these issues, this study focuses on advancing scholarly

understanding of consumer engagement, proposing it as the most adequate approach to
OBC participation and investigating its antecedents and outcome. The study argues that
consumer engagement offers a more potent and relevant framework for the understanding
of OBC participation (Brodie et al., 2013). By incorporating current developments in
 5

online and relationship marketing (Vivek et al., 2012), the approach adopted here
facilitates bringing OBC participation studies back to the forefront of research.

Building on key studies on consumer engagement (Brodie et al., 2011; Hollebeek, 2011;
Brodie et al., 2013), the present study defines consumer engagement in OBC as the state
that reflects consumers’ individual dispositions toward engagement partners. In an OBC
context, these partners are the OBC, representing the other consumers in the OBC, and the
focal brand. It is expressed through varying levels of affective, cognitive, and behavioural
manifestations that go beyond exchange situations. Consumer engagement is defined as a
psychological and motivational concept, which manifests itself for a specific engagement
subject, here the consumer. Furthermore, consumer engagement is expressed through three
types of occurrences, which are defined as being of affective, cognitive and behavioural
nature. These constitute what is referred to as the ‘dimensions’ of consumer engagement
(Brodie et al., 2011). Additionally, consumer engagement is understood as a context-
specific concept, which is therefore malleable and offers a broad range of contextual
adaptations, including OBCs (Brodie et al., 2013). One of the ways in which consumer
engagement can be adapted to specific consumption-related contexts is by adapting the
engagement ‘partner’. Relevant partners of consumer engagement in the context of OBC
are the brand (Brodie et al., 2013) and the community itself (Algesheimer et al., 2005).

Several arguments support the superiority of consumer engagement over other approaches
to capturing OBC participation. Firstly, consumer engagement takes into account the
specificities of the online environment (Christodoulides et al., 2011) that drive new forms
of consumer behaviour. Indeed, consumer engagement seems extremely suited to online
contexts for a number of reasons. For example, it is social and interactive by nature and
can only happen in the presence of the engagement subject (the consumer) and its
engagement partner (the brand or community) (Breidbach et al., 2014). This renders
consumer engagement particularly dynamic and interactive (Brodie et al., 2011), which is a
founding principle of Internet-mediated communication. Consumer engagement has the
potential to foster an improved understanding of focal consumer-brand and consumer-to-
consumer interactions within ICT-mediated environments (Sawhney et al., 2005;
Breidbach et al., 2014). Mollen and Wilson (2010, p.1) go as far as saying that engagement
should be understood as the ‘definite umbrella term’ for online interactions. Breidbach et
al. (2014) further support this stance by explaining how consumer engagement is the most
 6

adapted approach to understand consumer-to-consumer and consumer-firm interactions in

the context of ICT-mediated engagement platforms.

Secondly, although different schools of thought exist (Hollebeek, 2011a), consumer

engagement is largely conceptualised as a multi-dimensional concept, including emotional,
cognitive and behavioural facets (Brodie et al., 2011). In this sense, consumer engagement
as a concept resolves a number of the limitations inherent to existing approaches to OBC
participation, which tended to take a selective approach to participation. Consumer
engagement offers a way to address the need for a more refined and multi-dimensional
measure of brand-related behaviour, which is needed to strengthen the relationship
between group and brand-related behaviour (Bagozzi and Dholakia, 2006).

Thirdly, consumer engagement acknowledges the combined role of individual and social
consumer behaviour in brand-related contexts. Consumer engagement is positioned in the
extended realm of brand relationships and builds on existing relationship marketing
literature (Vivek et al., 2012). Virtual environments and related research have swiftly
evolved toward interactive, participatory and consumer-empowering modes of action and
investigation, forcing the relationship marketing paradigm to reinvent itself. Far from
allowing only one-on-one consumer-brand relationships to evolve, OBCs embed different
types of consumer-centred relationships (McAlexander et al., 2002), allowing social
consumer interactions to thrive (Christodoulides et al., 2011) and making brands
themselves increasingly social.

It is important to acknowledge that in spite of its obvious advantages, the concept of

consumer engagement in OBC is in itself underdeveloped, requiring further refinement.
For example, although Brodie et al. (2013) recognise the convergence of consumer
engagement and OBC research and call for further studies at the convergence of these two
domains, the application of consumer engagement in ICT-driven contexts ‘remains ill-
defined, and empirical contributions in this area are limited to date’ (Breidbach et al.,
2014, p. 594). The concept of consumer engagement seems to be still in its infancy,
compared to other conceptual frames of analysis of OBC participation, such as social
identification (Dholakia et al., 2004). Although consumer engagement is strongly rooted
into the relationship marketing paradigm and offers an expanded view of it (Vivek et al.,
2012), robust conceptualisations are only beginning to emerge and current empirical
operationalisations are divergent, fragmented and still lacking depth (Hollebeek et al.,
2014). As a result, empirical research of consumer engagement in online contexts is still
 7

scarce, thus restricting scholars’ understanding of this phenomenon (Hollebeek et al.,

2014). One of the key limitations of the existing research is the extremely narrow set of
options to measure consumer engagement at all, let alone in an online or OBC context.
Only two studies have attempted this exercise so far (see Hollebeek et al., 2014; Vivek et
al., 2014).

Moreover, if consumer engagement is to be used as a frame of reference to understand and

measure OBC participation, it is important to understand what drives it and results from it
(Brodie et al., 2013). Although it is conceptually clear that customer engagement is distinct
from other relational constructs (Hollebeek, 2011), the relationships and their directions
lack empirical testing and there is strong scholarly urge for further validation (Hollebeek et
al., 2014). Particularly in the field of OBC participation, a number of potential drivers and
outcomes have been identified in extant research, which would need to be tested against
the consumer engagement approach, and combined with conceptualisations of the drivers
and outcomes of engagement.

1.2. Research aim and objectives

Addressing the gaps in OBC participation and consumer engagement literature, the aim of
this research is to advance the concept of consumer engagement as valid approach to
conceptualising and measuring online brand community participation. Extending the
treatment of consumer relationships and engagement into the domain of OBC
participation, this study develops and refines the consumer engagement as a new concept
to conceptualise and measure OBC participation. Consumer engagement is an online-
relevant, multifaceted, holistic and interactive concept, which is well suited to OBC
research. Given the relative newness of the concept, the study offers insight into the role
that consumer engagement plays in bridging motives for participation and their outcomes
(Brodie et al., 2013).

More specifically, the study attempt to address the following research objectives:

(1) To clarify the conceptualisation and propose a measure of consumer

engagement as an approach to online brand community participation
 8

(2) To understand the role of consumer engagement as contributing to a cause

and effect network of relationships in the context of online brand communities.

The first objective is a prerequisite to the achievement of the second one. It aims to bring
clarity to the definition of consumer engagement, based on which a measure of consumer
engagement adequate to OBC contexts will be created. This step is necessary to enable the
application and measurement of consumer engagement in an OBC context.

The second research objective relates to the role of consumer engagement as a new
approach to OBC participation. Given the novelty of the approach adopted here, it is
necessary to understand under which conditions consumer engagement occurs in OBCs
and which effects it generates. To date, research in this area is fragmented and largely
conceptual, with only a minimal number of studies contributing to empirical evidence
(Brodie et al., 2013; Dessart et al., 2015). This first objective will therefore focus on the
antecedents and outcomes of consumer engagement in OBC and aim to identify the most
relevant ones to study in light of the current state of the literature.

1.3. Expected contributions

Based on the research aim and objectives developed above, this study expects to contribute
to the OBC and consumer engagement literature in four major ways.

Firstly, this study aims to advance the conceptualisation of consumer engagement and
positions it as a holistic, multidimensional and integrated approach to capturing and
conceptualising OBC participation. By reconceptualising consumer engagement in OBC,
this study will clarify the relevance and applicability of consumer engagement as a way to
study consumer-to-consumer and consumer-brand interactions in OBCs (Brodie et al.,
2013). Based on a deeper understanding of its dimensions, this study intends to propose a
novel conceptualisation of consumer engagement, one acknowledging concept
dimensionality and different focuses. The present research contributes mainly to the
literature on OBC participation (Casaló et al., 2008; Dholakia et al., 2004, Stockbürger-
Sauer, 2010) but the concept should be to be adapted seamlessly to other consumption-
related context.
 9

Secondly, this research seeks to quantitatively capture consumer engagement through the
creation of a scale. Due to the paucity of existing scales of consumer engagement
(Hollebeek et al., 2014; Vivek et al., 2014), this constitutes a major step forward to allow
further empirical research requiring a measure of consumer engagement. By providing
quantitative evidence of the existence of consumer engagement in an OBC context, the
study will strengthen consumer engagement research moving the flied beyond the
exploratory stage, answering Brodie et al.’s (2013) call. In this sense, the study will
contribute to empirically validating consumer engagement as a new lens through which to
study OBC participation thanks to the creation of a valid and reliable scale, an endeavour
that thus far lacks empirical support (Hollebeek et al., 2014).

Thirdly, this study aims to contribute to the OBC and consumer engagement literature by
advancing and testing antecedents of consumer engagement in OBCs. By doing so, it
partakes in the discussion surrounding the identification of the drivers of OBC
participation, which is rich in different theories and approaches (e.g. Dholakia et al., 2004;
Bagozzi and Dholakia, 2006; Füller et al., 2008; Matzler et al., 2011). Attention will be
given to antecedents so far under-researched in order to advance the understanding of OBC
participation from a consumer engagement perspective.

Lastly, the present research also seeks to consolidate knowledge regarding some of the
most relevant outcomes of consumer engagement in OBC through empirical testing of
these relationships (Hollebeek, 2011a; Hollebeek et al., 2014). Since consumer
engagement is proposed as a novel approach to OBC participation, it is important to
identify key positive consequences of this engagement (Brodie et al., 2013). The research
will focus on benefits most relevant to brand managers and scholars, as identified in the
OBC and consumer engagement studies.

In addition to addressing conceptual concerns, it is also the aim of this study to provide
valuable insight for OBC practitioners. Consumer engagement is highly topical for
marketing and social media managers, and this study seeks to provide them with guidance
on how to best manage their OBC and reap relevant benefits from it. The study should give
insight to how to best capture and generate consumer engagement and offer an alternative
to current practitioner approaches. In particular the study paints engagement as a more
complex concept than what leading social media consultancies make it out to be, yet that it
can also, on the other hand be captured in a simple and comprehensive fashion. By finding
out which elements contribute to higher engagement, this study will also clarify ways to
 10

efficiently increase and sustain it for OBC managers. Lastly, the importance of high levels
of engagement will be evidenced through an investigation of its key consequences. Focus
will be placed on relevant OBC efficiency indicators that will help justify long-term
investment in them. This study will then contribute to creating successful OBCs rich in
consumer engagement and helping brands achieve their long-term consumer relationship
goals.

1.4. Context of the study

Social media, and Facebook pages in particular, are chosen as the context of investigation
for this study. This choice was given careful consideration because the contexts in which
ideas occur are of general importance (McCracken, 1988), and consumer engagement is a
context-specific concept largely influenced by engagement platforms (Breidbach et al.,
2014).

Based on the premise that OBCs can develop on social media (Casaló et al., 2008; Healy
and McDonagh, 2010; Casaló et al., 2010; Cova and White, 2010; Garnefield et al., 2012;
Zaglia, 2013), several facts support their particular adequacy as OBC environments. Social
media foster the development of consumer relationships with and about a brand (Hennig-
Thurau et al., 2010). This is done through fanpages, social media advertising or sponsored
posts (Gironda and Korgaonkar, 2014). More specifically, the social nature of these
platforms, as well as their large scope give consumers more opportunities to directly
interact with brands, and with other consumers (Trusov et al., 2009; Casaló et al., 2010).
This trend is well recognised by brands, and more and more of them are flocking to social
media to engage with consumer, investing increasing amount of money into social
marketing (Forrester Research, 2014).

Moreover, consumers are more empowered by social media than other types of
communication channels: social media are collaborative and participative tools where
consumers can generate content (Kaplan and Haenlein, 2010), which is the translation of
their desire to engage (Christodoulides et al., 2011). Eager for empowerment and
interactivity, an increasing number of users still join OBCs on social media (Forrester
Research, 2014).
 11

Facebook is then chosen as the most adequate social medium for this study, focusing on its
‘page’ function as ideal manifestations of OBCs. Facebook pages are official accounts,
which are managed by and related to a specific brand (Janh and Kunz, 2012), exhibit key
OBC characteristics (Zaglia, 2013) and support relationships with other consumers and the
brand (Gummerus et al., 2012; Chang et al., 2013). Facebook is chosen over other forms of
social media, such as Twitter, because it is the most popular social network in existence
(Pew Research Centre, 2015). With over 1 billion active members and over 40 million
active pages for SMEs only (Facebook, 2015a), Facebook pages represent the most
adequate context for the investigation of consumer engagement in OBCs.

1.5. Structure and content

To achieve its intended contributions, this study addresses the above-mentioned aim and
objectives following a series of steps delineated in the six following chapters.

Chapter 2 of this thesis constitutes the literature review, which draws on the OBC and
consumer engagement streams of literature. The chapter first details current
conceptualisations and characteristics of OBCs. It then identifies and critically evaluates
different approaches to OBC participation. Next, the chapter focuses on consumer
engagement and its development in the marketing literature to date. It then provides a
critical analysis of the current studies integrating consumer engagement in OBC
scholarship. A review of the antecedents and outcomes of OBC participation and consumer
engagement is presented next. Based on this review, the chapter closes with a summary of
the research gaps and proposes three related research questions to address them.

Chapter 3 advances a conceptual model of consumer engagement in OBCs, which serves

as a basis for this study’s methodology design. A theoretical framework of consumer
engagement in OBC is presented first. The aim of this framework is to delimitate the
theoretical boundaries of the research, determine the key aspects of consumer engagement
in OBC, and provide a basis for its measurement, in support for the first objective of the
study. The second objective of the study is then conceptually addressed: a causal model of
the antecedents and outcomes of consumer engagement in OBCs is presented, and a series
of related research hypotheses developed. Based on the current state of OBC and consumer
 12

engagement research, a validation procedure is also stipulated to strengthen this study’s

findings.

Chapter 4 is dedicated to the methodology and process used to address the research
questions. Starting with its philosophical stance, it then explains the different steps taken to
collect data detailing the instrument design, data collection, and sampling procedures.
Features of the research sample are presented here, followed by the sample treatment and
an overview of its characteristics. The data analysis approach is lastly addressed in support
of the chosen analytical techniques.

Chapter 5 presents the measurement of this study’s constructs. Original measurement is

addressed first, detailing the steps followed to create one mirrored scale of consumer
engagement, in line with the first objective of the study. On this basis, items are developed
to measure online brand community and online brand engagement, and administered with
the survey procedure detailed in Chapter 4. Lastly, these items are subjected to a series of
psychometric tests resulting in two valid and reliable scales of consumer engagement in
OBC. The chapter concludes by presenting the existing scales chosen to measure the other
variables of the causal model.

Chapter 6 focuses on testing the hypotheses presented in chapter 3, using recognised

structural equation modelling techniques, including confirmatory factor analyses and
causal path estimation.

Chapter 7 presents a detailed discussion of the study’s findings, elaborating on the results
of Chapter 5 and 6. The results are presented in light of the existing OBC and consumer
engagement literature. Their congruence with, or departure from the findings of existing
studies are put forth and interpreted.

This thesis concludes in Chapter 8, with an account of its key contributions to the fields of
OBC and consumer engagement. Theoretical contributions are first presented, followed by
methodological ones, and the recommendations for managers of OBCs that this research
implies. The limitations of the present study are also discussed and suggestions to generate
further research on the compelling topic of consumer engagement in OBCs are detailed.
 13

Chapter 2: Literature review

2.1. Introduction

The concepts of OBC and consumer engagement are the core elements of this study’s
conceptual development. This chapter presents an extensive literature review of these two
concepts, starting with OBCs. It is structured in four main sections.

The first section focuses on describing the concept of OBC and presenting how it has been
conceptualised so far. Since an OBC is a specific form of brand community (Jang et al.,
2008), which is itself a particular manifestation of community (Muñiz and O’Guinn, 2001),
this section builds on the brand community literature and also defines the terms of
community and brand. The communalities between brand communities and OBCs are
presented and detailed, followed by an explanation of the online specificities that make
OBCs unique and provide an essential basis to this research’s objectives and contribution.

Secondly, as this study aims to advance research into OBC participation, core studies into
OBC participation are reviewed and their approaches synthesised according to a number of
criteria, including the theories, designs and approaches used. Different views of OBC
participation are presented, based on the treatment they received in the literature.

In the third part of this chapter, the concept of consumer engagement is presented as a
novel way to conceptualise OBC participation. This section details the concept of
consumer engagement and its congruence to the understanding of OBC participation. It
does so by first presenting the concept as treated in the social sciences and marketing
literature, followed by a critical evaluation of the current models of consumer engagement
in the OBC literature.

In the fourth and last section, the literature addressing the antecedents and outcomes of
consumer participation in OBC is reviewed, followed by the same exercise in the
consumer engagement literature. The aim is to map out the most relevant antecedents and
outcomes of consumer engagement to investigate in an OBC context in light of the two
streams of literature.
 14

The overall research gaps deriving from the extant OBC and consumer engagement
literature review are then presented, highlighting the problems inherent to previous views
of OBC participation and matching them with the strengths offered by the consumer
engagement view. This leads to the articulation of detailed research questions that drive the
empirical part of this research.

2.2. Online Brand Community

This first section of the literature review focuses on defining OBCs and all related terms,
including brand community, community and brand. It details key characteristics of these
terms and shows their relationships with one another, subsequently focusing on the
particular characteristics of the online environment.

2.2.1. Definition of OBC

The study defines an OBC as a ‘specialised, non-geographically bound community, based

upon social relationships among admirers of a brand in cyberspace’ (Jang et al., 2008, p.
57). This definition builds from a well-accepted brand community definition (Muñiz and
O’Guinn, 2001) and refines it adding the online element, showing that OBCs represent
virtual manifestations of brand communities. It further suggests that OBCs share a number
of characteristics with offline brand communities, but also exhibit specific characteristics
pertaining to their virtual nature. Considering commonalities with offline brand
communities, OBCs exhibit all three community markers identified by Muniz and
O’Guinn (2001), which are shared rituals and traditions, shared consciousness of kind and
a sense of moral responsibility. OBCs, like their offline counterparts are also centred on a
specific brand.

The elements that differentiate OBCs from brand communities concern the elimination of
space and time boundaries (Andersen, 2005), an increase of the level of social interactivity
(Quinton, 2013), a broader reach and improved accessibility (Kaplan and Haenlein, 2010;
Trusov et al., 2010), as well as more consumer power (Labrecque et al., 2013). These
represent challenges for OBC scholarship and essential considerations driving this thesis.
 15

2.2.2. Similarities between OBC and brand community

Brand community research is a growing field of interest in marketing since Muñiz and
O’Guinn’s seminal article in 2001. They introduced brand community as a ‘specialized,
non-geographically bound community, based on a structured set of social relationships
among admirers of a brand’ (Muñiz and O’Guinn, 2001, p. 412). Brand communities are
forms of communities specific to the marketplace that exhibit the characteristics of
traditional communities, with their own market logic and expressions (Muñiz and
O’Guinn, 2001).

The notion of brand community plays an important role in contemporary marketing theory
because it adds social dynamics into consumer behaviour and brand relationships, two
fields traditionally concerned with individual consumer activity (Muñiz and O’Guinn,
2001; Kates, 2003). Furthermore, brand communities reveal instances of social solidarity,
value co-creation and symbolic consumption (Arnould and Thompson, 2007). Generally,
brand community research is focused on the understanding of the bonds consumers form
around brand use and affiliation (Muñiz and O’Guinn, 2001; McAlexander et al., 2002;
Bagozzi and Dholakia, 2006). This section presents the characteristics and internal
mechanisms of brand communities.

In order to understand what a brand community is, it is first necessary to distinguish the
concept of communities from other forms of groups such as tribes or sub-cultures. Indeed,
marketing scholars have so far investigated numerous types of consumer groups and the
literature on this topic is made of different terminologies (Thomas et al., 2011). The
concepts of sub-cultures and tribes coexist with the concept of community, yet their
meaning appears to be different. This section endeavours to clarify the difference between
communities, tribes and subcultures of consumption.

The concept of community is rooted in the sociology literature. A community refers to a

grouping of individuals based on feelings of togetherness and mutual bonds, which aim to
be maintained by its members (Tönnies, 1887). In his seminal work, Tönnies (1887) puts
the community (Gemeinschaft) in opposition with the modern society (Gesellschaft). The
Gemeinschaft is a traditional, geographically local and familiaristic concept, whereas the
Gesellschaft represents a new social order characterised by competitiveness, self-interest
and formal relationships. The Gemeinschaft vs Geselschaft dichotomy led sociological
community thinking for decades on.
 16

However, influenced by multiculturalism, globalisation, urbanisation, and new

technologies, conceptualisations of communities have since then reconciled with
modernity (Delanty, 2003; Bruhn, 2005). For instance, in the current technological and
social context, a common geographical location is no longer a requirement for a
community to exist. Highly personalised, networked and remote social communications
now enact communities. Communities have evolved past a stage of geographic
embeddedness to become freed from geographic constraints (Castells, 1996).

Despite these contextual, societal and philosophical shifts, the core attributes of
communities seem constant, as evidenced by their repetitive appearance in sociology work.
Social relationships, communicative power and need to belong lie at the centre of modern
(and postmodern) community conceptualisations (Calhoun, 1983; Castells, 1997; Delanty,
2003). Communities are characterised by conscious and voluntary relationships among a
group of people with similarities that overcome their differences (Bruhn, 2005). This
voluntarism is enacted through mutual goals and values (Sarason, 1974) that generate
common commitment and responsibility toward the community. Community members
share a purpose, an identity and a belief that the community unites them (Loewy, 1993).

A subculture of consumption on the other hand is a ‘distinctive subgroup of society that

self-selects on the basis of a shared commitment to a particular product class, brand or
consumption activity’ (Schouten and McAlexander, 1995, p. 43). Harley Davidson
motorcycle riders, for instance, show that interactions with each other help members derive
understanding of the brand, and that the brand itself helps members substantiate their
position in society (Schouten and McAlexander, 1995). However, the riders’ connections
to each other go further than the brand: they are representative or a real ethos and way-of-
life that one could even refer to as religion or ideology.

Tribes are yet another way for consumers to associate. They emerge in the marketing
literature under a postmodern light as ‘micro-groups in which individuals share strong
emotional links, a common subculture and a vision of life’ (Veloutsou and Moutinho,
2009, p. 316). Tribes come together mainly on the basis of passions, shared irrational
emotions, lifestyles and consumption activities (Cova, 1997). In tribes, brands can act as a
support to the tribal link but not substitute for it. Cova and Cova (2002) highlight the
socialising role of tribes and the importance of brands in this context. Although brands do
not act as the central power in the tribe, the linking value they offer is important in
sustaining tribal relationships. Even though tribe members exhibit very strong emotional
 17

bonds that enable them to create meaning through shared experience and rituals, tribes are
inherently dynamic, unstable and small (Maffesoli, 1996).

Although subcultures and tribes have been conceptualised similarly to communities in

marketing, they differ from them in three major aspects. Firstly, tribes and subcultures
have a wider scope and range of activities than those directly linked to the brand. They
enact a way of life, an ethos (Schouten and McAlexander, 1995). Despite being an
important aspect of the connection among members, the brand does not necessarily act as
the essential focus of subcultures or tribes. A brand community, on the contrary, is based
on the pre-existence of the focal brand and it does not exist without it (Muñiz and O'Guinn,
2001). The case of the Apple Newton brand community shows that the brand community
can be as strong as to keep on existing even when the product has been withdrawn from the
market (Muñiz and Schau, 2005). Secondly, both tribes and subculture carry a certain
aspect of marginality from the mainstream culture, a concept that does not apply to brand
communities. In brand communities, even though the brand serves as a mean of
differentiation, it generally does not support the identification to a society ‘outlaw’ or
‘outsider’ status (Muñiz and O'Guinn, 2001). Thirdly, there is a strong element of
fuzziness and instability, particularly in tribes in contrast to brand communities which are
more stable, formal and committed than tribes (Cova, 1997).

This study focuses on brand communities and adopts the stance that they are distinct from
tribes and subcultures due to their higher stability and endurance, mainstream aspect and
brand-only focus (Muñiz and O’Guinn, 2001). More specifically it explores brand-centric
communities, that is to say communities that are dedicated from the onset to one specific
brand (hereafter ‘brand communities’) (Wirtz et al., 2013). The orientation of the
community toward a consumption object is referred to as the ‘community purpose’ (Porter,
2004), or ‘focus’ (Thomas et al., 2013). ‘Brand orientation’ in particular is defined as the
centrality of the brand as the focus of the community (Wirtz et al., 2013). This study
investigates online forms of brand-centric brand communities.

Following this assertion and because of the central role of the brand in this type of
communities, it is important to define what is meant by ‘a brand’ in the context of this
study. Innumerable conceptualisations of the brand exist in the marketing literature (de
Chernatony and Dall'Olmo Riley, 1998). The most basic understanding of a brand
conceptualises it as product, or set of functional attributes. In this line of thoughts, the
American Marketing Association defines the brand as a ‘name, term, design, symbol or
 18

any other feature that identifies one seller’s good or service as distinct from those of other
sellers’ (American Marketing Association, 2012). This rather narrow definition focuses
primarily on its ‘physical’ traits and features. It views the brand form the company’s
perspective as a product offering and aims to communicate the brand identity to the public.
Brand identity defines the way in which the brand is developed and perceived within the
organisation. It is an internal view that aims to be communicated to the target audiences
(de Chernatony, 1999).

Contending that this vision of the brand is limited, de Chernatony and Dall’Olmo Riley
(1998) provide a spectrum of the brand interpretation with 14 characteristics ranging from
the brand as a logo, to the brand as a risk reducer or a vision. Fournier (1998) advanced
brands as a relationship instrument for the creation of consumer—brand relationships.

Bringing these different approaches together, scholarship is increasingly recognising that a

brand needs to be managed through the full consumption experience, including the
physical features of the product, its emotional character (Egan, 2007), the relationship
between the consumers and the brand (Fournier and Lee, 2009), and the relationship
between the consumer and other consumer (Veloutsou and Moutinho, 2009). Moreover,
brand consumption moves beyond mere usage of the product or service, and expands
through active consumer participation in socially constructed environments, such as brand
communities (Muñiz and O’Guinn, 2001).

In this context, this study adopts a vision of the brand following the community paradigm
to brand management proposed by Quinton (2013). This paradigm acknowledges the
central role that the digital environment and brand communities now plays in consumers’
lives, and proposes that the way brands are understood should reflect this centrality. It also
refutes the traditional consumer-brand relationship model that views relationships as linear,
relational, exchange-based partnerships (Louro and Cunha, 2001), which is in line with the
position of this study.

The community paradigm views brands as semi-independent entities (Quinton, 2013),

which are influenced by both the environment and marketing managers (Jevons et al.,
2005). The influence from the environment includes increasing levels of consumer input
via brand communities, which are facilitated in online settings. Recognising the
fundamentally changing dynamics of brand management driven by digital technologies is
acknowledged as a stringent issue. For this reason, ways to appraise brand management
 19

and performance have moved toward more digital-oriented methods and approaches (e.g.
Christodoulides, 2009; Hennig-Thurau et al., 2010).

In the community paradigm, brands are defined as being ‘semi-independent entities

demonstrating ‘connectivity’ to consumers, heavily involved in providing ‘brand
experiences’ and developing ‘engagement’ with varied audiences’ (Quinton, 2013, p. 922).
Brands are considered as the ‘glue’ that hold together groups of consumers in digital
environments. They have a facilitating and connecting role in a set of non-linear
communications taking place between individuals, groups and/or companies (Kozinets,
2001).

Now that the general concept of brand community is clarified through an understanding of
the notions of community and brand, the following paragraphs focus on describing the
characteristics of brand communities, which are shared by OBCs. Building on sociological
work and empirical qualitative evidence, Muñiz and O’Guinn (2001) identified three core
brand community markers, which have been considered to be essential to the formation of
both brand communities and OBCs (Zaglia, 2013).

The first criterion is a shared consciousness of kind, which pertains to the feelings
members have for one another, and for the brand (Muñiz and O’Guinn, 2001). This feeling
of ‘we-ness’ has been integrated in several studies as a form of identification with the
group (see Carlson et al., 2008). It is an important community mechanism, as the self-
categorisation of a person as a member of a particular community is a form of social
identification (Algesheimer et al., 2005). This sense of belonging to an in-group transcends
geographic boundaries (McAlexander et al., 2002) and determines membership legitimacy
as well as oppositional brand loyalty (Thompson and Sinha, 2008).

The second criterion concern shared rituals and tradition, which represent accepted
and reproduced social processes surrounding the brand (Muñiz and O’Guinn, 2001). They
permit the production and transmission of the meaning of the brand within and outside the
community and are largely based on the shared consumption experiences with the brand.
Storytelling is an example of ritual and tradition in the Jeep brand community
(McAlexander et al., 2002).

The third marker denotes a sense of moral responsibility to the community and its
members, responsibility which is limited to the community boundaries and induces
collective action. Two communal duties exist: integrating and retaining members, as well
 20

as assisting members in the proper use of the brand (Muñiz and O’Guinn, 2001). These
have been reinforced as key community aspects by Schau et al. (2009), who refer to them
as practices of ‘social networking’ and ‘brand use’.

There is a consensus in the OBC and brand community literature that these three core
determinants of brand communities need to be present if a group of consumers is to be
called a brand community (Muniz and O’Guinn, 2001; Zaglia, 2013). Although there is a
lack of measurement to capture these three aspects of brand community quantitatively
(Madupu and Krishnan 2008; Madupu and Cooley, 2010), number of studies has found
evidence of the presence of these characteristics in brand communities in both offline and
online settings (e.g. Madupu and Krishnan, 2008; Zaglia, 2013). These findings support the
initial idea that OBCs are similar to traditional brand communities (Muñiz and O’Guinn,
2001).

However, beyond these unifying elements, brand communities can vary with respect to
other characteristics and considerable efforts concern the classification of communities
based on these characteristics. Extant studies have attempted to classify brand communities
according to different criteria, producing different typologies (e.g. Kozinets, 1999;
Stanoevska-Slabeva and Schmidt, 2001; Dholakia et al., 2004; Porter, 2004; Cova and
Pace, 2006; Fournier and Lee, 2009). These studies focus on various classifying elements
such as social structure (Kozinets, 1999), governance and relationship orientation (Porter,
2004), or geographic dispersion and size (Dholakia et al., 2004). None of these typologies
is however widely accepted in the brand community literature. This can be attributed to
fundamental issues of typology building, such as a lack of mutual exclusiveness or
collective exhaustiveness across criteria, or a lack of relevance (see Hunt, 1991).

Due to the difficulty to pinpoint ‘one’ accepted typology of brand communities, this study
rather presents the most prominent criteria that have been used in existing typologies and
explains how brand communities vary on their basis. The most cited classification criteria
are: (1) Governance, (2) Marketplace orientation, (3) Size, (4) Duration, (5) Relationship
structure and (6) Space. These six classification criteria of (online) brand community are
important to understand because they strongly affect community functioning and any
attempt to empirically explore it (Fournier and Lee, 2009). The table below summarises the
meaning of these classification criteria.
 21

Table 1: Brand community classification criteria

Criterion Definition
Governance The entity responsible for the creation, funding,
management and control of the community
Marketplace orientation The degree to which the marketplace plays a collaborative
role in communities.
Size The number of brand community members
Duration The amount of time elapsed since the brand community has
existed in its current form.
Relationship types The strength and structure of the brand community
relationships, and partners involved in it.
Space The geographic concentration and level of virtuality of the
brand community.

Governance refers to the creation, funding, management and control of the community. In
brand communities, these roles can be assumed by the company (or its representatives), by
a single consumer or by a group of consumers. Two main types of brand communities can
be delineated based on the principle of governance: either the community is ‘official’, i.e.
governed by the company, or initiated and run by consumers (Breitsohl et al., 2015).

Governance has important implications with respect to members’ selection, purpose and
scope of activities, expressive freedom, customer motivations and presence of community
markers, as detailed by Dholakia and Vianello (2009). Dholakia and Vianello (2009) argue
that company-lead communities fail to properly enact the three fundamental markers of
communities and that customer-lead communities are more participative, open and
beneficial for their members. McWilliam (2000) on the other hand contends that
corporations perform a number of important facilitating functions such as encouraging
dialogue, fostering relationships and active participation. In this sense, they can shape
communities according to their goals and objectives.

In light of this debate, a moderate point of view is taken by Jang et al. (2008). They
evidence that both consumer- and company-initiated communities have an impact on
community commitment and brand loyalty, however through different workings.
Consumer-lead communities should place an increased focus in creating quality content,
whilst companies should make sure to reward participation. This suggests that people join
consumer-governed and company-governed communities to satisfy different needs. Their
purpose and utility are different.
 22

Consumer communities can work with the purpose to support, or counter marketplace
dynamics, and consumers can decide to support or reject a brand, or company. The concept
of marketplace orientation captures these differences. Thomas et al. (2013, p. 1015)
define marketplace orientation as ‘the degree to which the marketplace plays a
collaborative role in communities’. A collaborative community therefore works along with
the marketplace, a neutral community has little interactions or rapport with the
marketplace, and an oppositional community opposes the marketplace directions (Thomas
et al., 2013). Although centred on a specific brand, anti-brand communities exist as well,
and have the purpose to oppose the focal brand, its ethical system or market dominance
(see Hollenbeck and Zinkhan, 2010). This is in contrast to brand communities which aim is
to show support and allegiance to the focal brand.

Brand communities exist independently of their size (Bagozzi and Dholakia, 2006). There
is however contention as per the optimal community size to generate full benefits for its
members and the brand. The size of the community has been linked with feelings of social
connectedness, members’ participation, identification and community value (Bagozzi and
Dholakia, 2006). Some studies propose higher levels of social influence in lager
communities (Algesheimer et al. 2005) and higher appeal in joining and sustaining
relationships (Thomas et al., 2013). Larger networks with weaker ties would also be more
conducive of innovation behaviours and product trials (Scott, 1991). Network theory
suggests that a community needs a critical mass of users who generate content and canvass
new members in order to be successful (Preece and Maloney-Krichmar, 2005). On the
other hand, a smaller community would perform better in terms of collective identity,
identification with the focal object and member equality (Bagozzi and Dholakia, 2006).

Departing from Muñiz and O’Guinn’s (2001) initial claim that communities are stable and
enduring, other researchers contend that the temporality of brand communities is dynamic
(McAlexander et al., 2002) and that it is to be considered on a case-by-case basis.
Community duration, thus, becomes an important feature. This dynamic and contextual
dimension of temporality is further accentuated in the OBC literature, since online media
allows instant creation and obsolescence of communities. Despite evidence against
longevity, it is still appreciated that communities surpass other forms of consumer
groupings in terms of stability, and this is a differentiating factor with tribes and
subcultures of consumption (Goulding et al., 2013). Factors such as heterogeneity of
 23

actors, collaboration and resource sharing are elements that help sustain brand
communities in the long term (Thomas et al., 2013).

Members’ relationships in (online) brand communities can vary according to a number of

aspects and relationship types are an important facet of community. These elements
encompass the relationships strength, structure, and the partners involved. The strength or
richness of relationship has been considered in a series of studies, and there is a general
consensus on its variability (McAlexander et al., 2002; Porter, 2004; Dholakia et al., 2004;
Fowler and Krush, 2008; Heere et al., 2011). More specifically, different elements are put
forth as impacting the strength of the relationship between members, such as their level of
identification to each other (Heere et al., 2011) or the size of the community being small or
based on a large network of ties (Dholakia et al., 2004). Brand communities might also be
made up of several tightly knit subgroups linked together by weaker ties (Fowler and
Krush, 2008).

Another aspect of the relationships in brand communities is their structure. Fournier and
Lee (2009) provide a typology of relationship structures, classifying them as hubs, pools
and webs. Webs are composed of people who have strong one-to-one relationships with
others who have similar needs. Pools, on the other hand, gather people who have strong
association with a shared activity, goal or values but loose associations with one another.
Lastly, in hubs, people have strong connections to a central figure and weaker associations
with one another. These types of community formations impact the way that information is
stored, dispersed and received in relational dynamics (Thomas et al., 2013), as well as the
quantity and quality of information exchanged (McAlexander et al., 2002). Depending on
how the relationship structure is set up, it might be easier or harder for new members to
join.

Although these classifications of the type of relationship might have their shortcomings,
they provide an idea of the different ways customer can gather around a brand, and build
on the premises that relationship and relationship quality exists between consumers and
corporations, or brands (Fournier, 1998). Different entities or relationship partners are part
of OBC, as proposed by McAlexander et al (2002): the customer, the product, the brand
and the company. The way relationships between these partners have been conceptualised
is addressed further in this chapter.
 24

The final feature of community concerns space. The spatial or geographical aspect of
brand communities is a recurring theme in the literature. There is an agreement that
different types of communities exist based on their spatial constraints. Brand communities
can be fully offline and relying only of face-to-face member interactions, they can be fully
online, where members never see each other, or a combination of both (e.g. Muñiz and
O’Guinn, 2001; McAlexander et al., 2002; Thomas et al., 2013; Wirtz et al., 2013). For
instance, McAlexander et al. (2002) explain that brand communities are dynamic when it
comes to their geographic concentration and online or offline presence. Brand
communities can be geographically concentrated, such as local Harley-Davidson Owner
groups (Algesheimer et al., 2005), or scattered, such as Warhammer communities, which
communicate across borders (Cova et al., 2007).

Information technologies play a crucial role in facilitating the geographic dispersion of

brand communities (Lin, 2007), and a large number of initially offline communities are
now complemented by an online presence. Similarly, the OBC literature equally evidences
that online interactions can, and are often, supplemented by offline relationships (Wirtz et
al., 2013). It is increasingly accepted that community ties can exist between individuals
that never see each other, and that they can share a psychological sense of community
which is as strong as in real-life environments (Carlson et al., 2008). Overall, brand
communities often exhibit a mix of online and offline interactions, allowing them to be
placed on a spectrum ranging from being fully online to being fully offline (Stockburger-
Sauer, 2010).

To summarise, this section has shown that (online) brand communities share a number of
core characteristics, largely rooted in the fact that they are both forms of brand-focused
consumer communities. Offline and online brand communities can however vary
according to a number of factors, including their size, duration, marketplace orientation,
relationship structure and space attributes. The following section focuses on the role played
by information technologies and how they set aside OBCs form offline brand communities
in a number of ways.
 25

2.2.3. Online specificities of OBCs

Communities have evolved past a stage of geographic embeddedness to become freed from
geographic constraints (Castells, 1996). The modern context characterised by mobility,
technology and multiculturalism affects the way communities are positioned in space and
time. Although they have long been considered as relying mainly on a common location
and physical proximity (Tönnies, 1887) the rise of globalisation and information
technologies has introduced a new global and transnational dimension to communities
(Delanty, 2003). Communities are now freed from geographic boundaries, spatial structure
and time constraints (Castells, 1996; Urban, 1996). Communities based on physical
proximity still exist, as the extant literature on offline brand community strongly
evidences, but physical proximity is no longer seen as a requirement for brand community
development, and OBCs are now in the forefront of community research in marketing
(Faraj and Johnson, 2011).

Online communities are aggregations of individuals in the cyberspace, where social

relationships are mediated by highly personalised technology (Rheingold, 1993; Castells,
1996; Mathwick et al., 2008; Fournier and Avery, 2011). Sproull (2003, p. 733) defines an
online community as a ‘large, voluntary collectivity whose primary goal is member or
social welfare, whose members share a common interest, experience or conviction, and
who interact with each other primarily over the Net’.

The raison d’être of these virtual groups of individuals—what brings them together—is,
like in brand communities, a shared interest, purpose, experience or goal (Williams and
Cothrel, 2000; Rothaermel and Sugiyama, 2001; Faraj and Johnson, 2011), which makes
members take part in the community voluntarily. This link among individuals is the
starting point for the community and the glue that holds online community members
together. The fundamental premises of offline communities therefore seem to hold in
online settings. This application of the offline definition to an online context is appropriate
to some extent, as the same characteristics of brand communities identified by Muñiz and
O’Guinn (2001) work to produce a brand community in the online context (Szmigin et al.,
2005; Madupu and Krishnan, 2008; Mathwick et al., 2008).

Despite the clear correspondence between offline and online brand communities, the
online aspect is a game-changer for the study of brand communities. Online communities
depart from their offline counterparts in several ways.
 26

Elimination of space and time boundaries represents a key distinguishing feature.

Internet mediation allows OBC to overcome same-place, same-time limitations inherent in
face-to-face settings (Andersen, 2005; Faraj and Johnson, 2011). In OBCs, brand fans from
across the globe can meet in one single virtual location with the sole requirement of having
Internet access, and sometimes, access to a specific platform as well.

OBCs have significantly increased the speed, convenience, reach and transparency of
communication by creating another complementary form of reality (Rheingold, 1993; Katz
and Rice, 2002) all of which lead to increased social interactivity. OBCs are recognised
as appropriate media for building consumer-brand relationships through new means of
communication and increased interactivity (Thorbjørnsen et al., 2002, Quinton, 2013).
Indeed, consumers are increasingly active participants in interactive online processes
involving multiple feedback loops, and almost synchronous communication (Hoffman and
Novak, 1996). Online tools are said to ‘enable and facilitate new and extended forms of
interactive consumer experiences, which may contribute to the development of customer
and/or consumer engagement with specific brands’ (Brodie et al., 2013).

Taking the opposite stance, some argue that online environments fail to convey rich
situational and interaction cues (e.g. facial expressions) due to their asynchronous
characteristic and are therefore inadequate for tasks associated with complex meaning and
reciprocal feedback (Faraj and Johnson, 2011). However, this thinking was based on the
idea that online communities are only using text-based communication and that no
feedback loop exists. Far from assuming that online communication is flawless, more
recent research shows that the variety of content type that can now be exchanged online, as
well as the possibility for users to share interests and bond in ever more complex ways
evidence the opposite trend (Fournier and Avery, 2011). Although social interactivity on
online communication technologies can be challenging, it is evident that it sustains
interactive environments where human social behaviour has ample opportunities to
develop (Murphy et al., 2014).

Technology allows sustaining bigger groups of individuals from dispersed locations (Katz
and Rice, 2002), and facilitates the broadcasting of information, which is then easily stored
and retrievable leading to broader reach. Marketers and consumers alike are empowered
by the Internet’s levels of reach and transparency (Kozinets et al., 2010). Online social
interactions are growing in size and relevance and consumers’ influential power is growing
exponentially (Trusov et al., 2010).
 27

Improved accessibility represents another distinguishing feature. OBCs are found on a

plethora of computer-mediated platforms and virtual spaces (Hagel and Armstrong, 1997;
Mathwick et al., 2008; Wang and Chen, 2012). Researchers have identified different online
tools or services that enable virtual communities, such as chat rooms, mailing lists (Kim et
al., 2008), blogs, forums (Shen et al., 2010), social networking sites (Porter et al., 2011) or
other forms of social media. Overall, the technological advances of the Web 2.0. (O’Reilly,
2007) are characterised by social media, which represent all web-based applications that
allow the creation and exchange of user-generated-content (Kaplan and Haenlein, 2010).

The presence of OBCs on so many different virtual platforms and outlets makes their
usage easier and more empowering to some extent for both marketers and consumers
(Kozinets et al., 2010), as technology is adaptable and transformable by users in order to
meet their personal needs. This is reflected in the use of a technological platform, which
has specific functioning, rules and interface. However, online platforms also bring with
them specific affordances, modes of functioning and interaction, which can be challenging
to use and apprehend (Dholakia and Reyes, 2013).

Online technologies and social media in particular have spurred both collective and
individual forms of consumer power resulting in increased consumer power (Labrecque
et al., 2013). This is a direct consequence of the above-mentioned factors: better access to
information across space and time, enhanced social interactivity and broader reach work
together to fuel consumer empowerment in the digital age (ibid, 2013). The voice of the
consumer becomes more forceful, as evidenced by user-generated content (Christodoulides
et al., 2011) and there is an obvious power shift from the marketers to the consumer
(Mathwick, 2002; Fournier and Avery, 2011). Understanding and accepting this shift in
power balance seems to be the key to successful brand management in the future, as social
media support consumer interactions that are often outside of the control of brand
managers (Quinton, 2013).

These various characteristics inherent to the online nature of OBC show that Internet
mediation brings unique features that are not found, at least with such intensity, in offline
brand communities. As Kozinets (1999) argues, ‘online interactions are becoming an
important supplement to social and consumption behaviour’. This tendency has only
grown stronger in the last two decades and has vast implications in terms of consumer
behaviour (Mathwick, 2002). OBCs thus require a specific treatment in comparison to
offline communities (Hagel and Armstrong, 1997).
 28

To summarise, OBCs hold clear parallelisms with brand communities when it comes to
their core characteristics, but their online nature strongly impacts their functioning, making
them unique and requiring a dedicated treatment. OBC participation is impacted by a
combination of elements which stem, on the one hand, from OBCs’ similarities with their
offline counterparts, and on the other hand, from their unique online characteristics.
Figure 1 summarises these similarities and differences.

Figure 1: OBC similarities with brand communities and online specificities

2.2.4. Classification of OBCs

The multiple advantages that virtual settings offer have fostered the emergence of a vast
amount of online consumption and brand communities (Kozinets, 1999). Examples of
online communities centred on consumption activities include wristwatch enthusiasts
(Rothaermel and Sugiyama, 2001), health and beauty conscious consumers (Kim, et al.,
2008), or coffee lovers (Kozinets, 2002). OBC research, on the other hand, has
investigated brands such as Timezone.com (Rothaermeal and Sugiyama, 2001), Ford and
Mustang cars (Dholakia et al., 2004) or Coca-Cola (Sicilia and Palazón, 2008).

Similarly to brand communities, OBCs have been categorised in typologies by a handful of

authors, putting forth their virtual nature. Dholakia et al. (2004), for instance, identify two
kinds of online communities based on their size, geographic dispersion and strength of
relationships: (1) Network-based virtual community, which are specialized,
geographically dispersed community, based on a structured, relatively sparse, and dynamic
network of participants and (2) Small-group-based virtual community, constituted by
individuals with a dense web of relationships, interacting together as an online group, in
order to achieve a wider range of joint goals, and to maintain existing relationships

Stanoevska-Slabeva and Schmidt (2001) take another approach by putting forward the aim
of the community and type of transaction. Four types of online communities exist under
 29

this typology: (1) Discussion communities: dedicated to exchange of information with

reference to a defined topic; (2) Task- and goal-oriented communities: striving to
achieve a common goal by way of cooperation; (3) Virtual worlds: providing a virtual
setting of complex world; (4) Hybrid: containing several types of communities.

Similarly, Kozinets (1999) emphasises the focus of the community as well as its social
structure. He claims that the group focus can be either information exchange or social
interaction. The social structure on the other hand can be loose of tight. This approach
results in four different categories of virtual communities of consumption: boards, rooms,
rings and lists, and dungeons.

Recently, Breitsohl et al. (2015) proposed a taxonomy of online consumption

communities, differentiating them on the basis of two criteria: (1) The community’s
content focus, or orientation (brand versus consumption activity) and (2) The
community’s type of host, or governance (the brand versus consumers). Their results show
that consumers’ posting behaviour is driven by different motives in different community
types: for instance, brand communities rely on consumers’ need to help others with brand-
related content. Overall, they report that there needs to be strong congruency between the
content posted on the community and the community purpose (Breitsohl et al., 2015).

To summarise the concept and characteristics of OBCs, research shows that OBCs hold
clear correspondence with offline brand communities regarding the core characteristics of
communities (Muñiz and O’Guinn, 2001). Their online nature however endows them with
specific characteristics that set them apart from offline contexts. Moreover, due to the
rapidly changing nature of the technological environment, it is hard to rely on one single
typology for the understanding of OBCs, or rely on it at all, due to the rapidly changing
nature and variety of criteria to include.

Research into OBCs therefore constantly needs to be in tune with the latest technological
advances and requirements, as well as have a strong understanding of the brand
community literature on which the foundations of OBC research lie.
 30

2.3. Consumer participation in Online Brand Community

This section addresses OBC participation. As the above definition and characterisation of
OBC evidenced the relevance of both offline and online characteristics, the brand
community and OBC-specific literature are reviewed conjointly.

2.3.1. Defining OBC participation

Consumer participation in OBC is a clear necessity for community survival (Koh and Kim,
2004; Casaló et al., 2007; Koh et al., 2007), and therefore been the focus of much research
in the past 15 years. Member participation sustains the development of members’
consciousness of kind (Casaló et al., 2010) and their relationship with the brand (Andersen
2005, Casaló et al., 2008). Participation is a reflection of communities’ success and
member satisfaction (Yoo et al., 2002; Casaló et al., 2008), and it fosters and enriches
members’ experiences (Muñiz and O’Guinn, 2001). In addition, studies suggest that
participation can have an impact beyond the community boundaries (Muñiz and O’Guinn,
2001; Andersen 2005). For these reasons, the study of OBC participation has been high on
the agenda of OBC researchers as a way to sustain, add value and vitalise the community.

Despite this intense interest in OBC participation, there is little emphasis on reaching a
common understanding of what OBC participation constitutes. To date, this field of
enquiry is composed of a plethora of theories, methods, research designs and worldviews,
providing an extremely intricate account of what OBC participation really represents.

The following section provides a critique of studies that concentrate on understanding,

characterising and modelling OBC participation. The aim of this section is to provide an
overarching view of the literature and highlight gaps. An overview of the key studies
addressing this issue is presented in Table 2. The selection of these studies rests on six
inclusion criteria: (1) A focus on online or offline brand communities; (2) A timeframe
starting from the first seminal study on brand community in 2001 until 2015; (3)
Emanating from journals strictly in the area of marketing which are (4) Ranked 3 or 4 stars
in the 2010 ABS list; (5) Dealing exclusively with brand-centric communities of consumer
and (6) Adopting a supportive rather than oppositional stance toward the brand and
marketplace. Supporting studies not referenced in this list are also used as secondary
sources.
 31

Table 2: Overview of selected OBC studies

Community Focus of Paradigmatic Design/ Core theories/

Paper Approach Research objectives
space the study stance Methods frameworks

Post-
Muñiz and Off and Interviews Sociology
Participation modernism/ Qualitative Reveal the idea of brand community
O'Guinn, 2001 online Netnography Relationship marketing
Interpretivism
McAlexander et Participation Ethnography Develop and integrated framework of brand
Offline Pragmatism Mixed Relationship marketing
al., 2002 Outcomes Survey community relationships (IBC)
Cross- Evaluate the role of group norms and social identity
Dholakia et al., Antecedents SIT
Online Positivism Quantitative sectional and consider their motivational antecedents and
2004 Participation UGT
(survey) mediators
Netnography Examine how a grassroots brand community
Muñiz and
Online Participation Interpretivism Qualitative (observation, CCT responds to the loss of the brand upon which it is
Schau, 2005
interviews) centred
Antecedents Cross- Model how different aspects of customer's
Algesheimer et SIT
Offline Participation Positivism Quantitative sectional relationships with the brand community influence
al., 2005 Relationship marketing
outcomes (survey) their intentions and behaviours
Case study
Analyse the power that a virtual brand community
Cova and Pace, Off and Participation Post- (Interviews, Collective consumer
Qualitative exerts over a brand of a mass-marketed
2006 online Outcomes modernism netnograghy, empowerment theory
convenience product
documents)
Antecedents Cross- TPB Investigate the social and psychological antecedents
Bagozzi and Off and
Participation Positivist Quantitative sectional Social intentions of group- and brand-related behaviours of small
Dholakia, 2006 online
Outcomes (survey) SIT group consumer communities
SIT Evaluate the effect of BC participation and
Thompson and Participation Longitudinal
Online Positivist Quantitative Product diffusion theory membership duration on the adoption of new
Sinha, 2008 Outcomes (survey)
Relationship marketing products from the preferred and opposing brands
Antecedents Cross- Explore the process of how a firm’s online
SIT
Kim et al.2008 Online Participation Positivist Quantitative sectional community enhances consumers’’ brand
Reciprocal action theory
Outcomes (survey) commitment
 32

Community Focus of Paradigmatic Design/ Core theories/

Paper Approach Research objectives
space the study stance Methods frameworks

Ouwersloot and Cross- Determine if a community population can be

Antecedents Relationship marketing
Odekerken- Offline Positivism Quantitative sectional meaningfully segmented based on different
Participation UGT
Schröder, 2008 (survey) motivations to join
Determine the effect of brand-related and
Antecedents Cross- Creativity theory
Füller et al., individual-specific traits in driving members'
Offline Participation Positivism Quantitative sectional Relationship marketing
2008 willingness and ability to engage in new product
Outcomes (survey) Personality traits
development
Cross-
Carlson et al., Participation Relationship marketing Determine the role of psychological sense of brand
Offline Positivism Quantitative sectional
2008 Outcomes SIT community on brand commitment
(survey)
Social Case study Reveal common processes of value creation among
Schau et al., Off and
Participation constructivis Qualitative (Interviews, Social practice theory networked firm-facing actors in brand-centred
2009 online
m netnograghy) communities
Netnography, Relationship marketing
Adjei et al., Participation Determine how C2C communications influence
Online Pragmatism Mixed Survey and Interpersonal
2010 Outcomes firms and how are firm goals reached
Experiment communication
Relationship marketing How can marketing management tools strengthen
Stokbürger- Off and Antecedents Experiments
Positivism Quantitative SIT brand communities by facilitating shared customer
Sauer, 2010 online Participation Survey
Theory of social capital experiences and multi-way interactions
SIT
Antecedents Cross- The effect of individual factors on brand
Matzler et al., Personality trait
Offline Participation Positivism Quantitative sectional community identification and product attachment in
2011 framework Relationship
Outcomes (survey) generating brand trust and loyalty, in a BC context
marketing
Case study
Social
O'Sullivan et Off and (observation, Explore the processes contributing to the genesis of
Antecedents constructivis Qualitative Consumer culture theory
al., 2011 online interviews, a brand community
m
netnography)
 33

Community Focus of Paradigmatic Design/ Core theories/

Paper Approach Research objectives
space the study stance Methods frameworks

Cross- Understand the role of interpersonal trust and

Hung et al., Antecedents Source credibility
Online Positivism Quantitative sectional platform credibility on consumer search
2011 Participation framework
(survey) and consumption behaviours in online communities
Antecedents
Cross- Examine the ways in which existing community
Heere et al., , Relationship marketing
Offline Positivism Quantitative sectional identities affect identification with a brand
2011 Participation SIT
(survey) community
Outcomes
Investigate the relative impact of community
Cross-
Marzocchi et Participation Relationship marketing identification vs company identification in
Offline Positivism Quantitative sectional
al., 2013 Outcomes SIT generating loyalty, through brand affect and brand
(survey)
trust, in an BC context
Cross-
Zhou et al., Antecedents Social cognition theory Investigate the transformation mechanisms that
Online Positivism Quantitative sectional
2013 Participation UGT convert visitors into members
(survey)
Cross- Understand how OBC types are likely to be
Gruner et al., New product
Online Outcomes Positivism Quantitative sectional associated with the success of new products in
2014 development
(survey) terms of sales volume and market shares
Cross-
Examine how consumers react to firms' active
Homburg et al., Participation Post- sectional
Online Quantitative Relationship marketing participation in consumer-to-consumer
2015 Outcomes positivism (survey) and
conversations in an online community setting
qualitative
 34

As evidenced in the previous section, the study by Muñiz and O’Guinn (2001) represents
the foundation of OBC and brand community literature and serves as a reference point for
many scholars. With the aim to reveal the idea of brand community, they refer to
community participation as a social construction largely resting on imagined social
relationships amongst members as well as three fundamental principles (see section 2.2.2.
above). These characteristics are evidenced through interviews and netnography of offline
and online manifestations of the same three communities in North America. From a
conceptual standpoint, this study relates largely to the sociology literature and communal
consumption studies. References are made to brand relationships literature too, making a
powerful case for the potential relational benefits of brand communities and signalling the
need for further investigation (Muñiz and O’Guinn, 2001). By revealing instances of social
relationships about a brand, this study constitutes a first departure point from the traditional
model of consumer-brand relationships, where one-to-one company-consumer ties are the
norm (Fournier, 1998).

This study has led to a number of related studies using similar approaches. Muñiz and
Schau (2005) also use a qualitative approach and adopt comparable methods (netnography,
observations, participant observation, interviews) to investigate OBC participation,
although focusing on a single case study. They examine how a grassroots brand
community responds to the loss of the brand upon which it is centred, namely the Apple
Newton personal digital assistant. They interpret community participation as
transformative experiences akin to religious involvement and evidence the existence of
Muñiz and O’Guinn’s (2001) three community markers. Their study is more closely
anchored in consumer culture theory (CCT), and it rests on a narrative analysis of
consumers’ tales.

Taking a similar case study approach and using interviews, netnograghy and document
analysis, Cova and Pace (2006) investigate OBC participation for the Nutella brand in
Italy. Their worldview is however different from Muñiz and colleagues, as they are clearly
anchoring their study in the postmodern marketing paradigm. They look at both offline and
online manifestations of the Nutella community participation. Taking a theoretical
perspective resting on the ‘collective consumer empowerment theory’ (Cova and Pace,
2006), they find that OBC participation shows a new form of sociality and consumer
empowerment resting on personal self-exhibition rather than social interactions. They
 35

however converge with Muñiz and O’Guinn (2001) in evidencing the role of rituals and
traditions engrained in OBC participation.

The theory of social constructivism in OBC participation is explored further by Schau et

al., (2009) who also relate it to the social practice theory. They offer a list of twelve
practices through which a community is enacted and which create value for the community
and the brand. These practices are evident in both online and offline settings, and deriving
from interviews, netnograghic and participant observation data. The practices are:
welcoming, empathising, governing, evangelising, justifying, documenting, badging,
milestoning staking, customising, grooming and commoditising. These practices are
grouped in four categories, which are: social networking, impression management,
community engagement and brand use (Schau et al., 2009). Through such practices, Schau
et al. (2009) show that OBC members co-create their individual and social OBC
experience and re-negotiate the brand value (see also Cova and Pace, 2006). Figure 2
below details the iterative process through which these four groups of twelve practices co-
exist to create value in the community.

Figure 2: Schau et al. (2009) framework of brand community practices

 36

This view of OBC participation as social constructed through practices is lastly taken on by
O’Sullivan et al. (2011), using the same conceptual and methodological approaches. They
focus on the Beamish beer brand in Ireland and, reflecting Muñiz and Schau (2005) and
Schau et al.’s (2009) studies, find out that OBC participation is made up of transcendent
sacralised experiences maintained through rituals and social practices.

In conclusion, several common features distinguish the qualitative studies of community

participation. In conceptual terms, the studies build largely on social constructivism and
consumer culture theories. Community participation is viewed as an ever-evolving,
experiential and iterative process whereby members make sense of the community and the
brand, which resonates with the ontological considerations made by social constructivism.
Members co-create value through rituals (Muñiz and O’Guinn, 2001) and negotiate the
meaning of the brand and group itself (Muñiz and Schau, 2005).

Congruent to this conceptual approach, these studies are based on inductive and iterative
approaches and rely strongly on qualitative methodologies involving case studies and
making use of netnographic, observational and interview data (e.g. Muñiz and O’Guinn,
2001; Kozinets, 2002; Muñiz and Schau, 2005). Netnography is considered to be suited to
the investigation of meaning and practices in online environments (Kozinets, 2002).
Through interviews and observational data, these studies also rely on OBC members’
narratives to understand brand meaning co-production in offline and online brand
communities (Muñiz and Schau, 2005; see also Kozinets et al., 2010). Considering the
amount of papers included in table 2, these studies however represent a small portion of the
total (5 out of 23).

Scholarship on brand communities seems dominated by deductive and quantitative

approaches anchored in the relationship marketing paradigm and the seminal work of
McAlexander et al. (2002). McAlexander et al. (2002) expand the triadic view of brand
relationship in brand communities proposed by Muñiz and O’Guinn (2001) and introduce
the concept of integration into a brand community (IBC) as a consumer-centric view of
brand relationship, which enacts OBC participation. For them, brand community supports
more complex webs of relationships and go beyond consumer-brands relationships. They
include links established between the consumers and other consumer, between the
consumer and the brand, but also between the consumer, product and firm.
 37

Figure 3 shows the evolution of customer-brand relationships in the brand community

literature as presented by McAlexander et al. (2002). The study shows that the more points
of attachment and experience with brand-related attributes such as the firm or the product,
the more the consumer feels integrated in the community and the stronger his/her
emotional and behavioural attachment to that brand is (McAlexander et. al, 2002).

Figure 3: Model of relationships in brand communities by McAlexander et al. (2002)

In contrast to Muñiz, O’Guinn and their colleagues, McAlexander et al. (2002) use mixed
data in the form of an ethnography and a large-scale survey, evidencing a more pragmatic
worldview. Their data are collected in North America at brandfests focused on the Jeep
and Harley-Davidson brands.
 38

A number of studies follow this relational view of OBC participation, in both offline and
online spaces. For instance, more recent research has verified that brand communities
foster these multi-way relationships with brands, consumers, products and the company
(Ouwersloot and Odekerken-Schröder, 2008; Stokurger-Sauer, 2010). After McAlexander
et al. (2002) most studies focusing on relationships in OBC use exclusively quantitative
data, mainly in the form of cross-sectional consumer surveys, assuming a largely positivist
take on the subject. They expand on the understanding of what a relationship in OBC is
and how it characterises participation. Algesheimer et al. (2005, p. 23), for instance,
defined brand-relationship quality in a brand community context as the ‘degree to which
the consumer views the brand as a satisfactory partner in an ongoing relationship’, which
represents an overall assessment of the strength of the relationship with the brand. Using a
web-based survey sent to German members of brand communities in the automotive
industry, they evidence the central role of consumer relationship with the brand and the
community in driving their intentions and behaviours.

Several authors have attempted to refine the meaning of relationships in OBCs from a
commitment perspective. In this stream of literature, community participation takes an
emotional bend and focuses on members’ attachment and affect toward the community
(Kim et al., 2008 - see also Mathwick et al., 2007; Lin, 2010; Wang and Chen, 2012). OBC
commitment refers to a member’s attitude toward the community, reflected by a high
degree of positive feelings (Jang et al., 2008). Commitment has been captured by various
emotional and psychological expressions such as a sense of belonging (Carlson et al.,
2008), a degree of emotional attachment and trust toward the community, and a need to
participate (Jang et al., 2008). Similar to Jang et al. (2008), Lin (2010) defines community
commitment as an affective manifestation of attachment to the community and
distinguishes it from community loyalty. Community commitment is reflective of the
‘stickiness’ of the community (Wang and Chen, 2012).

This stream of literature depicts OBC participation using a relational approach showing
how deeply anchored OBCs are within the relationship marketing paradigm. Building on
McAlexander et al. (2002), brand community studies have helped expand the realm of
brand relationships from one-to-one consumer-company interactions to a social, multi-
actor context (Szmigin et al., 2005; Veloutsou, 2007). This stream of literature views OBC
as social structures for the development of relationship between consumers, marketers and
 39

brands, as well as an opportunity to widen the traditional two-way consumer-brand

relationship to the social realm (McAlexander et al., 2002).

Another strand of literature in OBC participation has focused on social identification

theory (SIT). Dholakia et al. (2004) lead this stream of thought by centring their research
on social identity and group norms as two group-level characteristics of virtual community
participation. They adopt a positivist approach and use cross-sectional survey data
collected from North-American members of OBCs on different platforms to understand the
mechanisms that lead to OBC participation as understood from a social identity
perspective. A similar approach has been used by Algesheimer et al. (2005), although
targeting a German sample of automobile fans, and they define brand community
identification as a process whereby a person ‘construes himself or herself to be a member –
that is “belonging” to the brand community’ (p. 20). Later on, Bagozzi and Dholakia
(2006) show further validation of social identification as a form of community
participation, by surveying members of Harley Davidson and other non-Harley riders
groups in North America. More recently, Heere et al. (2011) evidence the existence of
social community identification among members of university football teams in North
America as well, using the same data collection methods.

The social identity approach to OBC participation is anchored in social psychology

theories and uses quantitative confirmatory methods to evidence its existence, antecedents
and outcomes. These methods are sometimes complemented by a behavioural approach to
participation, which emerged with it and grew in usage later on.

A purely behavioural approach to OBC participation is the last form of participation that
features in the literature. This view embraces the idea that participation percolates through
members’ actions, active states and activities within and with the community. It
emphasises measurement of participation based on members’ measurable actions. Driven
by a number of action-based principles, the behavioural approach is a widely accepted
view of OBC participation.

There is some debate whether participation represents a collective or individual endeavour

or whether participation is active or passive. For some, participation is best understood as
collective, or group action (Bagozzi and Dholakia, 2006), stemming from the social
psychology perspective. Others prefer measuring participation on an individual basis (Kim
et al., 2008). Concurrently, participation can also be active or passive (Koh et al., 2007).
 40

When active, it takes the form of community interactions, communication and part-taking
in activities (Kim et al., 2008). Posting and creating content (User Generated Content) is
probably one of the most active kinds of behavioural participation (see Koh and Kim,
2004; Cooke and Buckley, 2008; Fournier and Avery, 2011), as well as taking part in
group activities (Stokbürger-Sauer, 2010). Other studies argue that members can be
passive and only engage in lurking behaviour. Lurking is the passive behaviour of reading
others’ posts without contributing to the community (Nonnecke et al., 2006).

In more recent studies focusing on online settings, participation has sometimes been
approached from a purely quantitative, metrics-based approach, stressing the amount of
time spent on the community and number of views (Trusov et al., 2010), the membership
length (Mathwick et al., 2008), the number of ‘likes’, comments or visits on a Facebook
page (Gummerus et al., 2012; Chauhan and Pillai, 2013). An even more condensed
approach to OBC participation refers to the mere fact of being affiliated to the community,
and it is used in a number of studies as well (Füller et al., 2008; Thompson and Sinha,
2008; Zhou et al., 2013). All these metrics are easily gathered from the consumers
themselves, but can also easily be collected from website data.

Studies from the strictly behavioural approach to participation, whether they view it as
passive or active, collective or individual, and online or offline, all agree that these can be
quantitatively measured either through consumer self-reported data, experiments or
website metrics. It is important to note that studies looking at social constructivism as a
form of OBC participation also consider practices and behaviours as part of participation.
They however view OBC participation as a dynamic and continuous process (composed
partly of actions), rather than actions measurable at a point in time, which represents an
important ontological difference with the strictly behavioural approach.

Overall, the review of the different approaches to OBC participation reveals, in many
respects, a fragmented and complex stream of literature. Numerous approaches seem to
exist –the boundaries of which are however unclear– and they exhibit differences in terms
of paradigms, approaches, design and methods, and theories used. For example, major
ontological and epistemological differences exist across different approaches. Although the
epistemological stance of the studies is rarely explicitly mentioned, a review of the papers
highlighted paradigmatic preferences ranging from social constructivism (e.g. Muñiz and
O’Guinn, 2001; O’Sullivan et al., 2011) to postmodernism (Cova and Pace, 2006),
pragmatism (McAlexander et al., 2002) and positivism (Heere et al., 2011; Hung et al.,
 41

2011; Mazorcchi et al., 2013). These varying worldviews are crucial to determining the
conduct of the studies and therefore shape the literature as an intricate mix of methods and
approaches.

The literature evidences a plethora of theories used. Preece and Maloney-Krichmar (2005)
summarised the state of the literature rather adequately by saying ‘no particular theory or
set of theories currently dominates research on online communities. Rather, we see an
application of different theories (…)’ (p.00), an assertion which characterises the OBC
literature of the past 15 years. The theories most frequently used range from social
constructivism and consumer culture theory (e.g. Schau et al., 2009), to relationship
marketing theories and frameworks (Carlson et al., 2008), also making copious use of the
social identification theory (e.g. Bagozzi and Dholakia, 2006). Other theories also inform
OBC participation, such as the UGT (Dholakia et al., 2004), product diffusion theory (e.g.
Thompson and Sinha, 2008), TPB (Bagozzi and Dholakia, 2006), or social cognition
theory (Zhou et al., 2013). These last few theories however pertain more to the drivers and
outcomes of participation rather than capturing participation itself, and are thus addressed
later in this chapter.

Notwithstanding the relatively small number of contributions, it seems that the

methodological approach taken by most studies published in top journals is mainly
quantitative (e.g. Dholakia et al., 2004; Algesheimer et al., 2005), less often qualitative
(e.g. Muñiz and Schau, 2005), and on more scarce occasions, based on a mixed approach
(e.g. McAlexander et al., 2002; Adjei et al., 2010). The underlying assumptions signal
preference for deductive and confirmatory procedures, while induction, although strongly
defended in certain studies (Schau et al., 2009) seems to have gained less traction. The
overall picture is yet that of a disputed view on the appropriate approach to capture OBC
participation.

Accordingly, the type of data collected varies. Authors focusing on the purely behavioural
aspect of OBC participation have focused on survey data (e.g. Matzler et al., 2011;
Mazorcchi et al., 2013), similarly to those investigating social identification (e.g. Dholakia
et al., 2004) and participation behaviours. Studies taking a relationship stance toward
participation have also relied largely on survey data, also using ethnographic data at times
(e.g. McAlexander et al., 2002), and experiments on occasions (Stokbürger-Sauer, 2010).
Quantitative data analyses procedures vary, although they mostly rely on a form of
multivariate data analysis technique. When adopting a qualitative approach, single or
 42

multiple case studies were often preferred, mixing a series of data types ranging from
(n)ethnography, (participant) observations and interviews (e.g. Schau et al., 2009). These
are often analysed through interpretation or narrative structures. Data choices are driven by
the aim of the studies, their philosophical stances, as well as the conceptual frame and
theories that they rely on.

2.3.2. Key issues in the OBC participation literature

Overall, the literature attempting to capture OBC participation is characterised by a clear

lack of consensus on the appropriate way to approach this phenomenon. The review of the
literature suggests that there is an extremely fragmented treatment of this question. This
overall fragmentation is due to a plethora of co-existing, yet non-communicating
paradigms, theories, approaches, methods of data collection and analyses used. Although
the fragmented state of OBC research is recognised (Preece and Maloney-Krishmar, 2005),
it constitutes one of the major challenges in moving this agenda forward. For example, it is
unclear from the review whether OBC participation is a matter of sociological
identification (Bagozzi and Dholakia, 2006), behavioural manifestations (Gummerus et al.,
2012), emotional commitment (Kim et al., 2008) or something else. It is also unclear
whether OBC participation can be condensed to one single type of manifestation, like
identification alone, or whether several manifestations need to be combined (Casaló et al.,
2008). Capturing OBC participation therefore seems to be an exercise so far composed of a
multiplicity of uni-faceted and unidimensional views that fail to communicate or be
integrated.

Additionally, the literature remains undecided on whether a social or individual treatment

of OBC participation may be most appropriate. Despite an almost complete agreement that
sociality sits at the core of OBC participation across worldviews (Muñiz and O’Guinn,
2001; McAlexander et al., 2002; Bagozzi and Dholakia, 2006; Schau et al., 2009), ways of
capturing participation constantly shift from a social to an individual approach. The social
identification perspective naturally adopts a social perspective, as well as studies led by a
social constructivist view. However, these perspectives are challenged by research that
measures OBC participation as individual actions or emotions (e.g. Cova and Pace, 2006;
Füller et al., 2008), leaving the social to be a backdrop element of lesser importance. Given
the inherent and necessary social nature of OBC contexts, it is therefore surprising that
OBC participation would not be systematically approached as well as measured taking into
account this social perspective.
 43

Moreover, although most studies recognise the need to include and implement a social
aspect to OBC participation, attempts to capture OBC participation are often irresponsive
of the interactive and conversational nature of online contexts. A core distinctive
characteristic of online versus offline brand communities remains largely unaddressed in
this research. Conversations have only begun to emerge as an important aspect of OBC
participation and its measurement very recently Hombourg et al. (2015). Prior to this, only
the social constructivist approach to OBC participation viewed it as a socially interactive
concept (e.g. Schau et al., 2009). On the other hand, most positivists captured OBC
participation as individual actions, behaviours or sentiments (e.g. Carlson et al., 2008; Kim
et al., 2008; Thompson and Sinha, 2008). Even when acknowledging the social referent of
individual identification processes (with brand community identification or brand
identification in Dholakia et al., 2004, for instance) this social referent is not granted an
active role. This is evident in the way that Dholakia et al. (2004) express social identity: as
a psychological state, of an emotional and evaluative significance. Therefore, even in the
most social way to measure OBC participation to date, there seems no interactive aspect to
it, suggesting that the social referent of social identity in OBC (i.e. the community, or the
brand) does not need to be an active relationship partner for identification to occur.

To summarise, OBC participation is treated in a fragmented and incomplete manner,

leaving no indication as to its actual dimensions and manifestations. Moreover, it suffers a
lack of social referent in the way it is measured and, when social referents exist, there is a
lack of interactivity in the way it is approached. These issues not only weaken the
understanding of OBC participation, but they also show how much it is highly irresponsive
of the specificities of online contexts that make OBCs different from brand communities.

The overall OBC literature seems to strongly rely on seminal studies, theories and
approaches that were developed in offline contexts, such as McAlexander et al. (2002),
Muñiz and O’Guinn (2001) or Dholakia et al. (2004). Offline studies and frameworks,
methods and measures that work in offline contexts are often straightforwardly reproduced
in online contexts (e.g. Bagozzi and Dholakia, 2006), as evidenced by the lack of
interactivity of OBC participation measures denoted above. Furthermore, when attempting
to measure OBC participation in an online context, and in an online specific manner, OBC
studies often fall into the trap of using website metrics (e.g. Gummerus et al., 2012) which
evidence a highly unidimensional, behaviour-only measure of participation which are
organisation- rather than consumer-based.
 44

These problems demonstrate the need to account for OBC participation in a way that is not
only still in line which the core characteristics of offline brand communities, but also, and
most urgently, responsive of the highly interactive, social and rich online media. The next
section reviews the concept of consumer engagement and proposes it as a solution to the
current flaws of the literature on OBC participation.

2.4. Consumer engagement as an approach to OBC

participation

The aim of this section is to present consumer engagement as an alternative approach to

OBC participation, its strength residing in its clear potential to tackle the current
shortcomings of the OBC literature identified above. Given the relative novelty of the
concept, a somewhat divergent conceptual approach underlies the treatment of engagement
in OBC and in marketing studies at large, which this section aims to untangle. To do so, a
review of the broader marketing literature on engagement is presented first, also building
on other fields of the social sciences, as it paves the way to the specific treatment of
consumer engagement as an approach to OBC participation.

2.4.1. Defining consumer engagement

The concept of engagement has a long history. Marketing research on engagement has only
emerged in the last decade, whereas seminal research on this topic had started as early as
the early nineties (e.g. Kahn, 1990) in other fields of social sciences. Recognising this
theoretical and empirical precedence, as well as a high degree of transferability across
domains, marketing researchers (such as Bowden, 2009; Brodie et al., 2011; Vivek et al.,
2012) have thus tended to rely on expertise in other fields to ground their understanding of
the concept.

Over the last two decades, the term ‘engagement’ has been the subject of academic enquiry
in the fields of psychology (Achterberg et al., 2003), sociology (Jennings and Stocker,
2004), information systems (Wagner and Majchrzak, 2007), political sciences (Resnick,
2001), educational psychology (Bryson and Hand, 2007) and organisational behaviour
(Kahn, 1990). Although its interpretations are not always consistent across and within
 45

disciplines, a review of various conceptualisations in the social sciences is deemed

essential to assert the grounds of this study’s understanding of consumer engagement. An
overview of key studies in social sciences is thus proposed in this section, as well as key
takes from these studies that bring light on how to approach engagement in marketing.

Each of the social sciences discipline that has embraced the concept of engagement has
naturally done it with respect to discipline-relevant subjects and objects. Psychology has
examined the concepts of ‘social engagement’ and its group-oriented dimension (e.g.
Achterberg et al., 2003) along with ‘occupational engagement’ (e.g. Bejerholm and
Eklund, 2007). ‘Civic engagement’ has been the subject of investigation in sociology
(Jennings and Stocker, 2004), while ‘student engagement’ has attracted the attention of
scholars in the field of education (Bryson and Hand, 2007; London et al., 2007). Student
engagement literature interestingly suggests the existence of a continuum for the levels of
engagement and, hence, potential for student disengagement. In organisational behaviour,
the construct of ‘employee engagement’ has been reported as a state of connection to one’s
work and others (Kahn, 1990) which is linked to increased job satisfaction, low
absenteeism and high commitment (Salanova et al., 2005). Organisational behaviour, on
the other hand, is concerned with the concept of ‘stakeholder engagement’, as a context-
specific democratic process of involving all stakeholders of a company (Grudens-Schuck,
2000). Furthermore, the term ‘state engagement’ is specific to the political sciences where
a notion of interdependence with the engagement object exists (Resnick, 2001). Finally and
more closely linked to the subject of this thesis, the notion of ‘customer engagement’ is
addressed in the information systems literature. It represents a behavioural manifestation of
knowledge exchange between customers, companies and other customers (Wagner and
Majchrzak, 2007).

Table 3 below presents an overview of the engagement literature in the social sciences,
evidencing the rich pedigree of the concept in other areas than marketing. The overview
builds on existing reviews of the literature in marketing papers (Bowden, 2009; Brodie et
al., 2011; Vivek et al., 2012).
 46

Table 3: Definitions and dimensionality of engagement in the social science disciplines

Discipline Concept Authors Definition /Findings Dimensions

Achterberg et al., A high sense of initiative, involvement and adequate response to social stimuli,
Behavioural
Social 2003 participating in social activities, interacting with others.
engagement Emotional
Huo et al., 2009 Represented by group identification and group-oriented behaviour.
Psychology Behavioural
A lifestyle characteristic that describes the extent to which a person has a balanced
Occupational Bejerholm and Cognitive
rhythm of activity and rest, a variety and range of meaningful occupations/routines
engagement Eklund, 2007 behavioural
and the ability to move around society and interact socially.
Cognitive
Civic Jennings and Involvement in voluntary organisations and the performance of volunteer work,
Sociology Emotional
engagement Stoker, 2004 facilitating the development of social networks.
Behavioural
Stakeholder Grudens-Schuck, A democratic, context specific process that expresses aspirations of people and
Behavioural
engagement 2000 organisations that stakeholders are legitimised as important meaning-makers.
Political Resnick, 2001 Iterative process aiming to influence political behaviour of a target state through
sciences State maintained contacts with that state across multiple issue areas and focused on Behavioural
engagement generating a relationship of increasing interdependence

Information Customer Wagner and The intensity of customer participation with both representatives of the organisation
Behavioural
systems engagement Majchrzak, 2007 and with other customers in a collaborative knowledge exchange process.
Cognitive
Bryson and Hand, On a disengaged-engaged continuum, a student may exhibit differing engagement
Emotional
2007 levels to a particular task/assignment, module, course of study and Higher Education
Educational Student Behavioural
psychology engagement Student’s academic investment, motivation and commitment to their institution; Cognitive
London et al.,
perceived psychological connection, comfort, and sense of belonging toward their Emotional
2007
institution Behavioural
 47

Discipline Concept Authors Definition /Findings Dimensions

Catteeuw et al., The degree to which employees are satisfied with their jobs, feel valued and Cognitive
2007 experience collaboration and trust. Emotional
Cognitive
Employee’s desire/willingness to give discretionary effort in their jobs, in the form of
Frank et al., 2004 Emotional
extra time, brainpower/energy (includes cognitive, affective and behavioural aspects).
Behavioural
Kahn, 1990 The simultaneous employment and expression of a person’s ‘preferred self’ in task Cognitive
behaviours that promote connections to work and to others personal presence and Emotional
active, full performances. Physical
Organisational Employee
behaviour engagement Saks, 2006 Employee engagement is the amount of cognitive, emotional and behavioural Cognitive
components that are associated with individual role performance through attitudes, Emotional
intentions and behaviours. Behavioural
Schaufeli et al., A pervasive affective-cognitive state that acts to enhance organisational productivity.
Cognitive
2002 Positive, fulfilling, work-related state of mind that is characterised by vigour,
Emotional
dedication and absorption.
Salanova et Engagement acts to increase group morale, cohesion and rapport via positive
al., 2005 psychological contagion processes. It results in job satisfaction, low absenteeism, Emotional
organisational commitment and performance.
 48

Studies in the social sciences highlight several core characteristics of engagement that are
also evident in the marketing literature. Firstly, engagement involves different actors, or
parties. Engagement acts upon a specific subject (employee, student, stakeholder), with
respect to a specific engagement object, or focus (state, function, studies), denoting the
context-specific nature of engagement (Achterberg et al., 2003). In this sense, engagement
is inherently social and relational, since engagement cannot occur without the presence of
both parties (the student and his studies, the employee and his job, the citizen and the
state). Engagement relies on interactive connections (Kahn, 1990) and stimulus by another
party (Catteuw et al., 2007).

Secondly, engagement is a context-specific concept. Engagement has been shown to

emerge in a variety of different disciplines, as exemplified in the vast social science
literature. Additionally, engagement can manifest itself in a variety of social contexts
within the same discipline (e.g. Saks, 2006; London et al., 2007, Bakker et al., 2011). This
suggests that context-specific factors need to be taken into account when studying
engagement, relating directly to the engagement actors: as the context of engagement
changes, the actors involved change as well. For instance, in an organisational context,
employees are the subjects of engagement, whereas their engagement focus is their job or
task.

Thirdly, engagement is composed of a set of specific dimensions. As Saks (2006)

explains, employee engagement is the amount of cognitive, emotional and behavioural
components that are associated with individual role performance through attitudes,
intentions and behaviours. The cognitive, emotional and behavioural aspects of
engagement are often acknowledged (e.g. Jennings and Stoker, 2004), although there
seems to be no consensus on which dimension(s) to consider, denoting the dimensional
blur in which the concept stands across and within disciplines.

Fourthly, engagement has different levels of intensity, and a valence. Engagement can be
high or low. People can exhibit strong or weak engagement on a permanent basis, but
engagement level can also vary for the same individual through time (e.g. Catteeuw et al.,
2007; Bryson and Hand, 2007). Engagement can reflect a positive (Schaufeli et al., 2002)
or negative condition.

Although the engagement concept has received substantial attention across various
different academic fields, it is still in its development stage in the marketing literature
 49

(Hollebeek, 2011a; Hollebeek et al., 2014). The paucity of studies to date, and particularly
empirical studies, incurs a lack of consensus on the nature and scope of engagement in the
field. Given the conceptual fog surrounding the concept, it is not surprising that several
calls for a clearer definition of engagement have been made in the last few years (e.g.
Hollebeek, 2011b; Vivek et al., 2012).

Consensus concerning certain features of engagement is accompanied by disagreements.

For example, whether engagement constitutes a state or process is a first point of
dissention among marketing scholars. Bowden (2009, p.65) defines engagement as: ‘The
psychological process that models the underlying mechanisms by which customer loyalty
forms for new customers of a service brand, as well as the mechanisms by which loyalty
may be maintained for repeat purchase customers of a service brand.’ Implicitly, this
perspective models engagement as an overarching process comprised of the relationships
between customer commitment, trust and involvement, as impacting their loyalty levels.

No other evidence of considering engagement as a process was found in the marketing

literature, other than in Brodie et al.’s (2011) work, which however seems to hesitate in
classifying customer engagement (CE) as a state or process, using both terms in the same
definition (p. 260):

‘CE is a psychological state that occurs by virtue of interactive, cocreative customer

experiences with a focal agent/object (e.g. a brand) in focal service relationships. It
occurs under a specific set of context-dependent conditions generating differing CE
levels and exists as dynamic, iterative process within service relationships that
cocreate value. CE plays an essential role in a nomological network governing
service relationships in which other relational concepts (involvement, loyalty...) are
antecedents and/or consequences in iterative CE processes. It is a multidimensional
concept subject to a context and/or stakeholder- specific expression of relevant
cognitive, emotional and/or behavioral dimensions’.

Process-oriented studies imply that there is an iterative and cyclical nature to engagement.
Bowden (2009) in particular sees engagement as a higher-order phenomenon overarching a
series of engagement-building steps. In contrast to the limited support for the process
view, the notion of ‘engagement as a state’ seems to be widely accepted in the literature.
Supporters of the state view operationalise engagement as existing at certain intensity, at a
specific point in time (Patterson et al., 2006). Under this premise, engagement can
 50

therefore be operationalised and measured as a stand-alone construct, rather than composed

of a sequence and combination of others.

Both approaches have advantages and disadvantages. For example, it can be argued that
the state approach fails to encompass the dynamic and iterative nature of engagement, and
its variance for different segments of customers (Bowden, 2009). As a state, however,
engagement has antecedents, outcomes, a specific level, or intensity, and thus the potential
to be quantitatively operationalised, and measured as such (see Vivek et al., 2014;
Hollebeek et al., 2014). It also opens avenues for treating engagement as part of a
nomological network of relationships between relationship marketing constructs, discussed
further in this chapter.

2.4.2. The subject of engagement

As previously clarified, engagement involves both a subject (the engaged entity) and a
partner (the focus or object of engagement). There is a broad agreement that the unit of
analysis in marketing should be the individual customer (e.g. Bowden, 2009; Verhoef et
al., 2010; Hollebeek, 2012), or consumer (e.g. Brodie et al., 2013; Calder et al., 2013;
Wallace et al., 2014). Although the initiator of engagement can be the company, through
specific offerings or activities, the person whose engagement matters to marketing
researchers and practitioners alike is, understandably, the individual customer or consumer,
terms that are used interchangeably in the literature (Vivek et al., 2012). For the sake of
clarity, this review uses the word ‘consumer’ to refer to the subject of engagement, a
semantic choice explained in the next chapter.

2.4.3. The partners of engagement

Consumer engagement is interactive by nature; hence engagement can only happen if there
is a relationship partner to interact with, and use as an engagement referent. Although
studies sometimes refer to engagement ‘objects’ or ‘focuses’, this study adopts the term
‘partner’, as reflective of the interactive role of this engagement referent in OBCs.
Different positions exist regarding the relevant engagement partners to consider in
marketing, and this issue seems to be largely bound by the context in which engagement
occurs. The company or organisation is a recurring engagement partner (e.g. Patterson et
al., 2006; Verhoef et al., 2010), as well as the service offering or product (Vivek et al.,
2012). Extending beyond the product as engagement partner, Van doorn et al. (2010)
contend that the partner can be either the firm or the brand, so long as the relationship goes
 51

beyond the transaction. In fact, the prevalence of the brand as engagement partner is quite
clear. Out of 33 marketing articles in table 4, 19 of them include the brand as an
engagement partner. In fact, Gambetti and Graffigna (2010, p. 819) state that ‘customer-
brand engagement is the only really significant concept when considering engagement
from the marketing perspective’. Other engagement partners for consumer or customers
include types of media (Calder et al., 2009), or piece of advertising/content (Phillips and
McQuarrie, 2010).

Another engagement partner identified in the marketing literature, even before scholar
started paying much attention to engagement with a brand, is the brand community. Posing
the brand community as an engagement partner goes back to acknowledging the social
dimension of engagement and it is bound by community contexts (Vivek et al., 2014).

Brand community engagement was first put forward by Algesheimer et al. (2005, p. 21) as
‘the positive influences of identifying with the brand community through the consumer’s
intrinsic motivation to interact/co-operate with community members’. The group and social
aspect of engagement has thereafter been recognised by engagement scholars as an
important venue for increased levels of engagement (e.g Achterberg et al., 2003; Sawhney
et al., 2005; Gummerus et al., 2012), warranting a certain, yet limited attention to brand
community engagement.

Lastly, it is worth pointing out that most studies concentrate on one partner of engagement,
and there are very few studies that acknowledge multiple partner of engagement, at least
simultaneously. Gambetti and Grafigna (2010), provides a review of engagement and
recognise different engagement partners, supporting the idea that engagement can go in
different directions. Brodie et al. (2011) and Vivek et al., (2012) both support this premise
and evidence it with qualitative data. The former study focuses on engagement with a
brand and with other community members, while the latter includes all organisational
offerings or activities as potential engagement partners. This last proposition is verified
quantitatively in a study by Vivek et al. (2014).

Table 4 below provides an overview of the current studies in marketing that have focused
on consumer engagement. It details the type of paper and construct of interest as coined by
the authors and supports the assertion that consumers, or customers, are the preferred
engagement subject in marketing, whereas a wide array of engagement partners have been
considered, and mostly one at a time.
 52

Table 4: Key consumer engagement studies in marketing

Authors Paper type Construct Subject Partner Dimensions

Patterson et al., 2006 Conceptual Consumer engagement Consumer Service organisation Absorption, dedication, interaction, vigour
Bowden, 2009 Conceptual Consumer engagement process Consumer Service brand N/A
Sprott et al., 2009 Quantitative Brand engagement in self concept Consumer Brand Emotional
Mollen and Wilson, Conceptual Engagement Consumer Brand Affective, cognitive
2010
Engagement with a brand/firm/organisation

van Doorn et al., 2010 Conceptual Consumer engagement Consumer Brand or firm Behavioural
behaviours
Verhoef et al., 2010 Conceptual Consumer engagement Consumer Brand or firm Behavioural
Kumar et al., 2010 Conceptual Customer engagement value Customer Brand Behavioural, emotional
Brodie et al., 2011 Conceptual Consumer engagement Consumer Service Behavioural, cognitive, affective
brand/organisation
Hollebeek, 2011a Conceptual Consumer-brand engagement Consumer Brand Behavioural, cognitive, affective
Hollebeek, 2011b Qualitative Consumer-brand engagement Consumer Brand Behavioural, cognitive, affective
So et al., 20 13 Quantitative Consumer engagement Consumer Brand Behavioural, cognitive, affective
Gambetti et al., 2012 Qualitative Consumer-brand engagement Consumer Brand Experiential, social
Kaltcheva et al., 2014 Conceptual Customer engagement Customer Service firm Behavioural, cognitive, affective
Franzak et al., 2014 Conceptual Brand engagement Consumer Brand Behavioural, cognitive, affective
Hollebeek and Chen, Qualitative Brand engagement Consumer Brand Behavioural, cognitive, affective
2014
Hollebeek et al., 2014 Quantitative Consumer brand engagement Consumer Brand Behavioural, cognitive, affective
Jaakkola and Qualitative Consumer engagement behaviour Consumer Service brand Behavioural
Alexander, 2014
Sarkar and Sreejesh Quantitative Active customer engagement Customer Brand Behavioural and cognitive
2014
Wallace et al., 2014 Quantitative Consumer engagement Consumer Brand Behavioural
 53

Wirtz et al., 2013 Conceptual Online brand community Customer Online brand community members Behavioural, cognitive, affective
Engagement with a
brand community
engagement
Algesheimer et al., Quantitative Brand community Customer Brand community Motivational
2005 engagement
Gummerus et al, 2012 Quantitative Consumer engagement Consumer Community Behavioural
Kuo and Feng, 2013 Quantitative Brand community Customer Brand community Interactive
engagement
Habibi et al., 2014 Qualitative Brand community Customer Brand community Practices
Engagement with another type of

engagement
Higgins and Scholer, Conceptual Consumer engagement Consumer Task Sustained attention
2009
Calder et al., 2009 Quantitative Consumer engagement Consumer Media (website) Experiential, social
Phillips and Qualitative Engagement Consumer Advertising Behavioural, affective, immersive,
entity

McQuarrie, 2010 transporting, identification

Scott and Craig-Lees, Quantitative Audience engagement Audience Entertainment piece Emotional
2010
Calder et al., 2013 Quantitative Consumer engagement Consumer Product or service Civic, identity, intrinsic enjoyment,
social, utilitarian

Gambetti and Review Engagement Consumer Multiple entities N/A

Graffigna, 2010
Engagement with
multiple partners

Brodie et al., 2011 Qualitative Consumer engagement Consumer Brand and/or community members Behavioural, cognitive, affective

Vivek et al., 2012 Qualitative Consumer engagement Consumer Organisational offering or activities Behavioural, cognitive, affective, social

Vivek et al., 2014 Quantitative Consumer engagement Consumer Organisational object, consumption Behavioural, cognitive, affective, social
activity or event
 54

2.4.4. The dimensions of engagement

Both unidimensional and multidimensional definitions of consumer engagement exist in

the marketing literature. For example, Bowden (2009) contrasts engagement for new and
repeat customers as being respectively more cognitive or affective. Following a similar line
of thoughts, Gambetti and Graffigna (2010) put the relational (i.e. soft, affective)
dimension of engagement at the opposing end of a continuum with behavioural (i.e.
pragmatic, managerial) engagement. Other authors favour a unidimensional representation
of engagement, by emphasising its behavioural (van Doorn et al. 2010; Verhoef et al.,
2010; Gummerus et al., 2012) or emotional aspects (Sprott et al., 2009).

Despite the existence of a few unidimensional views of engagement, the majority of

customer engagement definitions in the marketing literature adopt a multidimensional
perspective, with combinations of two- or three-dimensional, as evidenced in table 4. This
approach is consistent with Saks’ (2006) multidimensional vision of employee
engagement. Hollebeek (2011a, p. 790), for instance, defines customer-brand engagement
as ‘characterized by specific levels of cognitive, emotional and behavioural activity in
direct brand interactions’. Extant research also identifies dimensions related to identity,
vigour, civism or absorption (Patterson et al., 2006; Calder et al., 2013). Going further into
this multidimensional perspective, Vivek et al. (2012) complement the cognitive-
emotional-behavioural triad by adding a social dimension to the equation. This fourth
dimension of engagement is supported by their qualitative empirical research involving
executives in various fields, and theoretically grounded within the service dominant logic
(Vargo and Lusch, 2004).

This varied treatment of the dimensions of consumer engagement is not surprising given
the emerging nature of the concept itself, and it provides a fertile ground for further
exploration. Although various dimensions have been suggested, a significant proportion of
the published work emanates from, or relates to Brodie and Hollebeek’s work (Brodie et
al., 2011; Hollebeek, 2011; Brodie et al., 2013; Hollebeek et al., 2014). They conceptualise
engagement as multi-dimensional construct with a cognitive, an affective and a
behavioural dimension, a view embraced by number of other engagement studies (e.g.
Mollen and Wilson, 2010; Wirtz et al., 2013). This view is also the one adopted in this
thesis.
 55

2.4.5. The valence of engagement

Much research in marketing and the social sciences focuses on engagement has as a
positive construct (e.g. Schaufeli et al., 2002; Hollebeek et al., 2014). However, faced with
increasingly critical and demanding consumer, marketing scholars also acknowledge the
existence of negative forms of engagement (Hollebeek and Chen, 2014), as well as
disengagement (Bowden et al., 2014). The difference between negative engagement and
disengagement is that negatively engaged consumers maintain a relationship with the
brand, whereas disengagement leads to the termination of the relationship (Bowden et al.,
2014). van Doorn et al. (2010) consider that the valence of engagement is one of its
primary properties, and that engagement has an equal potential to be positive or negative.

Both positive and negative aspects of engagement matter because they can have
dramatically diverging consequences for the brand, particularly in the context of OBCs
(Hollebeek and Chen, 2014). As suggested by Kumar et al. (2010), negative forms of
engagement such as negative Word of Mouth (WOM) can be detrimental for companies.
They recommend to track and address complaints behaviours in order to minimize the
effects of negative engagement (Kumar et al. 2010). Hollebeek and Chen (2014) support
these assertions with exploratory data, explaining that the valence of engagement affect the
outcome valence of brand attitude and WOM. In support to this, conceptual research also
shows that any form of customer behaviours (e.g. blogging, reviewing, recommending)
and emotions can take a negative valence in the instance of a poor fit between the
consumer and the brand, and as a result impact negatively the latter (van Doorn et al.,
2010). Scholars are encouraging further research to investigate both positive and negative
valences of engagement and how consumers behave toward their favourite but also least
favourite brands (Sprott et al.,2009; Gummerus et al.,2012; Hollebeek and Chen, 2014).

2.4.6. Engagement in the online brand community literature

The origins of the consumer engagement view in the OBC literature can be traced back to
the other approaches to participation. More specifically, the ‘practices and co-creation’
view detailed above, along with other key studies in the brand community literature
(Algesheimer et al., 2005; Brodie et al., 2011; Wirtz et al., 2013), show the emergence of
the engagement concept in OBC contexts.

In their article on community practices, Schau et al. (2009) use the term ‘engagement’ 77
times, refraining from using ‘involvement’ or ‘participation’ altogether. According to
 56

them, community engagement practices are those ‘that reinforce members’ escalating
engagement with the brand community’ (ibid., p 34), in the form of documenting, badging,
milestoning and staking brand use (see Figure 2). They also explain that each individual
practice has an anatomy, composed of understandings, or the knowledge and tacit cultural
template; procedures, or explicit performance rules and engagement, or emotional projects
and purposes (Schau et al., 2009). Interestingly, this anatomy of practices parallels the
three dimensions of consumer engagement, as the first one reflects cognition, the second
behaviours and the third is reflective of affect.

Schau et al., (2009) go further to explaining that, as practices evolve and are integrated
over time, they reinforce consumer engagement, which leads to further enactment of the
practices and value creation, both for the brand and the community. This account of
community practices gives indications as to the importance of engagement for community
sustenance and value. It suggests that community engagement is a form of social practice
related to experiential and shared brand use (Schau et al., 2009), but also that engaging
with a practice constitutes the emotional and motivational aspect of it.

Earlier on in the brand community literature, Algesheimer et al. (2005) had already
introduced the concept of community engagement, in a similar light, focusing on the
motivational aspect of the concept as well, but as a measureable variable part of a
nomological network of causal relationships. They defined community engagement as ‘the
intrinsic motivations to interact and cooperate with community members’ (p.21).

In this study, community engagement is an outcome of brand community identification. It

leads to increased levels of normative community pressure, as well as a number of
behavioural intentions: membership continuance, community recommendation and
community participation. The research indicated the potential key role of engagement in
creating consumer behaviour in the OBC context, as shown in Figure 4 below.
 57

Figure 4: Brand community model in Algesheimer et al. (2005)

The authors however acknowledge the limitations of their conceptualisation of engagement

and call for a better understanding of community engagement in brand community settings
(Algesheimer et al., 2005).

More recently, Brodie et al. (2013) made the relevance of the consumer engagement
construct in the OBC context much more explicit. They propose a framework of
engagement, its dimensions and objects within an OBC context depicted in Figure 5 below.
 58

Figure 5: Framework of consumer engagement in virtual brand communities by

Brodie et al. (2013)

They also provide a working definition of consumer engagement in an OBC context (p. 3):

‘Consumer engagement in a virtual brand community involves specific interactive

experiences between consumers and the brand and/or other members of the
community. Consumer engagement is a context-dependent, psychological state
characterized by fluctuating intensity levels that occur within a dynamic, iterative
engagement process. Consumer engagement is a multidimensional concept
comprising cognitive, emotional and/or behavioral dimensions, and plays a central
role in the process of relational exchange where other relational concepts are
engagement antecedents and/or consequences in iterative engagement processes
within the brand community’.
 59

This definition points out the unique characteristics of engagement as a way to capture
cognitive, behavioural and emotional aspects of interactive community participation.
Interestingly, these conceptualisations have been adopted in recent studies on Facebook
pages for instance, which characterise consumer engagement in these settings as a form of
interactive community participation (Janh and Kunz, 2012; Blazevic et al., 2014). To
illustrate, Brodie et al.’s (2013) study also recognises the existence of engagement in a
network of relationships with other relational variables, which have been the subject of
much debate in understanding the drivers and outcomes of OBC participation and are
detailed in the following section. Although possibly the most advanced study of consumer
engagement in OBC, the study suffers from several shortcomings. It relies on exploratory
approach, which is congruent with the stage of engagement research at the time, but limits
the applicability of the findings. The empirical data concern only one community (a
vibration training brand) and the data set consists of netnographic evidence and a limited
number of end-user depth-interviews (six in total). Consequently, it seems clear that
further research into engagement in OBC is warranted and that the concept bears high
potential into explaining OBC participation mechanisms beyond existing approaches (van
Doorn et al., 2010; Porter et al., 2011; Brodie et al., 2013).

In summary, this section has evidenced that consumer engagement offers evident strengths
in comparison to previous approaches to OBC participation, particularly through its social,
interactive and multi-dimensional aspects. Before elaborating further on this, a review of
the antecedents and outcomes of OBC participation, including the OBC and consumer
engagement literature is warranted to address the state of the literature related to this
study’s second research objective.

2.5. Antecedents and outcomes of OBC participation/

consumer engagement

This section aims to critically assess the current state of research in determining what
motivates and results from OBC participation. As evidenced in the section on OBC
participation, studies use different paradigms or worldviews, approaches, methods, data
types, theoretical frameworks and analysis tools, making up for an extremely rich stream
 60

of literature. For the sake of clarity, the antecedents of OBC and brand community
participation are reviewed first, followed by the outcomes. The review pays specific
attention to existing empirical models of participation.

Since the concept of consumer engagement is put forth as a way to advance OBC
participation, the models of antecedents and outcomes of consumer engagement are also
reviewed and evaluated. The literature on consumer engagement being less developed and
still more conceptual/exploratory than the OBC literature, the second part of this section
largely includes conceptual research or studies using qualitative data, and it presents
antecedents and outcomes of consumer engagement concurrently.

2.5.1. Antecedents of participation in the OBC literature

In terms of the antecedents of OBC participation, the relationship marketing and brand
relationship literature in particular can again serve as guiding frame (Algesheimer et al.,
2005). Research showed that existing consumer-brand relationships in brand communities
are an important vector of individual and collective identity construction and consumer-
brand relationship is viewed as an antecedent to brand community identification. For
Algesheimer et al. (2005), brand relationship quality leads to stronger identification to
the community (according to some, a form of OBC participation, as discussed earlier) and
subsequently community engagement. Accordingly, people are able to identify with the
community through the prior social relationships that they have with the brand (Kozinets,
1999). Interactions within the community realm help individuals build their individual and
collective self, a concept based on the social identification theory by Tajfel and Turner
(2004). In the same line of thought, Füller et al. (2008) show the impact of brand trust
and brand passion in consumer’s willingness to involve in brand community-based
innovation processes. Similar to Algesheimer et al. (2005), data were gathered in Germany
from a brand community in the automotive industry. These studies evidenced the central
impact that consumer-brand relationship plays on OBC functioning and sustenance, and
therefore on OBC participation.

Brand identification is a frequent antecedent of participation. This concept is often

used in brand-relationship literature (e.g. Tildesley and Coote, 2009) and deriving from the
social identification theory. Brand identification is also often cited as an antecedent of
OBC participation where members identify with the brand and with other community
members (Algesheimer, et al., 2005; see also Reto, 2012), and against out-groups.
 61

Existing conceptualisations of brand identification in the brand community literature tend

to take a psychological view to the concept, defining it as ‘the extent to which the
consumer sees his or her own image as overlapping with the brand’s image’ (Bagozzi and
Dholakia, 2006, p.49). Brand identification can also have social impacts, based on the
brand’s ability to act as a communication instrument, allowing the consumer to associate
or dissociate himself from a reference group (De Rio et al., 2001). The brand therefore acts
as a social marker of identity, which can be referred to as the brand signalling capacity
(Tildesley and Coote, 2009). This social aspect of brand identity is particularly suited to
research contexts where consumers are part of a group, such as OBCs (McAlexander et al.,
2002).

Brand identification is also an important part of the construction of the self (Belk, 1988),
building on the symbolic meaning of brands as ways to enhance one’s identity. Based on
Lasswell’s identification theory through a symbol (Lasswell, 1965), brand identification
can be seen as the perception of sameness between the brand and the consumer, where the
brand is an object with symbolic meaning that the consumer perceives as his own (Tuškej
et al., 2013). This approach however fits better within research contexts where the
individual is not part of a social group (Sen and Bhattacharya, 2001), where the perception
of the group is psychological only (Carlson et al., 2008), or where the consumer is not
working collectively toward group goals, or interacting with a group at all (Kuenzel and
Halliday, 2010).

The relationship between identification and engagement is not clear. Consumer-brand

identification has been found to be antecedent of community participation in some studies
(McAlexander et al., 2002), while other advocate for a moderating role of brand
identification between participation and its antecedents (Bhattacharya and Sen, 2003).
Others simply state that there is a correlation between the two phenomena, without arguing
for a directional causal relationship (Bagozzi and Dholakia, 2006). A slightly different
view is taken by Heere et al. (2011) who show that identification of the brand community
members with a city or state also works to produce stronger community affiliation. In the
context of university sports teams, the city or state can be considered as a brand. It is
interesting to note that all these studies emanate from North American research settings,
focusing on brand categories typically high in involvement (automotive and sports clubs).

Uses and Gratification Theory (UGT) also provides a rich source of antecedents to OBC
participation. The UGT originated from the communication and media literature as a way
 62

to understand people’s motivations for using specific media (McQuail, 1987). It rests on
the principle that people seek value in media usage and that this perceived value drives
them in their media interactions (ibid, 1987). McQuail (1983), who proposes five reasons
that drive media usage: (1) Information; (2) Personal identity; (3) Integration; (4) Social
interaction and (5) Entertainment. This media theory has been adapted to brand community
and OBC research in multiple instances, starting with Dholakia et al. (2004) (see also
Raacke and Bonds-Raacke, 2008; Pak et al., 2009; Calder et al., 2013). The basic premise
here is that members engage in OBC on the basis of perceived benefits, or value, that they
get from this participation (e.g. Wiertz and de Ruyter, 2007). Perceived value in the
context of OBC can therefore be defined as a community member’s overall assessment of
the utility of the OBC he is a part of, based on perceptions of benefits received from
community membership (Zeithalm, 1988).

For example, Dholakia et al. (2004) proposed a rather comprehensive model of the reasons
why people join and participate in virtual brand communities, based on the following types
of value they perceive: (1) Purposive; (2) Self-discovery; (3) Maintaining interpersonal
interconnectivity; (4) Social enhancement and (5) Entertainment. Purposive value is
composed of informational and instrumental value. Information value refers to getting and
sharing information, and knowing what others think (Rothaermel and Sugiyama, 2001;
Raacke and Bonds-Raacke, 2008), whereas instrumental value means accomplishing
specific tasks such as solving a problem, generating an idea, influencing others or
validating a decision. Self-discovery value is the understanding of salient aspects of one’s
self through social interactions. Self-discovery pertains to goal attainment and elaboration
of one’s values. Maintaining interpersonal interconnectivity means establishing and
maintaining contact with other people such as social support, friendship and intimacy. It is
different from social enhancement, which relates to gaining acceptance and approval from
other members and enhancing one’s social status on account of one’s contribution to it.
Lastly, entertainment value is the acquisition of fun and relaxing time through playing or
otherwise interacting with others.

The UGT has been used in multiple studies on OBC in recent years. For instance,
Gummerus et al. (2012) focus on Facebook brand community membership and refer to
social, entertainment, economic, practical and social enhancement value. Entertainment
and social enhancement values are defined similarly to Dholakia et al., (2004), and
practical value refers to their purposive value. Social value is similar to their maintenance
 63

of interpersonal connectivity although it goes further by including the brand as a

relationship partner. The authors also add another element, which is economic value and
pertains to gaining discounts, saving time, taking parts in raffles or competitions. These
material rewards are explicit incentives given by OBC managers to fuel member
participation (Garnefeld et al., 2012). Linking the UGT with social cognition and
observation learning theories, Zhou et al. (2013) show in a Chinese context that
informational and social value perceived from content posted on the OBC increase
members’ participation intentions.

To sum up, various perceived gratifications and benefits exist for members of OBC and
these are largely validated as key drivers of member participation. Benefits can be derived
from informational or cognitive value, social value, and personal status or
hedonic/entertainment value. These elements result in increased interaction and
participation in the OBC (Nambisan and Baron, 2007).

Yet another theory used to identify OBC participation antecedents is the Theory of
Planned Behaviour (TPB), which is an extension of the Theory of Reasoned Action
(TRA). The theory of reasoned action (TRA) was developed by Fishbein and Ajzen (1975)
with the intention to explain human behaviour. According to this theory, behaviour is
dependent on intentions to act, which are themselves a function of attitude toward the
behaviour and subjective norms (ibid., 1975). Attitudes are evaluative reactions to an
action and reflect predispositions to respond in a favourable or unfavourable manner
(Eagly and Chaiken, 1993). Subjective norms reflect the impact of expectations from other
people and are based on a need for approval (Bagozzi and Dholakia, 2006). TPB (Ajzen,
1991) builds on the TRA by adding an additional variable to predict intentions to act: the
perceived behavioural control the individual feels he/she has over the action. The TPB
explains that behaviours are directly influenced by one’s intention to act which itself
depends on three different factors: attitudes toward the act, subjective normative pressure
to act and perceived behavioural control (ibid, 1991).

The TPB has been applied in the context of OBC in order to explain OBC participation
(i.e. group behaviour). Critiquing its individual-oriented approach and the lack of
emotional variables, Bagozzi and Dholakia (2006) proposed a slightly revised version of
the TPB to apply to the context of OBC. They showed that group behaviour in an OBC is
influenced by social intentions (a socially revised version of individual intentions), which
itself depends on desires to act. The desire to act is driven by the attitude toward the act,
 64

positive and negative anticipated emotions (added to the initial TPB model) and subjective
norms. Similarly, Zhou (2011) combined elements of the TPB with social identity theory
and showed that subjective norms influence intentions to act and subsequent participative
behaviours in OBC.

The Technology Acceptance Model (TAM) is another adapted version of the TRA, which
is particularly relevant to virtual environments. The theory ignores the role of subjective
norm but argues that perceived usefulness and perceived ease of use are two main factors
affecting user acceptance of an information technology (Davis, 1989). With a focus on
information technologies, it has proven particularly suitable to the study of OBC
behaviour. In a Spanish context, Casaló et al. (2010) combine the TAM, TPB and social
identity theories to predict OBC participation in the tourism industry. They show that
attitude toward participation, perceived behavioural control, perceived usefulness and
identification with the online community all work toward consumer’s intentions to
participate.

Other studies in OBC contexts suggest that the way people participate in OBC is reliant not
only on social, relational, or technology-related attributes, but can also emanate from
members’ own traits and inner selves, therefore focusing on individual traits and
predispositions (e.g. De Valck et al., 2009). A first study placing important focus on
individual consumer traits and characteristics is Füller et al. (2008). In the context of an
offline German automotive brand community, they evidence that community members’
inherent creativity traits such as task motivation, skills, innovativeness and task
involvement determine their willingness to engage in open innovation projects through the
community.

Using the personality trait framework, Matzler et al. (2011) provide further evidence that
individual traits impact OBC participation. Using a similar context to Füller et al. (2008)
(German, offline, automotive industry), they evidence that consumers’ extraversion
influences identification with the community which in turns impacts brand loyalty.
Personality congruity influences product attachment, which in turn impacts brand loyalty
and brand trust. Matzler et al.’s (2011) study also shows that individual traits can be mixed
with brand relationship antecedents (brand trust and passion) in models predicting OBC
participation. This study is echoed by other research showing that inherent traits such as
demographics, personal preferences, traits or attributes do impact OBC participation
(Fournier et al., 2009, Chang et al., 2013). Similarly, Wiertz and de Ruyter (2007) advance
 65

online interaction propensity as an important individual trait, which can be seen as a proxy
for extraversion, and impacts OBC participation.

The literature suggests that several membership types exist, based on individual profiles
and characteristics, which determine the way people participate in OBC. Fournier and Lee
(2009), for instance, argue that brand communities can only be robust if they understand
people’s lives and that each of them has a role in the community. They identify 18
different roles (from mentor to learner and from celebrity to greeter), explaining that the
role one assumes can depend on age, gender, lifestyle, life cycle and occupation.

Following the same reasoning, De Valck et al. (2009) create a typology of six types of
OBC members: (1) Core members; (2) Conversationalists; (3) Informationalists; (4)
Hobbyhists; (5) Functionalists and (6) Opportunists. This classification is based on a
number of elements constituting the membership profile of individuals. Among other
criteria, De Valck et al. (2009) give a lot of importance to individual consumer
characteristics and traits in their classification. They include various demographic and
socioeconomic variables, as well as consumer’s Internet usage profile, their orientation
towards others, and their opinion leadership and expertise with respect to the community's
topic of interest. Consumer’s sociability and influenceability are also considered.

More recently, Chang et al. (2013) identifies an extensive set of individual traits that
influence the OBC members’ propensity to receive and send information. They classify
these driving traits under three headings: elementary (such as extraversion and
neuroticism), compound (such as need for information) and situational (such as
innovativeness and value consciousness). The study involved members of different
branded Facebook pages (Starbucks, 7–Eleven, Eslite Bookstore, and Books Online
bookstore). Their results show that information sharing is largely driven by a compound,
three-level series of antecedents made of different personality traits.

Looking closely into other types of antecedents of OBC participation, one notices that
individual perceptions tend to be integrated in other views as well. For instance, the UGT
shows that perceived benefits or value motivate people to act. The way these benefits or
values are perceived happens at an individual level. Similarly, social capital theory refers
to perceived social value, at an individual level (Mathwick et al., 2008; Faraj and Johnson,
2011). TRA, TAM and TPB also rely on the individual assessment of element such as the
site’s usefulness, or subjective norms (see Casaló et al., 2010). In this sense, they all refer
 66

to the individual member’s personal interpretations and motivations. Therefore, these

theories are based on individuals’ assessment of benefits and value that lead a person to
participate in an OBC. Inherent personality traits and characteristics are not accounted for
in these theories, except for the inclusion of consumers’ attitude toward a certain referent,
for instance the community, as shown in Bagozzi and Dholakia (2006).

Lastly, other studies take a more practical approach to the drivers of OBC participation by
focusing on the OBC management style of the brand that can trigger consumer
responses in the form of enhanced OBC participation. This line of enquiry seems to be
more relevant to online context and it emerged with Stokbürger-Sauer’s (2010) study. He
evidenced that OBC members of a health-related brand tend to be more receptive to offline
than online activities initiated by the brand, indicating that OBC participation is reliant on
offline participation as well.

Analysing over 115,000 posts on ten different OBCs, Homburg et al. (2015) focus on
consumers’ reaction to active brand participation in C2C conversations. Results indicate
that, for conversations that address the consumers’ functional needs and product support,
consumers show diminishing returns to active firm engagement which, at very high levels,
can undermine consumer sentiment. Considering the drivers of commitment, Kim et al.
(2008) found out that online community commitment is a function of support for member
communication, community value, recognition for participation and freedom of
expression. Using mixed methods and quantitative data gathered from marketing
professionals rather than consumers, (Gruner et al., 2014) report that OBC that are
managed by brands in a very or moderately open way (Open OBC and Discerning OBC)
tend to lead to higher success of innovative products.

The following section focuses on the outcomes of OBC participation.

2.5.2. Outcomes of participation in the OBC literature

The task of identifying the outcomes of OBC participation literature is less challenging
than the review of antecedents. When OBC participation is captured as social identification
with the community, its impact on community participation and engagement is well
documented. For example, Dholakia et al., (2004) build on traditional models of social
identity and influence (Bergami and Bagozzi, 2000) to show the impact of OBC
identification on group behaviours. Through the mediating effect of desires and group
 67

intentions, social identification works toward increased participation behaviours in an OBC

(Dholakia et al., 2004).

Furthermore, social identity is argued to be the basis of collective self-esteem and

evaluation of self-worth in online brand community settings (Dholakia et al., 2004), which
works toward increased groups behaviours. Algesheimer et al. (2005) go further into
explaining how identification with the brand community leads to positive consequences,
such as greater community engagement, or motivation to act, and ultimately greater
participation intentions and behaviours toward the community and the brand. Algesheimer
et al. (2005) however also pointed out a negative consequence of community identification
in the form of normative community pressure, which they define as the pressure to interact
and cooperate with the community. This first set of outcomes of OBC participation
assumes that participation is identity based, and as such views community behaviours as an
outcome. This view focuses largely on social behavioural outcomes of OBC
participation, also evidencing benefits for individuals such as self-esteem.

A much larger stream of literature has paid attention to the benefits of OBC participation
for the brand. The discussion started with McAlexander et al. (2002) and is largely
measured through constructs pertaining to the brand relationship paradigm. McAlexander
et al. (2002) pointed out that consumer-centric relationships with different entities within
brand communities have the potential to lead to, and potentially form, a single construct
akin to customer loyalty, through increased attachment to the product and the brand.

OBC contexts also show the various advantages of OBC participation in eliciting
favourable consumer intentions and behaviours with respect to the focal brand. Research
has demonstrated the strong impact of OBC participation (in the form of commitment,
identification or behaviours) on intentional and behavioural brand loyalty (Algesheimer et
al., 2005; Jang et al., 2008; Lin, 2010 – see also Fournier and Lee, 2009 and Muñiz and
O’Guinn, 2001 for theoretical discussions). Similarly, Stokbürger-Sauer (2010) evidences
how social interactions in OBC lead to higher levels of brand identification, which itself
has a positive effect on brand satisfaction, brand loyalty, and brand advocacy
(Stokbürger-Sauer, 2010). The mediating effect of brand identification on the relationship
between OBC participation and brand-related behaviour was also supported by Bagozzi
and Dholakia (2006) and Carlson et al. (2008).
 68

Overall, much emphasis seems to be placed on the (repeat) purchase or adoption of

products by consumers. Thompson and Sinha (2008) approached the issue from a product
adoption perspective supporting that membership in a brand community increases the
likelihood of adopting new products from the preferred brand but also decreases likelihood
of adopting new products from opposing brands (oppositional loyalty). Similarly focusing
on the adoption of new products, Gruner et al. (2014) show that more open communities
foster higher sales volumes and market share of new products. In this context, OBC
participation is also recognised as a driver of members’ participation in open
innovativeness, to generate new product ideas (Füller et al., 2008).

A significant number of OBC studies focus on the concept of brand loyalty to capture the
relational and sales benefits that companies can reap out of OBC participation. Brand
loyalty appears to be the end goal of many OBC studies, whether they are strictly taking a
relational perspective to OBC participation (McAlexander et al., 2002), deriving OBC
participation mainly from social identification (Algesheimer et al., 2005), or using a mix of
social, individual and relationship-based theories (Stokbürger-Sauer, 2010; Matzler et al.,
2011; Marzocchi et al., 2013). Specifically in online contexts, Casaló et al. (2010) show
that participation in an online travel-oriented community impacts attitudinal loyalty toward
the focal brand. Laroche et al. (2012) comment on this view and indicate that behavioural
loyalty is also an outcome of OBC practices in social media environments. Moreover, the
impact of OBC participation on brand loyalty is fully mediated by brand trust in this study
(ibid., 2012). Studies also seem to agree on the mediating effect of other brand-relationship
construct in linking OBC participation to brand loyalty, such as brand commitment
(Carlson et al., 2008), trust and affect (Matzler et al., 2011; Laroche et al., 2012;
Marzocchi et al., 2013) or identification (Stockburger-Sauer, 2010).

To sum up, the importance of generating brand loyalty among OBC members is high on
the agenda of OBC managers, as it appears to be an indicator of the return on investment
of the OBC. As such, it constitutes the ultimate goal of more than one brand managing an
OBC. This assertion was made from the early days of brand community research (Muñiz
and O’Guinn, 2001) and is still valid to date in online contexts (Gruner et al., 2014). It is
also a natural outcome of OBC participation: as members engage in group action, they are
more likely to gain brand-related value, develop a positive attitude toward the brand and
engage in repeat purchase (Jang et al., 2008).
 69

2.5.3. Antecedents and outcomes of consumer engagement

Understanding the role of consumer engagement in a network of nomological relationships

with other constructs has been a stringent issue on the engagement agenda, as posited by
Brodie and colleagues (2011). A few engagement researchers have attempted to tackle this
issue (e.g. van Doorn et al., 2010; Hollebeek, 2011) by identifying potential antecedents,
outcomes, or components of consumer engagement; some of which focus on OBC
contexts. Variations in their approach are nonetheless apparent and dominated by a strong
lack of empirical investigation and verification. This section presents and critically
evaluates studies which concurrently focus on the antecedents and outcomes of consumer
engagement, both in OBC and non OBC-specific contexts, aiming to see how it can
complement, or contradict existing research on the antecedents and outcomes of OBC
participation presented above.

Several antecedents and outcomes of consumer engagement have been proposed so far.
Table 5 gives a summary of the key potential antecedents, outcomes and components of
consumer engagement as conceptualised and empirically tested so far.

Four major conclusions can be drawn from this table. Firstly, it can be noted that consumer
engagement clearly holds links with other concepts deriving from the relationship
marketing paradigm. However, it is unclear whether some of them are antecedents,
outcomes or components of engagement (e.g. satisfaction, trust or commitment).
Furthermore, empirical validation of these relationships is lacking, particularly through
quantitative studies. Lastly, there is very limited treatment of the individual traits
potentially driving consumer engagement, as most of the focus is placed on other relational
and social constructs.
 70

Table 5: Link between engagement and other variables

Constructs References
Antecedents
Trust Van Doorn et al., 2010; Brodie et al., 2011; Hollebeek, 2011a
Commitment Van Doorn et al., 2010; Brodie et al., 2011; Hollebeek, 2011a
Satisfaction Van Doorn et al., 2010; Brodie et al., 2011; Hollebeek, 2011a
Sprott et al., 2009; Hollebeek, 2011a; Vivek et al., 2012;
Involvement
Hollebeek et al., 2014*
Participation Vivek et al., 2012
Interactivity Hollebeek, 2011a
Flow Mollen and Wilson, 2010; Hollebeek, 2011b
Rapport Hollebeek, 2011a
Identity Van Doorn et al., 2010
Consumption goals Van Doorn et al., 2010
Resources Van Doorn et al., 2010
Perceived costs/benefits Van Doorn et al., 2010
Outcomes
Self-brand connection Hollebeek et al., 2014*
Brand usage intent Hollebeek et al., 2014*
Loyalty Bowden, 2009; Hollebeek, 2011a
Consumption
Calder et al., 2013
intentions/ behaviours
Satisfaction Hollebeek, 2011a; Calder et al. 2013; Hollebeek et al., 2014
Experience Hollebeek, 2011a
Co-created value Holleebeek, 2011a
Trust Hollebeek, 2011a
Commitment Hollebeek, 2011; Vivek et al., 2014*
Self-brand memory
Sprott et al., 2009
links
Brand recall Sprott et al., 2009
Brand attention Sprott et al., 2009
Product preference Sprott et al., 2009
WOM/ advocacy Vivek et al., 2012; Wallace et al., 2014; Vivek et al., 2014*
Brand love Wallace et al., 2014
Value perceptions Vivek et al., 2014*
Benevolence
Vivek et al., 2014*
perceptions
Components
Commitment Bowden, 2009
Trust Bowden, 2009
Involvement Bowden, 2009
Experiences Calder et al., 2013*
* Asterisks indicate that the relationship has been quantitatively verified.
 71

In order to shed light into the underlying conceptualisations of these relationships, five
central models from the engagement literature are critically reviewed, namely Bowden’s
(2009), van Doorn et al.’s (2010), Hollebeek’s (2011a), Brodie et al.’s (2013) and Wirtz et
al.’s (2013).

The first study, Bowden (2009), develops one of the first conceptual models of the process
of engagement, which she conceptualises ‘as a psychological process that models the
underlying mechanisms by which customer loyalty forms for new customers of a service
brand as well as the mechanisms by which loyalty may be maintained for repeat purchase
customers of a service brand’ (p65). This view considers engagement as a process rather
than a state, and aims to differentiate new and existing customers in their progress toward
repeat behaviour and loyalty, as shown in Figure 6 below. Under this premise, engagement
is considered to be varying for existing and new customers. An interesting take from
Bowden’s study is the realisation that new customers rely strongly on calculative
commitment and functional evaluations in this process, whereas returning customers’
choice to repurchase is dominated by affective considerations.

Figure 6: Process of engagement in Bowden (2009)

 72

Shortly after Bowden, van Doorn and colleagues (2010) attempt to position customer
engagement behaviour in a causal model, however still conceptual. This model differs
from Bowden’s in three important ways. Firstly, it focuses on engagement behaviours
only, bypassing any attitudinal dimensions of the concept. Secondly, engagement is
viewed as a variable in the model, rather than a process. Thirdly, it is much broader in
scope, as it encompasses customer, firm and context-relevant antecedents and outcomes.
Focusing on the customer dimension, little clarity is given considering the outcomes. The
antecedents, on the other hand, are more precise and include satisfaction,
trust/commitment, identity, consumption goals, resources and perceived costs/benefits. In
terms of the firm-based antecedents, elements such as the brand characteristics, industry or
reputation are proposed. Firm outcomes on the other hand relate to the financial,
reputational or product aspects, among others. Overall, these considerations complement
Bowden’s approach, while however remaining quite broad and conceptual.

Figure 7: Process of engagement in Van Doorn et al. (2010)

 73

In line with these findings, and supporting the predictive power of engagement on loyalty,
Hollebeek (2011a) positions customer engagement in a network of nomological
relationships whereby she proposes that engagement impacts loyalty through the mediating
effects of relationship quality, which is composed of trust, commitment and satisfaction.
These relationships remain however at the conceptual level, and it is acknowledged that
commitment, trust and satisfaction might just as well be antecedents of engagement.

Figure 8: Conceptual model of engagement in Hollebeek (2011a)

These three models, however different, show that consumer engagement is linked with
several consumer-, firm- and context-related factors. Some of the most recurring
relationships that have been explored are those linking engagement with trust,
commitment, involvement and loyalty. It remains however unclear whether trust,
commitment and involvement are antecedents (Van Doorn et al., 2010; Brodie et al., 2011;
Hollebeek, 2011a), outcomes (Hollebeek, 2011a; Calder et al. 2013; Hollebeek et al.,
2014) or components (Bowden, 2009) of engagement.

More recent studies focus on consumer engagement in OBC contexts, and thus provide
increased precision into the interplay between consumer engagement and other constructs
in this context. Adopting a slightly different positioning, Brodie et al. (2011) start by
relating customer engagement to Vargo and Lush’s (2004) service dominant (SD) logic.
Customer engagement is reflective of co-creative experiences in networked service
relationships, replacing other relational terms like participation or involvement. In this
perspective, engagement sits within an iterative service relationship process, where its
consequences may act as antecedents in subsequent sub-processes or cycles.

Brodie et al. (2013) follow-up on this 2011 study by reporting qualitative case study
findings from an online ‘Vibration Training’ community, allowing them to settle on the
direction of the relationships between consumer engagement and other constructs, at least
 74

within the context of OBC. Their exploratory findings determine that engagement is
conducive of loyalty, satisfaction, empowerment, emotional bonds, trust and commitment
(Brodie et al., 2013). Triggers of consumer engagement are not as clearly expressed,
although they are said to derive mainly from information search.

Figure 9: Process of engagement in Brodie et al. (2013)

Lastly, Wirtz et al. (2013)’s model of the antecedents and outcomes of consumer OBC
engagement proposes a holistic approach to the issue. It includes brand-related, social and
functional drivers, moderating factors as well as outcomes of OBC engagement. This
model is the most comprehensive conceptualisation of the mediating role of consumer
engagement in an OBC context.
 75

Figure 10: Conceptual model of engagement in Wirtz et al. (2013)

2.6. Gaps and research questions

The issue inherent to the extant literature on OBC participation is threefold. Firstly, there is
an extremely fragmented theoretical approach to the treatment of OBC participation
(Preece and Maloney-Krishmar, 2005). This fragmentation leaves the reader to wonder
which approach is best, and if a single one can cater for the apparent complexity of the
phenomenon. This silo approach, with several views of OBC participation ranging from
affective to social and behavioural perspectives (e.g. Dholakia et al., 2004; Casaló et al.,
2008; Stockbürger-Sauer, 2010) indicates that more than one dimension of participation
might be needed to capture the full scope of OBC participation. In other words, this
suggests the need for a multi-dimensional approach to OBC participation, which could
encompass emotional, evaluative and behavioural elements (Brodie et al., 2013).
 76

Secondly, there is an oversight of social referents in the way OBC participation is

conceptualised and measured. The sociality of OBC participation transpires through some
of its perspectives, such as the social identification theory (e.g. Algesheimer et al., 2005),
whereby the individual construes his/her identity to be congruent with that of the brand
community. However, adopting a social referent to OBC participation is not systematic
(e.g. Kim et al., 2008), despite the clear social underpinnings permeating the whole OBC
participation literature from its inception (Muniz and O’Guinn, 2001). These
underpinnings are also inherent to the online context (Yadav and Pavlou, 2014).

Thirdly, when a social referent exists, a lack of interactivity is still denoted in the way
OBC participation is captured. Even when participation is envisaged with respect to the
rest of the group, the active engagement of the group in sustaining OBC participation does
not seem like a pre-requisite. The only approach truly acknowledging the interactive nature
of participation is social constructivism (e.g. Schau et al., 2009); yet is it bound by a
worldview that makes it inherently impossible to capture quantitatively.

To summarise, in addition to creating a confused account of OBC participation, studies

collectively denote a lack of integration of the very essence of online consumer behaviour
(Mathwick, 2002). Online consumers interact in rich, complex, social and highly
interactive ways, which are fostered by online platforms (Labrecque et al., 2013). There is
therefore an urgent need to address these three gaps in the treatment of OBC participation,
using an approach much more responsive of the uniqueness of online community
environments (Quinton, 2013).

In this respect, the recent introduction of consumer engagement as a form of OBC

participation offers clear benefits in comparison to other frames of analysis.
Multidimensionality is the first added value of consumer engagement, as it is largely
viewed in the marketing and social sciences as a multidimensional concept. Consumer
engagement encompasses an affective, cognitive and behavioural dimension. This
multidimensionality is even more evident and strongly acknowledged in OBC studies
(Brodie et al., 2013).

Additionally, consumer engagement is inherently social, extending the realm of

relationship marketing to the OBC context. Consumer engagement is always approached
as involving a subject (in marketing, often the consumer) and a partner (the brand, the
 77

community, a piece of media). In this sense, engagement can never happen, and can never
be measured without reference to both a subject and a locus of engagement.

Another asset of consumer engagement is that it is inherently interactive (Vivek et al.,

2012), which is congruent with online contexts. Consumer engagement is characterised by
an active two-way relationship between a subject, and an engagement partner (Brodie et
al., 2011). In this way, consumer engagement is more sensitive to the interactive nature of
online contexts, which shape consumer behaviour broadening the speed, reach and amount
of information shared (Brodie et al., 2013; Hollebeek et al., 2014). Consumer engagement
as a concept clearly offers strong advantages compared to existing approaches to OBC
participation. It shows potential to address all three key shortcomings of the literature.

The strength of consumer engagement lies thus in its holistic, social and interactive
perspective to the treatment of OBC participation. Consumer engagement seems to cover
under the same umbrella term all critical dimensions of OBC participation (Mollen and
Wilson, 2010): it involves behavioural manifestations and active participation in
community activities and interactions (see Dholakia et al., 2004; Wiertz and de Ruyter,
2007; Garnefeld et al., 2012), it has an affective dimension which has previously been
accounted through community commitment or attachment (see Jang et al., 2008; Kim et
al., 2008), and it also denotes a cognitive component in line with previous studies focusing
on attention and cognitive involvement in the OBC context. The iterative and experience-
laden nature of brand community practices is also accounted for (Schau et al., 2009), as
well as the social aspect of in-group actions (Bagozzi and Dholakia, 2006).

Moreover, consumer engagement appears to be highly congruent with online settings and
virtual engagement platforms (Breidbach et al., 2014), and brand experiences (Hollebeek
et al., 2014). The adequacy of consumer engagement to the OBC context is further
discussed in chapter 3 of this study.

Current research into consumer engagement in OBC settings is however still lacking,
particularly on the empirical front. Consumer engagement is a new concept, and most
studies addressing it are still conceptual or exploratory (e.g. Brodie et al., 2013). In this
context, little is known about the true applicability of the concept of consumer engagement
as a way to approach OBC participation. What is the exact nature of consumer
engagement? What does it multidimensionality really mean? More specifically, how can
we measure consumer engagement in a way to capture all the facets of OBC participation
 78

accurately? All these questions remain largely unanswered to date in the marketing and
OBC literature (Brodie et al., 2013) and calls for further operationalisation of the concept
of consumer engagement in online social environments have been made (Hollebeek et al.,
2014; Schivinski et al., forthcoming).

On the basis of this lack of conceptual clarity and empirical validation of a measure of
consumer engagement in OBC, the first research question of this thesis is:

RQ 1: How to measure consumer engagement in OBC?

In addition to measuring engagement, it seems important to identify its antecedents and

outcomes (Hollebeek et al., 2014). The review of the literature on the antecedents of OBC
participation revealed is a lack of consensus about the major drivers of consumer
participation in firm-hosted OBC (Casaló et al., 2010). In this endeavour, is has not been
uncommon for researchers to mix different theoretical frameworks and concepts. Casaló et
al. (2010), for instance, use elements of the TAM and social identity theory. Similarily,
Algesheimer et al. (2005) combine constructs from relationship marketing and social
identity theory, while Jang et al. (2008) add information technology elements to the mix.
Bagozzi and Dholakia (2006) on the other hand, combine the social identification theory
with the TPB to predict OBC participation. Studies investigation OBC participation, its
antecedents and outcomes clearly tend to mix and match theories and approaches based on
their specific OBC contexts and research agenda.

The literature review showed that social and motivational drivers for participation have
received an extended treatment (Campbell et al., 2014). Social identification (with the
community or with the brand) is understood as a clear driver of OBC participation, as well
as other social norms and attitudes (Bagozzi and Dholakia, 2006). The UGT also gave
ground to a broad stream of research on member’s motivation to participate based on
expected benefits and value, whether of an informational, social or personal nature
(Dholakia et al., 2004). The relationship marketing literature has also been adopted in
multiple instances, evidencing the role of elements such as brand relationship quality
(Bagozzi and Dholakia, 2006), brand trust, brand passion (Füller et al., 2008), or brand
identification (Stokbürger-Sauer, 2010) in the mix to reach higher OBC participation.

Individual traits and characteristics of consumers seem to have received a weaker and more
incomplete treatment than any other types of antecedents. A fragmented stream of research
however recognises the crucial importance of individual factors in influencing OBC
 79

participation, and their ability to be used as members’ segmentation or classification

criteria (see De Valck et al., 2009). Fournier and Lee (2009) provide a purely conceptual
approach to these individual traits, De Valck and colleagues (2009) base their findings on
qualitative data, and Chang et al. (2013) are the only ones using quantitative data in this
endeavour. Focusing on individual traits and specificities therefore seems to be a
particularly promising research direction, which has been rather overlooked (Wiertz and de
Ruyter, 2007) and bears high potential for managerial and theoretical contribution. Indeed,
recent studies call for further investigation of these types of participation drivers (Chang et
al., 2013).

Additionally, investigating individual traits and characteristics is relevant to the

understanding of consumer engagement (Wirtz et al., 2013). Firstly, because individual
traits such as personality tend to show a greater consistency through time than other
motivational or social factors, they should therefore be strongly related to repeat and long-
term engagement manifestations (Lastovicka and Joachimsthaler, 1988). Secondly,
conceptual studies in consumer engagement have recently introduced the need to study the
role of consumer-specific factors on their engagement (e.g. Wirtz et al., 2013), however
without giving clear indications on what these indicators might be.

Based on the current state of both OBC and consumer engagement literature, it therefore
appears to be a priority to further explicate the role of consumer-specific traits in driving
consumer engagement in OBC. The second research question of this thesis is thus
formulated:

RQ2: What are the individual traits and predispositions that drive consumer engagement
in OBCs?

The treatment of the outcomes of OBC participation seems less problematic and highly
centred upon the brand benefits of OBC participation. Brand relationship building seems to
be one of the most frequently researched outcomes of OBC participation (e.g. Matzler et
al., 2011; Marzocchi et al., 2013). This particular interest in understanding the brand-
relationship outcome of OBC participation makes sense both from a theory and practical
point of view. Theoretically, OBC research is anchored in the brand relationship paradigm,
whereby consumers and companies create and develop bonds (McAlexander et al., 2002).
The literature shows strong focus of OBC scholars in understanding relational antecedents
and outcomes of OBC participation (e.g. Algesheimer et al., 2005; Casaló et al., 2010).
 80

Although there is a disagreement among scholars as to whether consumer-brand

relationships should precede (Algesheimer et al., 2005), derive from (Jang et al., 2008) or
be an iterative circle with OBC participation, a significant number of studies have shown
that OBC participation leads to increased brand trust (Laroche et al., 2012) commitment
(Kim et al., 2008) and, most importantly, loyalty (Casaló et al., 2010; Stokbürger-Sauer,
2010; Matzler et al., 2011; Marzocchi et al., 2013). Studies also seem to agree on the
mediating effect of other brand-relationship construct in linking OBC participation to
brand loyalty, such as brand commitment (Carlson et al., 2008), trust and affect (Matzler et
al., 2011; Laroche et al., 2012; Marzocchi et al., 2013) or identification (Stockburger-
Sauer, 2010).

These important consequences of OBC participation are yet to be confirmed with reference
to the concept of consumer engagement. Indeed, the literature incorporating consumer
engagement in OBC studies still largely focuses on the conceptual relationships between
consumer engagement and its relational outcomes in an OBC context (Brodie et al., 2011).
Brand relationship benefits from consumer engagement in OBC are likely to exist
(Hollebeek, 2011); however the link between consumer engagement in OBC and brand-
relationship outcomes such as trust, commitment and loyalty has been largely overlooked.

Confirming the explanatory potential of the consumer engagement on consumer-brand

relationships in the context of OBC is thus an urgent requirement, which leads to the
formulation of the third research question driving this thesis:

RQ3: How does consumer engagement in OBC contribute to brand relationship

building?

2.7. Summary

This chapter reviews the literature on OBC participation and consumer engagement. The
aim is to uncover some of the most stringent gaps existing in the OBC literature, under the
guidance of the research aims and objectives presented in chapter 1. The chapter identifies
several key issues pertaining to the way OBC participation has been approached so far,
which make it largely irresponsive of the online environment in which they operate. The
review shows that OBC participation would benefit from a more social, interactive and
 81

multidimensional approach and measurement. On this ground, the chapter identifies

engagement as a valid route for conceptualising OBC participation. Engagement is a
promising concept because it is inherently social, interactive and multidimensional.
Research on engagement however remains conceptual and at best exploratory. A major gap
is identified with the need to provide a better conceptualisation of consumer engagement in
OBC, and develop a measure to capture this construct.

This chapter also identifies the need to focus on the individual traits and characteristics of
consumer in an effort to explain the formation of consumer engagement in OBC. Factors
such as consumers’ attitude toward participation, OIP and product involvement are
potential antecedents of consumer engagement. Lastly, a need for empirical validation of
the impact of consumer engagement on brand relationships in the context of OBCs is
identified.
 82

Chapter 3: Conceptual approach

3.1. Introduction

The introduction of this thesis determined two objectives of the present research: one
focusing on the capture and measurement of consumer engagement as an approach to OBC
participation, the second addressing the role of consumer engagement in a nomological
network of relationships with other constructs in the OBC context. The aim of this
chapter is thus to adapt the concept of consumer engagement to an OBC context, and
to build a conceptual model explaining the drivers and outcomes of consumer
engagement in OBC.

The literature review has highlighted three main gaps in the current understanding of OBC
participation and proposed consumer engagement as a way to advance this understanding.
The innovative concept of consumer engagement emerged as a sound holistic construct to
consolidate the currently fragmented state of research in OBC participation (Brodie et al.,
2013). From here on, consumer engagement is adopted as the denomination of the
phenomenon under study, in the place of participation. Using consumer engagement as a
way to conceptualise OBC participation is an effort to bring clarity and exhaustiveness to
the field of OBC participation.

This study is however sympathetic to the tradition of OBC studies, which combine
different theoretical strands (Preece and Malhoney-Krishmar, 2005). In identifying the
gaps regarding the antecedents and outcomes of consumer engagement in OBC, one notes
that there is no dominant theoretical approach: a variety of factors may affect and result
from engagement. Rather than settling on one theory, since there isn’t a preferred one, the
study integrates multiple theoretical strands in order to extend the current approach to
engagement in OBC. The study focuses on antecedents of consumer engagement that have
not been confirmed in empirical quantitative studies on engagement but transpired as
conceptually relevant: individual traits and characteristics (Wirtz et al., 2013). More
specifically, the OBC literature directs the choice of OIP (Wiertz and de Ruyter, 2007),
attitude toward OBC participation (Bagozzi and Dholakia, 2006) and product involvement
 83

(Sprott et al., 2009; Hollebeek et al., 2014) as promising concepts. Regarding the outcomes
of engagement, OBC studies have granted much focus on brand relationship concepts such
as trust, commitment and loyalty, and it seems relevant to test this assumption under a
consumer engagement approach (Hollebeek, 2011b). This broad integrative view allows
the confirmation of results from OBC studies, and the empirical validation of a lot of
exploratory and conceptual work in the field of engagement. In this sense, the study seeks
confirmation of existing work, but does not preclude discovery.

In this chapter, a framework of consumer engagement in OBC is proposed. This

framework builds on the existing consumer engagement literature and conceptual
frameworks, also reaching out to other streams of literature in the social sciences. This
framework presents the conceptual premises against which consumer engagement is
expected to operate, or be captured, in an OBC context. Secondly, a conceptual model is
proposed, placing consumer engagement, as conceptualised in the above-mentioned
framework, at the centre stage of OBC functioning. The individual factors driving online
consumer engagement are identified, as well as its resulting relational variables. The OBC
literature and theories are used to build this model, more particularly focusing on the brand
relationship paradigm, coupled with individual traits, preferences and attitudes. Thirdly,
the overall conceptual model is translated and operationalised into a set of hypotheses.

The last section of this chapter focuses on proposing a validation of the study across two
linguistic samples. The current state of OBC and consumer engagement research is
scrutinised and evidence of little contextual validation procedures is highlighted,
particularly when it comes to different languages.

3.2. Consumer engagement framework

It is clear from the literature review that consumer engagement is a context-specific

concept adaptable to a lot of different consumption-related marketing environments.
Depending on the context in which it appears, engagement has different referents (Brodie
et al., 2011). This section aims to define and characterise consumer engagement in the
OBC context.
 84

3.2.1. Defining consumer engagement in OBC

The literature review provided an account of the existing conceptualisations and definitions
of consumer engagement, highlighting the different approaches that exist. This study
focuses on consumer engagement in OBC and defines it as:

‘The state that reflects consumer’s positive individual dispositions toward

engagement partners. In an OBC context, these partners are the OBC, representing
the other consumers in the OBC, and the focal brand. It is expressed through
varying levels of affective, cognitive, and behavioural manifestations that go beyond
exchange situations’.

This definition closely follows Brodie et al. (2011), Brodie et al. (2013) as well as
Hollebeek’s (2011, 2014) conceptualisations of consumer engagement. It views consumer
engagement as a psychological concept expressed through multi-dimensional
manifestations, in line with most studies in marketing and broader social sciences
literature. Additionally, engagement involves a subject and a partner. In the OBC context,
the subject of engagement is the individual consumer, and its engagement partners are the
brand on the one hand, as personified by the OBC manager(s) and the OBC, as personified
by the other consumers in the OBC.

This definition also implies the enduring aspect of engagement. Consumer engagement
goes beyond one-off (or repeat) purchase situations. It is associated with an extended
vision of consumer-brand relationships that is not only concerned with exchange situations
but deep, on-going relationships (Vivek et al., 2012). This has important implications on
the way it needs to be conceptualised and measured, and echoes the need to recognise the
long-term nature of connections with brands online (Morgan-Thomas and Veloutsou,
2013).

Additionally, this definition focuses on engagement as a positively-valenced concept. The

study acknowledges the existence of positive and negative forms of engagement in OBC
(Hollebeek and Chen, 2014). However, most research on engagement in OBC has thus far
initiated conceptual and exploratory work on the positive side of the phenomenon, calling
for further research on positive engagement (Brodie et al., 2013; Wirtz et al., 2013).
Although work is needed on both sides of the engagement coin, this study first attempts to
 85

bring further research on positive engagement, which also allows focusing on positive
outcomes (Hollebeek and Chen, 2014).

Moreover, in line with the first stated objective of this research, it adopts a definition of
engagement as a state, which allows it to have varying levels, or intensity, and thus the
potential to be operationalised, and measured (see Vivek et al., 2014; Hollebeek et al.,
2014). It also opens avenues for treating engagement as part of a nomological network of
relationships between constructs, which is the second objective of this study.

The meaning of the different elements of this definition is further clarified below, showing
how this definition is relevant and suited to the OBC context, how it advances the meaning
of OBC participation. Since consumer engagement is a context-specific concept, and that
this study chooses to focuses on the context of OBC, it is important to delineate the
contextual boundaries that this focus implies. More specifically, the choice of engagement
subject and partners is explained in line with the OBC context. Subsequently, the
dimensionality and sub-dimensionality of engagement are conceptually clarified – these
are however considered to be static and not to vary according to the context of
investigation, since they derive from different literature streams and thus contexts. These
clarifications allow reaching a conceptual framework of consumer engagement in OBC.

3.2.2. The subject of engagement in OBC

The notion of the subject of engagement represents a key aspect of conceptualisation. As

clarified previously, engagement involves both a subject (the engaged entity) and a partner
(the active focus of engagement). Clearly determining and defining the subject of
engagement is an important boundary assumption of this study, as it is understood from the
literature review that it varies according to the context of investigation. There is a broad
agreement that the unit of analysis in marketing should be the individual customer (e.g.
Bowden, 2009; Verhoef et al., 2010; Hollebeek, 2011), or consumer (e.g. Brodie et al.,
2013; Calder et al., 2013; Wallace et al., 2014), regardless of the context of investigation.
Although the initiator of engagement can be the company, through specific offerings or
activities, the person whose engagement matters to marketing researchers and practitioners
alike is, understandably, the individual customer or consumer (Vivek et al., 2012). It seems
however that ‘customer’ and ‘consumer’ are used interchangeably in the literature, as
evidenced in Table 4 in the literature review. Clarification of the engagement subject is
therefore needed. Since this study focuses on a specific consumption-related environment,
 86

the OBC, a clear understanding of what it meant by ‘consumer’, or ‘customer’ is a

prerequisite.

This study focuses on consumers as the subject of engagement and level of analysis as it
appears to be the best terms to qualify members of an OBC. Consumers are considered
here in their capacity of OBC members, or users, and can be defined as the persons who
consume the community content, contribute to it and participate to the community to some
extent (Kaplan and Haenlein, 2010). Traditionally, a consumer is defined as ‘the user of a
product, service or other form of offering’, whereas a customer is ‘the person who
purchases and pays for a product, service, or other form of offering from a company or
organisation’ (Baines et al., 2011, pp. 729-30). In OBC embedded on social media
however, being either a consumer or customer of the brand is not a prerequisite to OBC
membership or participation (Ruiz-Mafé and Blas, 2006).

This study therefore views brand usage in a broader sense than pure product or service
consumption and rather sees it as going beyond purchase situations (Vivek et al., 2012).
Being member of an OBC focused on a brand is therefore akin to consuming the product,
in a non-purchase sense. This is what is intended in the definition of consumer engagement
when referring to situations ‘beyond purchase’ (ibid., 2012). On this basis, this thesis refers
to consumers as the subject of engagement and level of analysis, adopting a view of
consumers as any existing member of an OBC, irrespective of past, previous or current
brand ownership.

This conceptualisation is congruent with both the OBC and consumer engagement
literature. Individual community members can be the subject of engagement (Algesheimer
et al., 2005) and can feel engaged toward any particular organisational activity or event
(Vivek et al., 2014). Moreover, both existing purchasing consumers and non-existing
purchasing consumers of a brand are capable of engagement (Bowden, 2009). Indeed, one
may or may not have brand ownership in order to visit and partake in an OBC (Zhou et al.,
2013), and ‘administrative’ association with a brand is not considered as a prerequisite for
OBC participation (Owersloot and Oderkerken-Schröder, 2008).

Based on this understanding of what being a consumer in an OBC context means, this
study refers to ‘consumer’ engagement throughout, rather than ‘customer’ engagement.
 87

3.2.3. The partners of engagement in Online Brand Community

Consumer engagement is interactive by nature (Vivek et al., 2012). Hence, engagement

can only happen if there is a relationship partner to interact with, and use as an object of
engagement. Similarly to specifying the subject of engagement, boundary assumptions of
this study need to be determined in terms of the engagement partners.

The literature review showed that in the context of OBCs, the relevant and key engagement
partners to consider are the brand and the community itself (Algesheimer et al., 2005;
Wirtz et al., 2013). In line with its context of investigation, this research is centred on
consumer engagement with two engagement partners: the OBC, as personified with other
OBC members on the one hand, and the brand itself on the other hand. Consumer-to-
consumer relationships as well as consumer-brand relationships are core constituents of the
OBC functioning, as evidenced through the strong stream of literature in OBC focused on
relationship marketing (McAlexander et al., 2002).

The reasons for focusing on OBC and brand engagement as engagement partners in an
OBC context stem from both the consumer engagement and OBC literature. In the OBC
literature, it clearly transpired that consumer develop relationships with the focal brand,
which are indicative of their participation in the community (e.g. Kim et al., 2008). The
relationships that they form with other members of the community are equally indicative of
their commitment to the community, and researchers have thus coined ‘brand community
engagement’ as a construct of interest when tapping into OBC participation (Algesheimer
et al., 2005).

This duality is also apparent in the consumer engagement literature. Although most of the
consumer engagement studies focus on one engagement partner at a time, studies focusing
on engagement in an OBC context have clearly acknowledged the need to consider
multiple engagement partners: the community of other members, and the brand (Brodie et
al., 2013; Wirtz et al., 2013). It is even argued that the former can impact the latter (Wirtz
et al., 2013), an assertion which will be developed later on in this chapter.

The remainder of this study therefore refers to consumer engagement in an OBC context
as an overarching concept, which always has the individual consumer as a subject, but can
have two different engagement partners, the OBC and the brand. This thesis refers to
‘online brand community (OBC) engagement’ and ‘online brand engagement’ to
differentiate when consumer engagement is directed at the OBC or at the brand.
 88

Depending on the scope of analysis, this study either refers to of the concept consumer
engagement as whole, or focus on one or two of its partner-specific constructs.

3.2.4. The dimensions and sub-dimensions of consumer engagement

This study aligns with the multi-dimensional, cognitive-affective-behavioural, view of

consumer engagement (Brodie et al., 2011; Hollebeek, 2011). Not only is it the most
accepted view in the extant literature on consumer engagement (Brodie et al. 2011), but it
is also one of the factors that distinguish engagement from other unidimensional
approaches to OBC participation, therefore evidencing its strength. A clarification of the
exact meaning of these dimensions in an OBC context is however required, as existing
marketing literature has been providing a rather confused account of the existence and
meaning of different engagement dimensions.

The development of the dimensions of engagement proposed hereafter builds on relevant

literature in 1/consumer engagement 2/ (online) brand community and 3/ other fields of the
social sciences such as employee engagement. Each dimension is first approached and
defined conceptually, followed by its sub-dimensions.

Affective engagement
Affect represents the first dimension of engagement. Marketing research is rich in
references to the concept of affect, as well as emotions and feeling, and the role of these
constructs in consumer behaviour is well referenced (Bagozzi et al., 1999). Among
established affective processes, emotions, feelings and moods can and need to be
distinguished. Emotions and feelings are both responses to specific stimuli, however
emotions are usually intense and enduring, whereas feelings are more transient (Agarwal
and Malhotra, 2005). An emotion is a valenced mental state of readiness, having a specific
referent and strong intensity, and it may result in specific actions. Moods on the other hand
are lower in intensity, and usually not associated with a stimulus object (Bagozzi et al.,
1999). Based on these considerations, affect is defined in the context of this study as ‘the
summative and enduring level of positive emotions experienced by a consumer with respect
to his/her engagement partner’. Affect is composed of different and complementary
emotional stimuli, relevant to the subject, object and context of the affective occurrence.

Engagement is generally perceived as a diffuse, long-lasting state rather than a trait

(Schaufeli et al., 2006), hence, it would be inaccurate to dimensionalise it in terms of
transient phenomena. Rather than time-bound emotions, affect refers to a summative
 89

mechanism of emotion processing (Sojka and Giese, 1997), and it therefore reflects the
enduring aspect of engagement. When considering affective dimensions of engagement, it
is therefore important to encompass different relevant emotions that reflect positive
engagement and consider these emotions from a long-lasting and summative perspective.

The first aspect of the affective dimension is enthusiasm. Enthusiasm represents an

individual’s strong level of excitement and interest regarding the engagement partner
(Vivek et al., 2014), and several researchers have found enthusiasm to represent a positive
affective engagement state in both employee and consumer engagement literature. In an
organisational context, engagement refers to the employee’s sense of significance,
enthusiasm, inspiration and pride (Schaufeli and Baker, 2004), which indicates that he/she
feels enthusiastic and passionate about his/her job. This is consistent with the dimensions
of vigour (Schaufeli et al., 2002; Patterson et al., 2006) and activation (Hollebeek et al.,
2014), which are representative of the motivational nature of engagement (Higgins and
Scholer, 2009; Pham and Avnet, 2009; Brodie et al., 2011).

Enthusiasm also differentiates engagement from other similar marketing constructs, such
as satisfaction (Macey and Schneider, 2008). Despite the fact that satisfaction also
constitutes a summative affective response, it is an evaluative process based on past
consumption experiences and the performance of the offering (Johnson and Fornell, 1991).
Enthusiasm is characterised by a strong feeling of excitement (Bloch, 1986), which is an
enduring and active state, and does not encompass performance evaluations.

Enjoyment is the second aspect of the affective dimension of engagement. It refers to an

individuals’ intrinsic and effortless pleasure in being in touch with the focus of his/her
engagement. In the employee engagement literature, researchers refer to employee’s
absorption as the fact of being carried away, immersed and happy when engaged with a
task (Schaufeli et al., 2002). Engagement is therefore associated with a pleasurable state
(Mollen and Wilson, 2010). Calder et al. (2013) define intrinsic enjoyment as ‘the
experience of having a break and forgetting about everything else, of being transported
into a better mood or state of mind’ (p.12) and consider it an integral part of consumer
engagement, based on a series of qualitative interviews with newspapers readers.
Engagement is thereby associated with a form of hedonic pleasurable state, with underlies
the motivation to remain engaged.
 90

Enjoyment and pleasure are well-documented affective outcomes of brand use and
consumption, as demonstrated in the brand affect concept and scale (Chaudhuri and
Holbrook, 2001). Brands are thus recognised to have the potential to generate pleasurable
states for consumers, and this can be expected from other engagement objects as well. The
construct of brand affect as defined by Chaudhuri and Holbrook (2001) is however not
applicable to engagement as such, and the scale cannot be adopted because there is a clear
conceptual and operational boundary to brand affect which limits it to consumptive
situations, and it is a fundamental characteristic of engagement to be going beyond
purchase situations.

The following table summarises this study’s conceptualisation of the affective dimension
of consumer engagement, and its two sub-dimensions: enjoyment and enthusiasm. All
definitions are applicable to both engagement partners this study investigates: the OBC and
the brand. Definitions can therefore be duplicated by simply substituting the engagement
partners with the appropriate referent.

Table 6: Affective dimension and sub-dimensions

Affective dimension of consumer engagement

Definition References
The summative and enduring level of positive Sojka and Giese, 1997; Bagozzi, et al.,
emotions experienced by a consumer with 1999; Chaudhuri and Holbrook, 2001;
respect to his/her engagement partner. Brodie et al., 2011; Hollebeek, 2011a;
Hollebeek, 2011b.
Sub-
Definition References
dimension
Enthusiasm A consumer’s intrinsic level Schaufeli et al., 2002; Schaufeli and
of excitement and interest Baker, 2004; Patterson et al., 2006; Vivek,
regarding the engagement 2009; Hollebeek, 2011a.
partner.
Enjoyment A consumer’s pleasure and Chaudhuri and Holbrook, 2001; Schaufeli
happiness derived from et al., 2002; Mollen and Wilson, 2010;
interaction with the Calder et al., 2013.
engagement partner.

Cognitive engagement
Cognition is a well-recognised dimension of engagement in different streams of literature,
and it is often combined with the emotional dimension of engagement. To illustrate, in
organisational behaviour, Saks (2006, p. 600) defines employee engagement as ‘‘the
amount of cognitive, emotional and physical resources an individual is prepared to devote
in the performance of one’s work roles.’’ Consistent with this emphasis on the
 91

psychological elements, engagement is ‘a positive, fulfilling, work-related state of mind’

(Schaufeli et al., 2002, p. 74), suggesting that employee engagement is a persistent and
pervasive affective–cognitive state (Schaufeli and Bakker, 2004). More specifically,
organisational behaviour research has emphasised absorption (Schaufeli et al., 2002) and
attention (Rothbard, 2001) as cognitive aspects of employee engagement.

In the consumer engagement literature, several researchers refer to the cognitive aspect of
the concept (e.g. Patterson et al., 2006; Mollen and Wilson, 2009; Sprott et al., 2009;
Brodie et al., 2011; Hollebeek, 2011a; Hollebeek, 2011b; So et al., 2013; Vivek et al.,
2012; Brodie et al., 2013). Specific examples illustrating the cognitive dimension of
customer engagement are active, sustained and even complex cognitive processing (Mollen
and Wilson, 2010); immersion (Hollebeek, 2011), attention (Higgins and Scholer, 2009)
and absorption (Patterson et al., 2006). Hollebeek (2011b) further exemplifies occurrence
of cognitive engagement through individual’s level of concentration and/or engrossment
with a brand, which, again, is happening beyond and irrespective of any exchange (Vivek
et al., 2012). Based on these grounds, this study defines cognitive engagement as ‘a set of
enduring and active mental states that a customer experiences with respect to his/her
engagement partner’, and includes attention and absorption as sub-dimensions.

Attention is the first aspect of the cognitive dimension of engagement. In the context of
work, attention has been found to be an important dimension of employee engagement
whereby the employee focuses and is mentally preoccupied with work (Rothbard, 2001).
Attention represents an immaterial limited resource that individuals can allocate in a
number of ways. Higgins and Scholer (2009) exemplify this point in the consumer
engagement realm in the following way: the more a spectator is engaged in a movie, the
less likely he is to pay attention to noises in the audience, and the more likely he will be to
follow the central point of the plot. Engagement is about sustained attention, which is a
limited resource. A consumer who is engaged with an object will also be more attracted by
information about it (So et al., 2013), and therefore more prone to further engagement
behaviours and knowledge co-creation, such as active information search and sharing
(Brodie et al., 2013). Attention is therefore considered an important aspect of cognitive
engagement and it is defined as the ‘cognitive availability and amount of time spent
actively thinking about and being attentive to the focus of engagement’.

Absorption represents the second sub-dimension of cognitive engagement. In

organisational behaviour, absorption is recognised as an indicator of employee engagement
 92

(Rothbard, 2001). Absorption with one’s work is characterised by being fully concentrated
and engrossed, whereby time passes quickly and one has difficulties detaching oneself
from work (Schaufeli and Bakker, 2004). Absorption refers to a distortion of time, loss of
self-consciousness and effortless concentration with respect to the object of engagement.
In contrast to attention, which represents the amount of cognitive availability of an
individual, absorption means being engrossed in a role and refers to the intensity of one’s
cognitive focus on a role.

Patterson et al. (2006), draw on this organisational behaviour research by Schaufeli and
colleagues to define customer engagement in terms of the cognitive ‘absorption,’
emotional ‘dedication’ and behavioural ‘vigor’ and ‘interaction’ dimensions. Absorption,
in particular, is viewed as the level of customer concentration on and engrossment with a
focal engagement object, such as a brand/organisation. This is congruent with Hollebeek’s
(2011) aspect of immersion with, and concentration on a brand. In line with these
definitions, this study views absorption as ‘the level of customer concentration and
immersion with an engagement partner’.

Table 7: Cognitive dimension and sub-dimensions

Cognitive dimension of consumer engagement

Definition References
A set of enduring and active mental states Rothbard 2001 ; Schaufeli et al. 2002 ;
that a consumer experiences with respect Patterson et al., 2006; Higgins and Scholer,
to his/her engagement partner. 2009; Mollen and Wilson, 2009; Sprott et
al., 2009; Brodie et al., 2011; Hollebeek,
2011a and 2011b; Vivek et al., 2012; Brodie
et al., 2013.
Sub-
Definition References
dimension
Attention The cognitive availability Rothbard, 2001; Higgins and Scholer, 2009;
and amount of time spent So et al., 2013.
actively thinking about
and being attentive to the
engagement partner.
Absorption The level of consumer’s Schaufeli and Bakker, 2004; Patterson et al.,
concentration and 2006; Hollebeek, 2011; Hollebeek, 2011.
immersion with an
engagement partner.

Behavioural engagement
Extant studies differentiate behavioural aspects of engagement from the affective and
cognitive aspects. Indeed, behavioural consumer engagement has been the only dimension
 93

of engagement in focus in many studies (e.g. Sawhney et al., 2005; Libai et al., 2010; van
Doorn et al., 2010; Verhoef et al., 2010; Gummerus et al., 2012). This impetus for research
on behavioural engagement might be in part linked to the call of the Marketing Science
Institute identifying customer engagement as a research priority, and defining it as
‘customers’ behavioural manifestation toward a brand or firm, beyond purchase, which
results from motivational drivers […]’ (MSI, 2010, p.4).

Despite this call for a better understanding of engagement behaviours, marketing research
is still hesitant when it comes to delineating what represents consumer engagement
behaviours, or how to encapsulate them in a framework. Examples such as complaining,
participating in events or giving suggestions to other customers or staff members can be
found as manifestations of customer engagement behaviours (van Doorn et al., 2010). In
online settings, actions like blogging, giving ratings or spreading word-of-mouth have also
been identified (Verhoef et al., 2010). Online engagement behaviours can be narrowed
further down to social networking sites by looking at metrics such as the number of ‘likes’,
comments, posts, group visits or number of interactions with an in-group application
(Gummerus et al., 2012).

Engagement being a context-dependent and motivational construct, engagement

behaviours depend greatly on the context and the ways they can be enacted in a particular
setting (van Doorn et al., 2010). As this research focuses on OBC settings, relevant
frameworks pertaining to OBC participation are used to approach behavioural engagement.
The ways individuals behave in an OBC has been approached in multiple instances and
several categorisations have been proposed to understand customer participation and
engagement. These categorisations reveal the motivational aspect of engagement
(Hollebeek, 2011), as they reflect the benefits, value, or goals the consumers seek to attain
through their actions (Mollen and Wilson, 2010; Verhoef et al., 2010).

Based on netnographic qualitative data Brodie et al (2013) identify 5 sub-processes of

consumer engagement in virtual brand communities, namely: sharing, co-developing,
socializing, advocating and learning. Similarly, Hennig-Thurau et al. (2004) report eight
specific factors of online community contribution (i.e. engagement): venting negative
feelings, concern for other consumers, self-enhancement, advice-seeking, social benefits,
economic benefits, platform assistance and helping the company. In their seminal study on
brand communities (not specifically online), Schau, et al. (2009), using the term
‘engagement’ or a derivative 77 times, propose a framework of community practices,
 94

which can be combined into four categories: social networking, community engagement,
brand use and impression management. This behavioural aspect of engagement is also
found in the idea of interaction, involving the sharing and exchanging of ideas, thoughts
and feelings about experiences with the brand (So et al., 2013).

The combined analysis of these different frameworks gives ground to the development of
what constitutes behavioural engagement and what are its sub-dimensions. In line with the
MSI (2010) and other customer engagement behaviour researchers (Van Doorn et al.,
2010; Verhoef et al., 2010), we define the behavioural dimensions of consumer
engagement as ‘the behavioural manifestations toward an engagement partner, beyond
purchase, which results from motivational drivers’.

An interesting framework to understand customer engagement behaviour is that of van

Doorn et al (2010), who propose that engagement behaviours are characterised by their (1)
valence; (2) form or modality; (3) scope; (4) nature of its impact and (5) customer goals.
Although each of these dimensions is expanded upon hereunder, the key categorisation
criterion used here is the customer’s goal. Although the three sub-dimensions of
behavioural engagement hereunder all materialise through interactive actions and thus
have a social dimension, their fundamental differentiating characteristic is what motivates
them, i.e. the customer goals (van Doorn et al., 2010). In this respect, their purposes are
clearly different. The underlying purpose of sharing is the exchange of resources; the aim
of endorsing is to sanction, exhibit approval, give weight to the engagement focus, whereas
the purpose of learning is to seek help or information. These three sub-dimensions are
detailed hereunder.

Sharing resources is a key determinant of engagement behaviours. Sharing can be driven

by individuals seeking to provide resources. Providing resources can take the form of
exchanging experiences, helping other customers (Hennig-Thurrau et al., 2004; Mathwick
et al., 2008), providing recommendations and ideas to the company (Sawhney et al., 2005;
Brodie et al., 2011), or making suggestions to improve the brand experience and use
(Schau et al., 2009). Customer engagement therefore relies heavily on the exchange of
experiences (Vivek et al., 2012), as research on co-creation indicates (Vargo and Lusch,
2004). The act of sharing is defined in this study as a collaborative and interactive
exchange, driven by the motivation to provide resources. Social media environments and
online brand communities are both contexts particularly prone to the development of
sharing behaviours, as they are based on usage, which is in large part driven by content,
 95

and people feel a need to engage in altruistic behaviour (Breitshol et al., 2015). In social
media contexts, sharing can manifest itself through shares, comments, posts, tweets,
replies, or even direct messages.

Learning is an important aspect of behavioural engagement, which pertains to consumers

seeking help, advice or ideas from their engagement object. In this sense, it constitutes the
counterpart of sharing, which is seen here as an act of giving. The search of resources
represents the other side of the coin, whereby consumers seek help, ideas, resources and
information from the company or other customers (Hennig-Thurrau, 2004). Learning is an
important part of consumer engagement (Brodie et al., 2013), as shown by the increased
focus on content strategies. Through brand use practices (Schau et al., 2009), consumers
can learn how to improve their experience with the brand. Learning can also refer to
seeking help when one faces an issue, or is dissatisfied in their brand use. By searching to
improve their experience, learn more, or fix issues, consumers show engagement, as a
disengaged customer would in contrast defect from brand use (Lee et al., 2009). In social
media contexts, like sharing, learning can manifest itself through comments, posts, tweets,
replies, or even direct messages.

Endorsing is another aspect of positive consumer engagement, which has been found in
both conceptual and empirical work on the topic. Endorsing can have a smaller or larger
scope, and it refers to the act of showing support to the focus of engagement by
sanctioning their actions or ideas. The scope of endorsement can be small or large, whether
the aim of the endorser is to show approval only to their engagement focus or to act as a
referent of the engagement focus toward the outside (Van Doorn et al., 2010). In an OBC
setting, for instance, members can sanction group activity, content or ideas through the
Facebook ‘like’ mechanism, as exemplified by Gummerus et al. (2012). Depending on the
group settings, this endorsement is visible to non-group members, or not. Recommending,
or engaging in word-of-mouth that reaches beyond a specific setting, is also a form of
endorsement, except that it has an external focus. Schau et al. (2009) refer to it as
‘impression management’ in the context of a brand community. It includes all practices
that have an external focus on creating a favourable impression of the brand. This idea of
‘impression management’, also referred to as ‘influence impression’ or ‘word-of-mouth’ is
recurrent in the social media literature and considered as a particularly potent form of
earned media (Ashley and Tuten, 2015). Again, in the context of a virtual brand
community, Brodie et al. (2013) propose that a behavioural sub-dimension of engagement
 96

is ‘advocating’, which occurs ‘when consumers actively recommend specific brands,

products/services, organisations and/or ways of using products or brands’(p7).

It is worth mentioning that all three behavioural sub-dimensions share a common social
aspect. Social benefits are recognised to act as important motivational factors of
engagement, (Calder et al., 2013) and a social dimension has been identified in the
consumer engagement literature (e.g. Patterson et al., 2006). Patterson et al. (2006) define
engagement interactions as the various interactions and connections ‘between the customer
and the front line service employees, between the customer and the organization, between
the customer and the brand, and among the customers themselves’ (p. 3). It can therefore
be said that a social aspect underlies all engagement behaviours, because they entail
interactions and creation of bonds between the engagement subject and partners. The
following table summarises the behavioural dimension and sub-dimensions of consumer
engagement.

Table 8: Behavioural dimension and sub-dimensions

Behavioural dimension of consumer engagement

Definition References
The behavioural manifestations toward an Sawhney et al., 2005; Van Doorn et
engagement partner, beyond purchase, which al., 2010; Verhoef et al., 2010;
results from motivational drivers. Brodie et al., 2011; Gummerus et al.,
2012; Hollebeek, 2011a, 2011b.
Sub-
Definition References
dimension
Sharing The act of providing content, Hennig-Thurau et al., 2004;
information, experience, ideas Sawhney et al., 2005; Mathwick et
or other resources to the al., 2008; Vivek et al., 2012; Brodie
engagement partner. et al., 2013.
Learning The act of seeking content, Hennig-Thurau et al., 2004; Schau et
information, experience, ideas al., 2009; Brodie et al., 2013.
or other resources from the
engagement partner.
Endorsing The act of sanctioning, showing Schau et al., 2009; Van Doorn et al.,
support, referring. In a 2010; Gummerus et al 2012 ; Brodie
community context, et al. 2013.
endorsement can have an
internal or external focus.

To sum up, the above conceptualisation of the dimensions of consumer engagement show
the deep meaning associated with each of each dimensions and sub-dimensions. The
 97

conceptualisation developed here draws on research in the fields of education,

organisational behaviour and other social sciences, as well as conceptual studies in
marketing research, and some exploratory work. The outcome provides a rich and
multifaceted conceptualisation the concept of engagement, one that seems highly
applicable to the context of OBC.

The following figure represents the concept of consumer engagement, focusing on two
different engagement partners in an OBC context, namely the OBC and the brand. It
depicts the dimensions (affective, cognitive and behavioural) and sub-dimensions of the
two constructs. This conceptualisation of consumer engagement in an OBC context is used
throughout the remaining of the study.

Figure 11: Consumer engagement in an Online Brand Community

3.3. Proposed model of consumer engagement in OBC

The review of the literature has identified gaps in the treatment of both antecedents and
outcomes of OBC participation when approached from a consumer engagement
 98

perspective. This section develops a model of consumer engagement in OBC, integrating

its key drivers and outcomes. The general logic concerning the identification of
antecedents and outcomes is presented first, also conceptualising a relationship between
consumer engagement with the OBC and with the brand. Based on this general view, the
specific research hypotheses are developed and a causal model proposed, based on three
main types of relationships.

3.3.1. Drivers of engagement

Consumer engagement can have varying intensities (e.g. Patterson et al., 2006; Hollebeek,
2011). This assertion is in line with the OBC literature, where OBC members can vary
depending on their level of activity in the group. For instance, research differentiates
lurkers from contributors (Kozinets, 1999; Preece et al., 2004). Despite this acceptance that
varying levels of consumer engagement can occur, there is still very little understanding as
to which factors, in particular individual ones, influence engagement levels in an OBC
context.

Studying the individual factors that may affect levels of engagement is of particular
importance, in line with the gap identified in the literature review on OBC. Firstly, because
the individual consumer is the subject of engagement, it is important to gain understanding
into which individual characteristics that play a role in the formation of engagement and
why this is likely to occur. Such insight should allow a better understanding of the
emergence, and sustenance and variations of engagement levels for consumers. Secondly,
consumer engagement being a psychological concept (Brodie et al., 2011), it is fair to
assume that it would be influenced by internal individual pre-dispositions, attitudes and
preferences (Campbell et al., 2014). Thirdly, from a managerial perspective, the individual
drivers of engagement have the potential to help OBC managers better manage their
consumer based through more effective segmentation (Campbell et al., 2014). Although
current research in consumer engagement acknowledges the driving potential of individual
factors over consumer engagement (van Doorn et al., 2010; Wirtz et al., 2013), research in
this direction is scant, largely conceptual and lacking detail.

Given the current state of research in consumer engagement and OBC alike, this study
adopts a focus on individual traits and predispositions as drivers of consumer engagement
in OBC. More specifically, OIP (Wiertz and de Ruyter, 2007), attitude toward OBC
 99

participation (Bagozzi and Dholakia, 2006) and product involvement (Hollebeek et al.,
2014) are studied. These three elements are chosen because they all have an enduring,
long-term and stable aspect congruent with the nature of consumer engagement, which is
further detailed and explained in the hypothesis development section.

3.3.2. Relationship between consumer engagements with different partners

As explained above, consumer engagement can be directed at different engagement

partners, and this study focuses on OBC engagement and online brand engagement.

Little research has so far investigated the interplay between OBC engagement and online
brand engagement, and most studies have approached them separately. However, the
specific nature of OBC allows relationship formation between consumers and different
OBC actors, such as the other consumers, and the brand (McAlexander et al., 2002) within
the same settings. On this ground, Wirtz et al. (2013) identify that OBC engagement
conceptually leads to increased levels of brand engagement, a relationship which this study
aims to test.

3.3.3. Outcomes of engagement

The conceptual model integrates several outcomes of engagement. Existing frameworks

converge in showing that brand benefits derive from consumer engagement, whether from
a conceptual (van Doorn et al., 2010) or empirical standpoint (Brodie et al., 2013). More
specifically, this study aims to show the impact of consumer engagement on brand trust,
commitment and loyalty. In contrast to other studies on consumer engagement and its
brand-related outcomes, brand satisfaction is not taken into account since members of the
online community are not supposed to be existing paying customers of the brand, and have
therefore not necessarily formed brand satisfaction.

Based on these arguments, this study’s conceptual framework is proposed in Figure 12 and
related hypotheses subsequently developed.
 100

Figure 12: Conceptual framework of consumer engagement in an Online Brand

Community

3.4. Hypothesis development

3.4.1. Drivers of consumer engagement

Given that individual traits and preferences have the potential to impact social actions
(Wiertz and de Ruyter, 2007), this section proposes three antecedents to consumer
engagement in an OBC: Online Interaction Propensity (OIP), attitude toward OBC
participation and product involvement. Although membership in an OBC presupposes
some level of interest in group membership and engagement, research has shown that OBC
members can greatly vary in terms of their level and type of participation, ranging from
passive lurkers to active contributors (Nonnecke et al., 2006). Hammond (2000) contends
that there are two types of memberships in online groups: communicative and quiet.
Communicative membership displays frequent interactions, ask questions and answer to
them, whereas quiet membership is translated mainly in the simple reading of messages.
These different types and levels of engagement in an OBC have strong implications in
terms of community vitality and management. The synthesis of existing literature shows
that three individual characteristics could impact consumer engagement in OBC: OIP,
attitude toward OBC participation and product involvement.
 101

Online Interaction Propensity

Online Interaction Propensity (OIP) is a potential indicator of the type of membership that
one has in an OBC. The concept is based on the assumption that people exhibit different
preferences when it comes to interacting with others online. OIP is defined as a trait-based
individual difference in the predisposition to participate in online interactions (Wiertz and
de Ruyter, 2007). OIP is a behavioural predisposition, which is rooted in personality and
individual characteristics. In communication and psychology disciplines, it is referred to as
the willingness to communicate with others and people can exhibit different levels of
propensities to interact (Wiertz and de Ruyter, 2007).

Online settings such as OBC advance the study of interaction propensities in three ways
(Hoffman and Novak, 1996). Firstly, communication can be synchronous or asynchronous
in OBC. Secondly, OBC interactions usually occur between relative strangers, in other
words people that one has never met in real life and that might be located anywhere in the
world (Wiertz and de Ruyter, 2007). Often, the only common denominator is their passion
for the brand. Thirdly, online interactions in OBC might be less rich than offline ones, but
typically have a broader reach (Blazevic et al., 2014).

OIP is a fairly under-researched individual trait when it comes to understanding drivers of

OBC participation. Research has so far focused on the role of participation intention in
driving OBC participation (Algesheimer et al., 2005; Casaló et al., 2010; Zhou et al.,
2013). OIP however differs from participation intention in an OBC. Participation intention
is a narrower concept focused on intended membership in a specific OBC. OIP, on the
other hand, is an enduring characteristic engrained in people’s personality, rather than a
transient intention to do something at a point in time (Wiertz and de Ruyter, 2007).

Therefore, OIP is potentially a particularly potent driver of consumer engagement in OBC.

Although they define engagement as a sum of experiences, Calder et al. (2009) provide a
first basis to support this assertion, as they show that the need for social interactions is a
driver of consumer engagement. Recently, Blazevic et al. (2014) also determined that
General Online Social Interaction Propensity (GOSIP) is an explanatory factor for
consumer engagement and online interaction behaviours. They work on the basic premise
that the core modus operandi of social media platform is interaction, underpinning levels
of engagement. They report that GOSIP has the ability to explain why one person will
engage in online interactions and another will not, under identical circumstances. Despite
 102

these advances, research into OIP remains scarce (Blazevic et al., 2014) and requires
further attention, particularly given the ever-changing nature of OBC settings. Moreover, it
is so far limited to explaining participation behaviours only rather than the full
multidimensional spectrum of consumer engagement.

Based on previous findings, we hypothesise that OIP has a positive impact on consumer
engagement, both with the OBC and the brand. This leads to the articulation of hypotheses
1a and 1b:

H1a: OIP is positively related to OBC engagement.

H1b: OIP is positively related to online brand engagement.

Attitude toward OBC participation

Following Wu and Chen (2005), the attitude toward OBC participation reflects in this
study the favourable or unfavourable assessment a consumer makes of participating in the
OBC. Particularly developed in the TPB (Ajzen, 1991), attitude is a psychological concept,
which aims to explain individual decision making. The attitude toward an act is one of the
key determinants of the intention to act, leading ultimately to actual action. Attitude is a
valenced evaluative (cognitive) reaction to an action. It is based on the belief that a certain
action, behaviour or state is going to lead to favourable or unfavourable consequences
(Eagly and Chaiken, 1993). An attitude therefore ranges on a spectrum from negative to
positive. In this sense, it departs from OIP, which is not valenced, or evaluative.

Attitudes are created through long-term learning and acquired over time. They are
relatively stable, learned predispositions that are retrieved or activated to influence
decisions or behaviours (Bagozzi and Dholakia, 2006). In this sense, they are, like
engagement, going beyond purchase or consumption situations and built in the long run,
through an accumulation of experiences. They are not contingent on the occurrence of a
specific event, but are retrieved when needed. Since attitudes and engagement are both
relying on enduring mechanisms, attitudes toward OBC participation can be considered as
an adequate driver of consumer engagement.

Attitude’s impact on intentions and behaviour has been examined in online settings, using
the TPB as an overarching framework (e.g. Hsu et al., 2006; Pavlou and Fygenson, 2006).
More specifically, social media and OBC literature report the necessity to consider it as a
driver of the intention to act and actual actions in group settings. In an extended version of
 103

the TPB, Bagozzi and Dholakia (2006), find that attitude toward brand community
participation is a driver of the desire, intention and behaviour of community participation.
Similarily, Casaló et al. (2010), show that members’ attitude toward participation in a firm-
hosted online travel community is a potent driver of the actual intention to participate.
These findings find further conceptual validation in Hennig-Thurau et al. (2010) who
assert that consumers with high positive attitudes toward new media are more likely to
exhibit high levels of new media brand engagement. New media brand engagement can be
enacted through active types of engagement such as creating content, or watching a
branded video. This leads to the development of the following two hypotheses regarding
consumer engagement in OBC:

H2a: Attitude toward OBC participation is positively related to OBC engagement.

H2b: Attitude toward OBC participation is positively related to online brand

engagement.
 104

Product involvement

Product involvement is defined in this study as an on-going concern or interest for a

product class, regardless of situational influences (Richins and Bloch, 1991). The concept
has an emotional and a cognitive dimension, and stems from a holistic experience with
products (Chaudhuri, 2000). In line with this definition and within the broad context of
consumer engagement in OBC, this study focuses on enduring involvement, in contrast to
situational involvement. Unlike situational involvement, which tends to be highly related
to purchase situations and choices, enduring engagement goes beyond purchase and has a
long-term scope (Laurent and Kapferer, 1985), which renders it more relevant than
situational engagement as a potential predictor of ongoing consumer states such as
engagement.

The general relevance of the concept of involvement for marketing scholars lies in its
central mediating role in the prediction of consumer behaviour (Mittal and Lee, 1989).
Involvement is also a useful segmentation criterion, as consumers can experience different
levels of involvement (e.g. Traylor and Joseph, 1984; Beatty et al., 1988; Warrington and
Shim, 2000). More specifically, the relevance of product involvement in determining
consumer engagement can be attributed to several factors.

Firstly, numerous studies recognise the mediating role of involvement in predicting

consumer behaviour, attitudes or intentions akin to consumer engagement. These can
broadly be related to the extensiveness of the purchase decision process, enacted with
constructs such as information search, interest in and consumption of advertising (Mittal
and Lee, 1989; Flynn and Goldsmith, 1993), shopping enjoyment (Mittal and Lee, 1989),
shopping effort, brand comparison or selectivity (Traylor and Joseph, 1984). Frequency of
purchase and amount spent, product usage and loyalty are other facets of involvement
response (Mittal and Lee, 1989, Flynn and Goldsmith, 1993). Word-of-mouth (Richins et
al., 1992) and satisfaction (Richins and Bloch, 1991) are also influenced by involvement
levels. Finally, the strong link that product involvement and brand commitment exhibit has
been the subject of much research, making brand commitment a well-acknowledged
outcome of high involvement (Traylor and Joseph, 1984; Mittal and Lee, 1989; Pritchard
et al., 1999).

Secondly, there is an understanding that consumer engagement might be related to the type
of product under consideration. It seems however unclear how product type can impact
 105

engagement and most of the research in this domain remains conceptual. So far,
engagement scholars in marketing have placed much emphasis on service brands (e.g.
Bowden, 2009; Brodie et al., 2011) and some authors propose that a change in the type of
engagement product can lead to different levels of cognitive, emotional and behavioural
engagement (Brodie et al., 2011). Whether the focal brand is hedonic or utilitarian and
whether involvement with this kind of products is high or low could impact consumer
engagement (Hollebeek, 2011). Other scholars propose that consumers are more quickly
disengaged with utilitarian services than with co-creative services (Bowden et al., 2014).
Based on the idea that different types of products or services can elicit different levels of
consumer involvement and that engagement might be related to it, the study of product
involvement as an antecedent to consumer engagement allows shedding light on the
relationship between consumer engagement and the type of products a brand sells.

Thirdly, although different from it, consumer engagement is closely related to the concept
of involvement. Involvement differs from engagement in major ways. For instance, one
might be highly involved in selecting a new computer, but this does not necessarily entail
the kind of interactive experiences that Calder et al. (2013) perceive as engagement.
Taking a similar perspective, Vivek et al. (2012) explain that whereas involvement is a
form of interest and implies high levels of information processing, it is different from
engagement because it does not entail a behavioural aspect, when engagement does.

Recent studies document the link between the two concepts and involvement is
conceptualised as an antecedent of consumer engagement (e.g. van Doorn et al., 2010). To
date, this relationship has been empirically confirmed once, in a study developing a scale
of customer engagement. In this study, Hollebeek et al. (2014) prove the empirical
distinctiveness of engagement and involvement, further validating the mediating role of
engagement in the relationship between involvement and self-brand connection. Although
empirical evidence is scant, there are strong conceptual foundations to support the
assertion that product involvement precedes consumer engagement in OBCs. This leads to
the development of the following set of hypotheses:

H3a: Product involvement is positively related to OBC engagement.

H3b: Product involvement is positively related to online brand engagement.

 106

3.4.1. Relationship between online brand engagement and OBC engagement

Existing literature on OBC identifies clear causal effects between the relationships
members form with other members of the community, and the ones they have with the
focal brand (e.g. Bagozzi and Dholakia, 2006; Veloutsou and Moutinho, 2009).
Understanding the dynamics of consumer interactions and relationships in an OBC is
crucial because these interactions are what contribute to the OBC creation, sustenance and
vitality (Casaló et al., 2008) and ultimately impact customer relationships and brand
management strategies (Hennig-Thurau et al., 2010).

On the one hand, community participation, behaviours, commitment and attitudes have
been approached as drivers of subsequent brand-related relationship building. For instance,
OBC participation is known to lead to stronger product relationships (Stokbürger-Sauer,
2010), intentions to recommend the host firm (Casaló et al., 2010), and dedication to the
brand’s success (McAlexander et al., 2002). Feelings elicited from interactions with other
community members shape consumer’s attitudes about the host brand and future
interactions with it within the community: for instance, consumers engage in product
support by giving feedback to the brand through the community interface (Nambisan and
Baron, 2007). Similarly, community commitment leads to higher levels of brand
commitment (Kim et al., 2008). Schau et al. (2009) also find that community practices
foster engagement with the focal brand. One can thus assume that individuals who are
more committed to an OBC are more likely to develop positive attitudes and behaviours
toward the brand. These identified relationships are not directly related to the concept of
engagement as defined in this study; however each of them taps into certain aspects of
engagement, such as enthusiasm for the brand, OBC engagement behaviours or brand
endorsement. In this sense, these relationships propose a baseline to support the hypothesis
that OBC engagement is conducive of brand engagement.

Further evidence of this link can be found in research dealing directly with consumer
engagement applied to OBC settings. Although limited to a conceptual framework, Wirtz
et al. (2013) state that OBC engagement has a positive influence over brand engagement.
Their conceptualisation of engagement is close to the stance taken in this study, as it
encompasses attitudinal and behavioural aspects, of engagement, views it as a
motivational, interactive and experiential constructs, and refers to specific instances of
engagement, which resonate with our sub-dimensions. Going further into this direction,
Jahn and Kunz (2012) empirically prove the impact of fanpage engagement on Facebook –
 107

the integrative and interactive participation in the community – on a number or brand-

relationship concepts, including brand WOM, loyalty, purchase and commitment.
Although these relational outcomes place little focus on intra-community engagement with
the brand, they are an indication of the potential of OBC engagement to foster brand
engagement. These studies clearly show support for the precedence of OBC engagement
on brand engagement both in the OBC and consumer engagement literature.

Nevertheless, it is important to also note that the reverse relationship has been validated.
Studies show that consumer relationships with the brand precede brand community
memberships. More specifically, brand relationships impacts brand identification, which in
turn leads to community engagement (Algesheimer et al., 2005). Despite the validation of
this causal link, the authors however also show that community engagement leads to
further brand-related behaviours such as repeat purchase, showing support to the positive
impact of community participation over brand outcomes.

Hypothesising the positive impact of OBC engagement over brand engagement is

congruent with the scope and focus of this study, for several reasons. One of the boundary
assumptions of this study is its focus on the engagement of existing consumers of the
OBC. Understanding how prior relationship with the brand leads the individual to become
a member of the community therefore does not fall within the scope of the research.
Instead, how being a member of the community can foster one’s relationship with the
brand is of interest. Moreover, studies have shown that the relative importance of brand
tribalism (i.e. a neighbour concept to brand community participation) over brand reputation
is stronger when it comes to creating brand relationships (Veloutsou and Moutinho, 2009).
This suggests that studying the impact of OBC engagement over brand engagement is of
particular interest and relevance. For these reasons, the following hypothesis is presented:

H4: OBC engagement is positively related to online brand engagement.

3.4.2. Outcomes of engagement

This study focuses on two potential direct outcomes of online brand engagement, brand
trust and brand commitment, and subsequently proposes that online brand engagement
drives brand loyalty through the mediating effect of these two constructs. Because trust
and commitment have often been paired, from foundational brand relationship research
(e.g. Chaudhuri and Holbrook, 2002) to recent brand engagement advances (Hollebeek,
2011), they are approached together as two facets of brand-related outcomes.
 108

Brand commitment and brand trust

In this study, brand trust is defined in line with Chaudhuri and Holbrook’s (2001, 2002)
conceptualisation, which builds on Moorman et al. (1992). Brand trust is the willingness of
the consumer to rely on the ability of the brand to perform its stated function. In this
situation, the brand is a relationship partner in which the customer has confidence. The
brand shows integrity and honesty, and because any relationship might entail risk, it is seen
as a safe relationship partner (Morgan and Hunt, 1994). Brand commitment, on the other
hand, is the enduring desire to maintain a valued relationship with a brand in the long term
(Morgan and Hunt, 1994). It is based on the emotional or psychological attachment to and
preference for a brand within a product category (Lastovicka and Gardner, 1977). Brand
commitment is considered as a unidimentional, attitudinal construct in this study and it has
an important long term, enduring aspect (Moorman et al., 1992; Morgan and Hunt, 1994).
Both commitment and trust have been conceptualised as elements of relationship quality
(Rafiq et al., 2013) and the mediating role of trust and commitment in marketing
relationship exchanges is well understood (Morgan and Hunt, 1994; Chaudhuri and
Holbrook, 2001).

There are several reasons why brand engagement leads to brand trust and brand
commitment. Firstly, the very nature of brand engagement is based on reciprocity, which is
a necessity to build trust. Brand engagement is a two way-relationship in which customers
invest time, energy and effort (Vivek et al., 2012). As a counterpart of the consumer’s
efforts, the level of investment and dedication to the relationship from the part of the brand
will also influence consumer trust (Rafiq et al., 2013). This can be translated in the amount
of formal and informal sharing of meaningful and timely information between the brand
and the consumer, which leads to higher levels of trust (Morgan and Hunt, 1994). In
contrast to this, a brand exhibiting opportunistic behaviour, i.e. self-interest seeking with
guile, leads to lower levels of trust (Morgan and Hunt, 1994). In general, consumers with a
high relational orientation place higher importance on trust than low-relational customers
(Garbarino and Johnson, 1999).

Secondly, brand engagement lowers the risk associated with a brand and increases
preference for it, because it creates familiarity and closeness. We know that identification
with the brand or the brand community to which the customer belongs impacts trust
(Marzocchi et al., 2013). Moreover, shared values between the consumer and the brand
foster trust (Morgan and Hunt, 1994). By engaging with the brand, the consumer creates a
 109

strong bond with it, on an emotional, cognitive and behavioural level. This close
connection, built over repeat interactions, is likely to lead to increased trust and
commitment (Chaudhuri and Holbrook, 2002).

The link between brand engagement and brand trust and commitment is also prominent in
the consumer engagement literature, although the direction of causality seems unclear.
Most of the conceptual research so far takes a careful stance by proposing that brand trust
and commitment can either be outcomes or antecedents of brand engagement (van Doorn
et al., 2010; Hollebeek, 2011; Brodie et al., 2011). More specifically, some advance that
there is a feedback loop between brand engagement and brand commitment and trust,
which continuously evolves as consumers keep engaging with brands and reinforcing their
relationship with them (van Doorn et al., 2010). Whether engagement precedes trust and
commitment or derives from them could also be attributed to whether the individual is an
existing customer of the brand or not (Hollebeek, 2011). Existing customers would have
developed trust and commitment prior to engaging with the brand, whereas new customers
would only develop trust and commitment after being engaged with the brand. Despite this
conceptual blurriness, according to Gambetti and Graffigna (2010) engagement, in the
relational sense of the term, is a way to build trust and commitment with the brand, rather
than the other way around.

To conclude, a clear link between brand engagement and brand trust and commitment
emerges from the brand relationship literature, feeding through OBC and consumer
engagement research. The boundary conditions in which this study operates allow
clarifying the hypothesised direction in which this relationship is likely to operate.
Consumers of interest in this research are the members of the community, and can be
either existing or non-existing paying customers of the brand. A prior strong relationship
with the brand can therefore not be assumed, as the OBC might be the first touch point the
consumer has with the brand. Based on this understanding, this study is therefore interested
in the relational outcomes built from online brand engagement in the OBC settings, and the
following hypotheses are proposed:

H5: Online brand engagement is positively related to brand trust.

H6: Online brand engagement is positively related to brand commitment.

 110

Brand loyalty

Loyalty is a complex construct, which has been conceptualised as both attitudinal and
behavioural (Jacoby and Chesnut, 1978). This study focuses solely on behavioural
manifestations of brand loyalty, and repeat purchase behaviour is used as its measure. In
agreement with Odin et al. (2001), this study proposes that the most direct way to measure
loyalty is to focus on the repeat purchasing behaviour of the same brand, as declared by
consumers. It is acknowledged that repeat purchasing behaviour does not represent the full
picture of loyalty, and that if it is a function of inertia, it is less valuable. However,
considering the context of engagement, one can assume that the level of customer inertia
toward a brand is relatively low, therefore rendering repeat purchase behaviour a
satisfactory indicator of brand loyalty. Moreover, this study takes the stance that brand
commitment is the manifestation of attitudinal loyalty and is thus distinct from behavioural
brand loyalty (Jacoby and Chesnut, 1978).

Behavioural loyalty has been the focus of much research in the brand community and OBC
literature, and has come through as one of the key outcomes of OBC participation (e.g.
Algesheimer et al., 2005; Bagozzi and Dholakia, 2006; Jang et al., 2008; Thompson and
Sinha, 2008; Casaló et al., 2010). The more community members become integrated into a
community through frequent and repeat interactions with other members and the brand, the
more brand loyalty they are likely to display (McAlexander et al., 2002). From the
perspective of the TPB, intention and desires can trigger group behaviour, which
ultimately leads to increased brand loyalty (Bagozzi and Dholakia, 2006). More
specifically, antecedents of behavioural brand loyalty in OBC contexts include brand
community identification (Heere et al., 2011), brand identification (Stokbürger-Sauer,
2010), emotional support and encouragement from the brand (Fournier and Lee, 2009),
uncertainty reduction through C2C communication (Adjei et al., 2010). These studies all
support the notion that OBC participation can strongly trigger brand loyalty among its
members.

Whether loyalty outcomes are the same when approaching OBC participation from a
consumer engagement perspective is less obvious. Looking at the consumer engagement
literature and its take on brand loyalty allows crossing this bridge. In an extensive literature
review, Hollebeek (2011a) highlights the complex relationship between loyalty and
engagement. It appears, from a conceptual standpoint, that brand loyalty is the ultimate
stage of any consumer engagement process (Bowden, 2009; Calder et al., 2013), and that
 111

the causal link between consumer engagement and loyalty is mediated by other relational
constructs, including trust and commitment (Hollebeek, 2011a).

The mediating role of trust and commitment in driving loyalty from engagement seems
generally supported (Hollebeek, 2011a). The relationships between brand commitment and
brand trust with brand loyalty are verified in a number of relationship marketing studies
(e.g. Garbarino and Johnson, 1999; Gruen et al., 2000). In a brand community context,
Marzocchi et al. (2013) also support this assertion, showing that attitudinal loyalty (a
synonym for brand commitment) and brand trust lead to increased levels of behavioural
loyalty. Aligning early relationship marketing studies with OBC and consumer
engagement research allows positing that, with the precedence of consumer engagement
over brand trust and commitment, the following hypotheses hold in an OBC context:

H7: Brand trust is positively related to brand loyalty.

H8: Brand commitment is positively related to brand loyalty.

3.5. Summary of the hypotheses

The proposed relationships between the independent, mediating and dependent constructs of
this study are visually represented in presented in Figure 13 below and summarised in Table 9.

Figure 13: The causal model

 112

Table 9: The research hypotheses

Antecedents of OBC engagement

H1a OIP is positively related to OBC engagement.
H2a Attitude toward OBC participation is positively related to OBC engagement.
H3a Product involvement is positively related to OBC engagement.
Antecedents of online brand engagement
H1b OIP is positively related to online brand engagement.
Attitude toward community participation is positively related to online brand
H2b
engagement.
H3b Product involvement is positively related to online brand engagement.
H4 OBC engagement is positively related to online brand engagement.
Outcomes of online brand engagement
H5 Online brand engagement is positively related to brand trust.
H6 Online brand engagement is positively related to brand commitment.
H7 Brand trust is positively related to brand loyalty.
H8 Brand commitment is positively related to brand loyalty.

3.6. Validation of the study

Validating measures and models across contexts is an important endeavour in quantitative

studies, which aims at proving their generalisability. Consumer engagement studies in
particular have called for assessment of the concept across contexts (Gambetti and
Grafignia, 2010; Brodie et al., 2011) and recent scaling literature has started addressing
this gap (Vivek et al., 2014). This section proposes a way to validate the expected results
of this study. More specifically, a cross-contextual validation of the study will be
performed, focusing on different languages.

OBC and engagement studies can seek cross-contextual validation in different ways: it can
for instance be done using data collected for different brands (e.g. Phillips and McQuarrie,
2010), product categories (e.g. Wallace et al., 2014) or in different languages (Ouwersloot
and Odekerken-Schröder, 2008). This section explains why validating the study’s results in
another language is particularly relevant given the context of investigation and the state of
 113

the literature. It also explains why two specific languages were chosen, namely English
and French.

This study will attempt to validate the conceptual model with its measurement and
structural components across different languages, for several reasons. The key reason
concerns the relative paucity of cross-cultural studies in OBC research. A review of the
current cultural contexts of investigation in both OBC and consumer engagement literature
show the lack of cross-cultural or linguistic validation existing so far, as well as a strong
bias toward English cultures. Table 10 details the context of investigation of the key OBC
and consumer engagement studies using empirical data in terms of the brands, product
categories and cultural contexts. The table reveals that out of 39 selected papers, 24 of
them collected data in English, 13 of which were situated in the United States. A few
studies were realised in German, Italian, Chinese or Hindu, and a small amount of them
did not specify the country or language of investigation. Out of this pool of studies, only
one collected data in two different languages (French and Dutch, in Ouwersloot and
Odekerken-Schröder, 2008). However, this was done in the aim of pooling the two
samples together rather than for validation purposes.

Even though it seems that validation has been sought in OBC and consumer engagement
studies by including several brands or product categories in the data collected (see Table
10, e.g. Vivek et al., 2014), a complete lack of validation using different languages is also
evident. Although it might be understood for studies operating on a local, real-life basis
(e.g. Jaakkola and Alexander, 2014) it is surprising that no study investigating
communities or engagement in online settings has to date sought such validation, given the
lack of geographic boundaries inherent to online environments.

Yet, a basic premise of OBCs operating on social media including Facebook is that they
are global by nature, bound only by the language in which they are set up. In 2014,
Facebook supported 70-plus languages (Facebook, 2014). Social media allow
communication and interactions regardless of geographic and cultural boundaries and this
global nature of social media brings challenges for brand page management, if a brand is to
ensure consistent brand positioning at a global level. Gensler et al. (2013) pinpoint the
importance for brand managers to be able to ensure consistent brand stories on a social
media site when this site is used around the globe by consumers who might have
completely different interpretations of brand meaning. The global nature of social media
(Kaplan and Haenlein, 2010) therefore brings about a key cultural consideration: should
 114

consumers from different linguistic backgrounds be engaged with differently on social

media?

Two opposing stances are taken when considering the global and cross-cultural aspect of
social media and Facebook in particular. Some authors are proponents of the view that
consumers around the world behave differently and that social media marketing need to be
adapted on this basis (Jackson and Wang, 2013). Particularly, highly collectivist cultures
such as China seems to greatly differ from highly individualistic cultures like the US
(Jackson and Wang, 2013) when it comes to social media behaviour. Others argue that
there is a convergence in the way people from different cultural backgrounds behave
(Douglas and Craig, 2006). Since geographic boundaries are eliminated on social media,
people from all cultural background and languages are able to use the same platforms in
the same ways. They are able to interact with like-minded people located worldwide, using
the same protocols or interaction (Zaglia, 2013). Although cultures might be different, they
access the same services and use the same platforms.

Moreover, social media and Facebook page management need to be in line with the
company’s global marketing approach. Depending on whether the brand has a unified
global or differentiated local strategy, it should be able to use Facebook in a way that
reflects their global strategy. Recognising this need, Facebook has made its pages highly
customisable and companies have since 2012 the ability to embed local pages under their
global page, thus allowing them to communicate in the consumer’s language and have
different local content, but communicated and structured in a coherent manner (Facebook,
2012). Despite these functionalities, it is still unclear whether Facebook page managers
should manage their pages globally or locally to foster increased levels of engagement.

Based on these practical and theoretical considerations, this study aims to validate its
results in another language than the initial language of investigation, namely English.
French is chosen as an appropriate language for the purpose of cross-language validation
because these two languages and the cultures they represent exhibit enough variance on
specific cultural dimensions to be considered different, while at the same time remaining
comparable in a number of ways. The GLOBE framework (House et al., 2004), Hofstede’s
cultural dimensions (Hofstede et al., 2010), as well as existing cross-cultural marketing
research are used to compare the two cultures.
 115

Firstly, according to the GLOBE study and societal clusters (House et al., 2004), any
English-speaking sample belongs to the ‘Anglo’ cultures cluster, which are competitive,
result-oriented and have low levels of in-group collectivism, whereas a French-speaking
sample belongs to the ‘Latin European’ cluster, which tend to highly value individual
autonomy and have average levels of in-group collectivism.

Considering Hofstede’s cultural dimensions, differences appear between French and

English-speaking countries in terms of power distance, uncertainty avoidance and
pragmatism. French-speaking countries tend to have a much higher power distance,
meaning that they are more equalitarian. They also rank higher in terms of uncertainty
avoidance, preferring clear structures and rules than English-speaking societies. Lastly,
they have higher pragmatism, showing a better ability to adapt to changing situations and
contexts (Hofstede, 2014).

Focusing on the uncertainty avoidance, this particular cultural trait also bears impact on the
level of technology penetration and innovation levels. With higher levels of uncertainty
avoidance, countries like France or Belgium tend to exhibit lower levels of technology
penetration, lower amount of innovators in the population (de Moij, 2011), as well as lower
levels of consumer co-creation through online channels and more reluctance to sharing
personal information on the Internet for French-speaking cultures (Garnier and McDonald,
2009). A mapping of cultural differences in Internet usage reveals that France and Belgium
are clear laggards, whereas the UK, USA, and Ireland are on the innovators side (Hofstede
et al., 2010)

Some similarities however exist between the two cultures. Belgium, France and the UK are
all characterised as having a ‘middle’ context in terms of communication style (Hall 1976),
meaning that they are neither extremely explicit nor extremely implicit cultures. This
implies that the explicitness of their communication on social media should be about the
same. Moreover, French-speaking countries and English speaking countries have similar
levels of individualism according to Hofstede (2014), both being quite high. This is an
important factor to consider when studying social media interactions and it suggests that
both cultural samples will behave similarly in terms of social interactions, which might be
different with highly collectivist cultures such as China (Jackson and Wang, 2013).
 116

Table 10: Research contexts in key OBC and consumer engagement studies

Paper Language (Country) Brand Product category

Muñiz and O'Guinn, 2001 English (US) Ford Bronco, Macintosh, and Saab Automotive and Technology
McAlexander et al., 2002 English (US) Jeep and Harley Davidson Automotive
Dholakia et al., 2004 English (US) Unknown Unknown
Muñiz and Schau, 2005 English (US) Apple Newton Technology (hardware)
Bagozzi and Dholakia, 2006 English (US) Harley Davidson Automotive
Carlson et al., 2008 English (US) US theme park Entertainment
Technology, Food and Beverage, Culture,
Schau et al., 2009 English (US) Multiple brands
Automotive, Cosmetics
Adjei et al., 2010 English (US) Unknown Machinery and telecoms
Heere et al., 2011 English (US) University football teams Entertainment
Calder et al., 2009 English (US) Several newspapers and magazines Media
Multiple brands including Pantene,
Phillips and McQuarrie, 2010 English (US) Fashion and beauty
Michael Kors, Dolce & Gabbana, etc.
Calder et al., 2013 English (US) Several newspapers and magazines Media
Multiple brands including Costco, Multiple categories including retail,
Vivek et al., 2014 English (US)
Apple, Xbox. entertainment and technology
Gruner et al., 2014 English Unknown Durable goods
Habibi et al., 2014 English Jeep and Harley Davidson Automotive
Hollebeek and Chen, 2014 English (New-Zealand) Apple and Samsung Mobile Telecom
Hollebeek et al., 2014 English (New-Zealand) Facebook, LinkedIn and Twitter Social media
Scott and Craig-Lees, 2010 English (Australia) Unknown Entertainment
So et al., 2013 English (Australia) Unknown Tourism
O'Sullivan et al., 2011 English (Ireland) Beamish Beer Food and Beverage
 117

Paper Language (Country) Brand Product category

Multiple categories including beverages,
Wallace et al., 2014 English (Ireland) Multiple brands
fashion, retail, music, etc.
Brodie et al., 2013 English (New-Zealand) Vibra-Train Health and Fitness
Multiple brands including Nivea, Multiple categories including cosmetics,
Hollebeek, 2013 English (New-Zealand)
Mercedes-Benz or Disney automotive, entertainment etc.
Jaakkola and Alexander, 2014 English (Scotland) Scotrail Transportation services
Matzler et al., 2011 German Volkswagen Automotive
Füller et al., 2008 German (Austria) Volkswagen Automotive
Volkswagen, Mercedes, BMW, and
Algesheimer et al., 2005 German (Germany) Automotive
other car brands
Hung et al., 2011 Chinese (China) Multiple brands Multiple categories
Zhou et al., 2013 Chinese (China) Unknown Unknown
Cova and Pace, 2006 Italian (Italy) Nutella Food and Beverage
Marzocchi et al., 2013 Italian (Italy) Ducati Automotive
Ouwersloot and Odekerken- Dutch and French (The
Settlers of Catan and Swatch Entertainment and Fashion
Schröder, 2008 Netherlands and Belgium)
Sarkar and Sreejesh 2014 Hindi (India) Unknown Automotive
Intel, AMD, ATI and NVIDIA
Thompson and Sinha, 2008 Unknown Technology (hardware)
microprocessors and video cards
Kim et al., 2008 Unknown Herbal product brand Cosmetics
Stokbürger-Sauer, 2010 Unknown Unknown Online service
Homburg et al., 2015 Unknown 10 different brands Unknown
Multiple brands, including Citroen, Multiple categories including beverages,
Gambetti et al., 2012 Unknown
Coca-Cola, Dove, etc. automotive, personal care etc.
Gummerus et al, 2012 Unknown Unknown Gaming
 118

3.7. Summary

This chapter has presented the development of the conceptual framework of the study.
Firstly, the concept of consumer engagement has been defined in terms of the OBC
context. By doing so, three boundary conditions of this study have been determined: (1)
The subject of engagement is identified as the consumer, who is defined as the individual
member of the OBC; (2) The partners of engagement are on the one hand the OBC,
representing the community of other OBC members, and on the other hand the brand,
enacted by the page manager(s) and (3) The three dimensions and seven sub-dimensions of
consumer engagement are defined and detailed. On this basis, Figure 9 depicts the partners,
dimensions and sub-dimensions of consumer engagement in an OBC.

Based on this understanding, the conceptual model of consumer engagement in OBC has
been proposed, offering a range of antecedents and outcomes of consumer engagement, as
well as the relationship between consumer engagement with the OBC and with the brand.
Specifically, the antecedents include OIP, attitude toward OBC participation and product
involvement. The outcomes encompass brand trust, commitment and behavioural loyalty.

The chapter closed with a discussion concerning the validation of the study’s hypotheses.
Validating the original English survey results in another language has been deemed
necessary and French is proposed as a validation language. The benefits and rationale for
this choice have been explained in detail.
 119

Chapter 4: Methodology

4.1. Introduction

This chapter outlines the key methodological choices and the rationale driving the
empirical design of the study. A series of decisions regarding methodology are made
(Bryman, 2008) and presented in this chapter. The process of research and different steps
implemented in order to seek out knowledge in the context of this study are detailed
(Schwandt, 2007). Care is given to each aspect of the methodology, ensuring that they are
consistent with one another and with the research question articulated in this study.

To establish a framework guiding the methodological choices of the study, a thorough

literature review allowed clarifying the concepts under investigation and their
relationships. This leads to the development of a conceptual framework of consumer
engagement in OBC and a series of hypotheses, making up the causal model of this study.
These were presented in Chapter 3 along with the specification of the domain of interest
and boundary assumptions of the study.

This chapter details each of the methodological choices made in order to achieve the
study’s objective and provides a rationale for these choices, at strategic and tactical levels.
The detailed research process is presented in Figure 14 below. Firstly, in line with the
study aims and objectives, a paradigmatic stance is chosen and argued for in detail. The
context in which the study is set is then outlined. Data collection methods are then
presented, focusing on the research instrument creation, administration, pretest and pilot.
The sampling issues are then addressed, presenting this study’s approach to sampling and
response bias, as well as the sample characteristics. A presentation of the chosen data
analysis techniques concludes the chapter1.

Moreover, the chapter addresses elements of method equivalence and bias related to
validation of the study with a French-speaking sample. These permeate every aspect of the
methodology (instrument design, administration and sampling) (van de Vijver and Tanzer,
2004) and are thus detailed throughout the chapter.

1
Although measurement of existing scales and the consumer engagement scale development are part of the
research process, these aspects of the research are presented in a dedicated chapter. Since they represent a
substantial part of this thesis, focusing a full chapter to these issues was deemed appropriate.
 120

Figure 14: The research process

 121

4.2. The research paradigm

A paradigm, worldview or ‘weltanschaüng’ is a set of linked assumptions about the world,

which is shared by a community of scientists investigating that world (Kuhn, 1962).
Paradigms are fundamental because they serve as guides, or frameworks for any given
issue under investigation. They allow researchers to find out which problems are worthy of
exploration and provide organising principles and the criteria to choose the appropriate
research tools (Filstead, 1979). To each worldview corresponds a set of philosophical
assumptions. Denzin and Lincoln (2011) refer to these assumptions as basic beliefs, and
include ontology, epistemology and methodology. Ontology answers the question ‘what is
the nature of reality?’ while epistemology addresses the relationship between the
researcher and the researched (Creswell, 2007).

Different philosophical assumptions regarding epistemology and ontology lead to different

paradigms, and these are closely related to the stated aims and objectives of a piece of
research. The main research paradigms are summarised by Denzin and Lincoln (2011)
under five overarching philosophical stances: positivism, post-positivism, critical realism,
constructivism and participatory.

In line with the research questions and study objective, the paradigmatic stance driving this
research is post-positivism. This study adopts an objective ontology and a modified
objective epistemology. Following an objective ontology, this study is based on the belief
that there is an objective reality in the world, independent of what we think about it, and
the purpose of this study is thus to reach an accurate representation of this reality.
However, in line with Johnson and Duberley (2000), the contingent, negotiated and
dynamic nature of social structures is acknowledged, as well the active role of individuals
in creating what they apprehend. For this reason, this study adopts a modified objectivist
epistemology in the sense that objectivity remains a regulatory ideal, which is however
difficult to maintain (Healy and Perry, 2000). On these ontological and epistemological
grounds, the study aims to apprehend reality as closely as possible but admit that this
cannot be done perfectly due to the fallibility of observations.

This reasoning is further supported by the nature of the research questions. RQ 2 and RQ 3
are primarily concerned with causal relationships and seeking to identify clear, objective
relationships between consumer engagement, its drivers and outcomes. These questions are
 122

thus rather guided by a positivist agenda, which can be defined as seeking truth in
causality.

Positivism ‘traces its origins back to the great social theorists of the nineteenth century
and especially to Auguste Comte and Emile Durkheim. The positivist seeks the facts or
causes of social phenomena with little regard for the subjective states of individuals’
(Bogdan and Taylor, 1975, p. 2). For positivists, there is a single truth that can be
measured and studied with total objectivity, with no interaction between the researcher and
the researched (Guba and Lincoln, 2005). Because positivism assumes that social reality is
made up of objective facts, it entails that value-free researchers can precisely measure and
use statistics to test causal theories. Methods used are grounded in the conventional natural
sciences and quantitative methods are often, but not exclusively used. Positivism often uses
a deductive approach of inquiry to test general law (Bryman, 2008), in which the
replication and falsification principles play an important role. In this sense, RQ2 and RQ3
seem clearly in line with a positivist perspective. However, the central role of consumer
engagement in this exercise shifts the adequate paradigm toward a post-positivist approach.

Indeed, RQ1, upon which RQ2 and RQ3 are dependent, requires a slightly different
treatment. There is an exploratory and subjective aspect associated with the exercise of
creating a measure of consumer engagement. Prior to reaching an adequate measurement
of the concept, it is important to engage in qualitative research in order to inform the
quantitative phase and create adequate items to measure the concept. This research
question therefore requires a modified approach to the positivistic epistemology inherent to
RQ2 and RQ3.

As a critical offspring of positivism, post-positivism bears similarities and differences with

it. Like logical positivism, post-positivism is often associated with quantitative research
and favours deductivism and hypothesis testing for theory verification (Creswell and Clark,
2008). In this sense, post-positivists pursue objectivity. However, they also acknowledge
that we can only approximate nature. While positivists believe that the researcher and the
researched person are independent of each other, post-positivists accept that theories,
background, knowledge and values of the individuals can influence what is observed
(Robson, 2011) and that nature can never be fully understood. Scientific method and
hypothetical deduction are still favoured, but structured qualitative approaches and
questions are more prominent than in positivism (Guba and Lincoln, 2005; Merriam et al.,
 123

2007). Overall, post-positivists consider one reality, but several perceptions of that reality
must be combined to obtain a better picture of it (Healy and Perry, 2000).

Consistent with post-positivism, the study follows a series of logical steps where methods
are predominantly quantitative, but purposive qualitative data are used, and essential to
informing and fine-tuning quantitative phases (Creswell, 2007). As Bryman (2008) holds,
deductivism carries elements of inductiveness as well, and this research therefore carries
some elements of qualitative enquiry to support a predominantly quantitative focus
(Ritchie et al., 2013).

A cross-sectional design is used, with an embedded sequential approach for the purpose of
the consumer engagement scale development: the core quantitative phase is complemented
in this study by a supporting qualitative phase (Ritchie et al., 2013) for the main purpose of
item generation and domain specification (Denzin and Lincoln, 2011). The qualitative
phase of this study thus has in input role and is mostly instrumental to addressing RQ1,
whereas the quantitative phase has an output role and informs the empirical findings.

4.3. Study settings

4.3.1. Social media

This study is set in the context of social media, focusing more specifically on Facebook as
an ideal social media platform for the study of OBC. Social media are firstly chosen as an
appropriate context adapted to the investigation of consumer engagement in OBC.

‘Social Media is a group of Internet-based applications that build on the ideological

and technological foundations of Web 2.0, and allow the creation and exchange of User
Generated Content’ (Kaplan and Haenlein, 2010, p. 61).

The extant OBC literature shows that social media represent adequate environments for the
investigation of OBC dynamics. More specifically, studies have investigated the
functioning and benefits of OBCs in environments such as proprietary brand websites and
forums (Casaló et al., 2008; Casaló et al., 2010; Cova and White, 2010; Healy and
McDonagh, 2010; Garnefield et al., 2012), bulletin boards (Dholakia et al., 2004; Faraj and
 124

Johnson, 2011), online chat systems and multiplayer virtual games (Dholakia et al., 2004)
and social networking sites (hereafter SNS) such as Facebook, MySpace, YouTube or
Twitter (De Vries et al., 2012; Chang et al., 2013; Chauhan and Pillai, 2013; Ewing et al.,
2013; Fielder and Sarstedt, 2014).

There are two key reasons why social media are particularly suited to the investigation of
OBC participation. Firstly, social media foster relationships between individuals and
between individuals and brands (Hennig-Thurau et al., 2010). They allow users to connect
with friends and like-minded individuals and engage in asynchronous and synchronous
communication. Relationship building is also achieved through fanpages, social media
advertising and sponsored posts (Gironda and Korgaonkar, 2014). Unlike traditional
media, social media offer abundant scope for users to communicate back with the brand
(Chauhan and Pillai, 2013; Colliander et al., 2015) and embed brand-interactions in a
social network of relationships with other users (Trusov et al., 2009). As a result, social
media are increasingly considered as adequate tools to build consumer relationships with
and about a brand (Casaló et al., 2010).

Secondly, social media’s fundamental premise is User-Generated Content (UGC), whereby

content is produced and developed by different participants in a continuous and
collaborative manner (Kaplan and Haenlein, 2010). UGC, which achieved broad popularity
in 2005, is an umbrella term that contains various forms of media content that are publicly
available and created by end-users from text to video and audio materials (ibid, 2010).
UGC satisfies information and expressive needs, which are core to OBC participation
(Kaplan and Haenlein, 2010). By fostering UGC, social media give ground to rich OBCs
by satisfying consumers’ need to give and get information. Lastly, Christodoulides et al.
(2011) point out that UGC is the most recent manifestation of consumers’ desire to engage.

For these two reasons, and with the support of existing OBC research in this context, social
media are considered an appropriate environment for the investigation of consumer
engagement in OBC (Brodie et al., 2013; Chauhan and Pillai, 2013).

4.3.2. Facebook pages

Considering the range of social media platforms and environments, Facebook Pages are
chosen as the specific context of investigation of this study for several reasons. Firstly,
they qualify to the basic defining criteria of OBC: they are focused on a specific brand
 125

(Janh and Kunz, 2012), exhibit instances of consciousness of kind, moral responsibility
and shared values, and support relationships with other consumers and the brand
(Gummerus et al., 2012; Chang et al., 2013; Zaglia, 2013). Secondly, Facebook Pages are
chosen over other forms of social media branded account (such as branded Twitter profiles,
for instance) because they share closer similarities with the definition of OBC. Facebook
Pages have more clear-cut boundaries. Being a member of the Page requires a voluntary
choice of affiliation (Zaglia, 2013), which is a stronger commitment than ‘following’ a
branded profile on Twitter, for instance.

Facebook is a Social Networking Site 2 (SNS), which functions as a network of

interconnected individual profiles. It allows the creation of individual profiles and gives
the possibility to post content in the form of text, video or pictures (Raacke and Bonds-
Raacke, 2008). The business model of Facebook is based on certain levels of information
disclosure, which allows monetisation of the site. On Facebook, individuals connect with
friends or colleagues, but other forms of relationship affiliation also exist (Trusov et al.,
2009). In particular, individuals can connect with brands through the so-called ‘Pages’.
Indeed, brands, like any other member of Facebook, have the ability to create a profile,
called a ‘Page’. According to Facebook, ‘pages allow real organisations, businesses,
celebrities and brands to communicate broadly with people who like them. Pages may only
be created and managed by official representatives’ (Facebook, 2015a). Facebook Pages
are different from Facebook ‘Groups’, which can be created by anyone. In 2014, Facebook
has more than 30 million pages on which people interact. This makes it the biggest hub for
companies to create an official OBC (Chauhan and Pillai, 2013).

Research shows that Facebook pages are reflective and supportive of the consumer’s
relationship with a brand (de Vries et al., 2012; Malhotra et al., 2012). Pages allow brand
fans to share their enthusiasm about the brand and be united by their common interest in
the brand, and conversely provide a source of information and social benefits to the
members (Dholakia et al., 2004). On these brand fan pages, companies can create brand
posts containing anecdotes, photos, videos, or other material; brand fans can then interact
with these brand posts by liking or commenting on them (de Vries et al., 2012).

2
A Social Networking Site (SNS) is a type of social media (Kaplan and Haenlein, 2010) which has experienced extreme
success and growth in the last ten years (Blazevic et al., 2014).
 126

Twitter was also considered, as it is the second most popular SNS (Smith et al., 2012).
However, it was rejected on several grounds. Firstly, Facebook is more popular than
Twitter, therefore granting a bigger scope to this research. In 2015, Facebook is, with over
1 billion active users (Facebook, 2015b), the most popular SNS in existence and it still
dominates the market over competitors, such as Twitter, who has just over 284 million
active users (Statista, 2014). Moreover, according to the PEW research centre, 71 percent
of adult Internet users use Facebook while 23 percent of Internet users are on Twitter in
2014. Although Facebook’s overall growth has slowed and other sites continue to see
increases in usership, Facebook remains the most popular social media site (Pew Research
Centre, 2015).

4.4. Data collection

This section is concerned with the different stages involved in collecting data to answer the
research questions and test the related hypotheses. The design of the research instrument is
addressed first, followed by the method of administration and the pre-test and pilot phases.

4.4.1. Instrument design

The survey was designed using a rigorous process, involving a number of decisions
regarding the question content (measurement, wording and type of questions used) and
their sequence, as well as the physical form of the questionnaire (Churchill, 1999). These
steps were then re-examined in order to ensure coherence and consistency of the final
instrument.

The content of the questions is largely addressed in the measurement sections detailed
further on. That section deals with sources of the existing scales that are used and explains
the development of the consumer engagement scales. Both in de adaptation of existing
scales and development of the new ones, care was given to having purposeful, precise and
complete questions (Czaja and Blair, 2005). Churchill (1999) and Bryman (2008) provide a
checklist of question utility criteria which guided the researcher’s assessment of each item
content. Following these guidelines, it was ensured that each question measured some
 127

aspect of the research questions, which all required information was gathered and that
wording was clear enough for all respondents to understand the questions in the same way.

In terms of the response strategy, multiple-choice, closed questions were used in the
survey. This type of response strategy is considered the most appropriate for self-
administered surveys (Czaja and Blair, 2005). They ensure ease of information recoding,
save analysis time, and ensure response format homogeneity (Buckingham and Saunders,
2004). More specifically, seven-point Likert scales were used to tap into the strength and
direction of respondent’s attitudes and behaviours. Strictly speaking, Likert scales are
ordinal scales. There is however a widespread tendency to use Likert scale for interval
based techniques such as factor analyses or structural equation modelling (DeVellis, 2012),
based on the premise that the psychological distance on a Likert scale, if not equal, are
extremely close (Kennedy et al., 1996). This treatment of the Likert scale as an interval
scale is common (Buckingham and Saunders, 2004) and is considered best for self-
administered and online surveys (Hair et al., 2006).

Seven-point scales were used for several reasons. Firstly, this amount of points is required
in order to perform a successful factor analysis (Malhotra and Birks, 2006). Secondly,
scales with an uneven number and thus providing a neutral option in the middle are
preferable. Respondents might not feel strongly about an issue (Czaja and Blair, 2005) and
might otherwise feel that they are forced to take a stance (Cox, 1980). Finally, seven points
seem the appropriate number, as scales with a larger amount of point do not tend to
improve reliability or validity (Dawes, 2008).

An ‘agree’ scale was used following Dillman (2000), anchoring it with 1 equals
‘completely disagree’ and 7 equals ‘completely agree’. Only end-points were labelled to
elicit a finer variance of responses (Czaja and Blair, 2005).

In addition to a majority of Likert scales, a semantic differential scale is used as well to tap
into the construct of attitude toward OBC participation, as detailed in the measurement
section. Providing a different type of response strategy on some answers is also a way to
ensure continued attention of the respondent and avoid automated responses that a single
response strategy can elicit.

The next aspect of instrument regards the questions wording. In line with leading research
on survey design (Churchill, 1999; Buckingham and Saunders, 2004; Czaja and Blair,
 128

2005; Bryman, 2008), great care was taken to avoid the general pitfalls of question
wording: double-barrelled or ambiguous questions, use of jargon, long items, leading or
too general questions. In a limited amount of instances, existing scales exhibited tangential
wording, which could be ambiguous or double-barrelled (e.g. ‘In general, I thoroughly
enjoy exchanging ideas with others online’). Due to reliance on existing scales, the risk of
such problems was minimised. Throughout the questionnaire, simple words were used,
bearing in mind that the smallest common denominator in terms of respondent profile was
an 18 year old person. The researcher also avoided theoretical jargon (e.g. ‘brand
community’) to use context- and user-relevant wording (e.g. ‘page’). These considerations
were further taken care of in the pre-test and pilot phases.

Reverse-coding items were avoided as much as possible. Their advantage is to allow the
avoidance of response sets, which is a common issue with Likert scales (Bryman, 2008).
However, experience shows that negatively-worded questions confuse the respondent
(Buckingham and Saunders, 2004), who then do not tend to answer in line with their
overall attitude. Overall, it is considered that the disadvantages of items worded in an
opposite directions outweigh any benefits (DeVellis, 2012).

Another major element of the question wording was the translation of the English
questionnaire in French to allow validation of the study. In this process, translation
equivalence had to be ensured to control as much as possible for this form of instrument
and item bias (Douglas and Craig, 2006). Translation equivalence means that the translated
questions convey the same meaning across cultures. In order to achieve translation
equivalence, the questionnaire was translated using iterative team-based translation
principles in line with existing guidelines (van de Vijver and Tanzer, 2004; Douglas and
Craig, 2006). The questionnaire was first translated from English to French by the lead
researcher of this study, who is bilingual. This version was then submitted to a bilingual
management researcher (French living in the UK for 15 years), a bilingual marketing
professional (Belgian working in English), a linguist (Belgian English teacher with 25
years’ experience), a bilingual student (Irish student fluent in French) and a bilingual
Scottish journalist. Each individual reviewed the translation and fed back to the lead
researcher with comments and modifications to the translation, who then made changes
and fed them back to the team. An iterative procedure was used whereby the lead
researcher and the translation team collaborated until all parties agreed upon a satisfactory
version (Craig and Douglas, 2006).
 129

Using this procedure, item translation, ambiguity in the original item wording,
appropriateness and familiarity of item wording or connotations associated with item
wording could be accounted for, in line with van de Vijver and Tanzer’s (2004) most
common sources of item bias. For instance, it was noted by one of the translators that
‘interagir’ was a poor French translation of ‘interact’ that would likely confuse the
respondent. It was replaced by ‘participer activement’, which reflect active participation.
Another translator commented on the meaning in French of terms such as ‘transported’,
‘punishing’ and ‘to detach oneself’. Although these terms make sense in English, their
literal French translation is seldom used, calling for a more common way to phrase them,
which was then decided based on the translator’s suggestions. On the other hand, other
elements of the questionnaire, such as the reference to a Facebook ‘fan’ or ‘page’ were
kept in English in the French version of the questionnaire despite their linguistic
inaccuracy, because this is the wording used on Facebook in both languages. Translating
these words literally in the French questionnaire would have made no sense, and instead
confuse the respondent. Based on these considerations, the final version of the French
questionnaire was agreed upon by all members of the translating team.

The sequence of the questions was also subject to careful consideration. The survey
started with a screening question ensuring that the respondents were all above 18 years of
age, to comply with the University of Glasgow’s ethics requirements. An introductory
statement then explained the purpose and content of the survey (Bryman, 2008), as well as
a link to the Plain Language Statement. The sequence of the survey then followed general
rules of survey structure, making sure that earlier items were simple (Baker, 2003),
immediately associated with the subject matter (Dillman, 2000) and going from general to
specific, and simple to more complex. The survey was broken down into 5 main parts,
ensuring that transition between each section was clearly explained and logical; to facilitate
the respondent’s mental processing. The instrument opened with simple questions about
Facebook overall participation. After asking the respondent to communicate which brand
page they were answering about, all brand-related items were introduced, including brand
engagement, followed by all category-related items. All community-related questions
constituted the next part, covering brand-community engagement. The survey concluded
with demographic questions. Placing ego-involving questions as the end is a common
practice to avoid the participant feeling threatened (Breugelmans, 2008).
 130

As response rate is a major concern for online surveys, particular care was taken to make
the survey experience as positive as possible. Previous researchers (e.g. Oppenheim, 2000)
have suggested a number of techniques to increase response rate, some of which are used
in this research. The physical aspect of the questionnaire plays an important role in
motivating respondents to participate and finish the survey (Churchill, 1999). Bearing this
in mind, the survey was designed to be limited to eight full pages, however with breaks
between sections, numbered pages, and arrows for skip instructions, as recommended by
Dillman (2000). A progress bar was also visible on each page, in order to reduce the drop-
out rate (Couper et al., 2001). An enticing and clear introductory paragraph was written,
which detailed the purpose of the study, content of the questionnaire, reason for selection
of the respondent, and confidentiality approach. Instructions throughout the questionnaire
were as clear as possible (e.g. ‘Indicate on a scale from 1 (completely disagree) to 7
(completely agree) to which extent you agree with the following statement’) (Bryman,
2008; Das et al., 2011). The visual aspect of response options was also taken care of. Likert
scales were presented in an item by scale point table, with radio buttons. Numeric scales
with a large amount of possible responses (e.g. year of birth, country of residence, daily
amount of time spent online) used a drop-down menu. Those with a small amount of
possible responses (five and below) were presented with horizontal radio buttons. Overall,
the layout of the questionnaire was conservative, in the shades of blue and time new roman
font, because it is considered more professional (Oppenheim, 2000).

An incentive was also given to participants, in the form of three £100 Amazon vouchers, to
be won through a lucky draw at the end of the data collection. The researcher felt it was
necessary to give a big incentive given the length of the questionnaire (up to 20 minutes if
done properly), and incentives are known to improve response rates (Churchill, 1999).

Contacting page administrators was achieved with a clean, professional and interesting
cover letter, as shown in Appendix 1. Affiliation with the University was stressed, and a
condensed version of the results was offered upon request. Follow-up was done within one
week of non-response; however it did not prove to successfully enhance the response rate
of page administrators. This was mostly due to the fact that most of them could not post
third-party information on the page as a rule, and no amount of convincing could change
their internal policy.
 131

Lastly, an iterative approach was taken to creating the survey instrument, as the first draft
of a questionnaire seldom leads straight to the final instrument (Churchill, 1999).
Numerous rounds of revisions occurred, based on re-examination by the researcher,
comments from supervisors, and reviews from colleagues. These rounds of review allowed
identifying major problems in the sequence, wording and usefulness of questions. The
major steps subsequently taken to evaluate the quality of the drafted instrument are
detailed in the pre-test and pilot section.

4.4.2. Administration

A self-completion online survey was used to collect data. More specifically, a web-based
and communication-based method was used (Bryman, 2008). ‘Web-based’ signifies that
the data were collected through the Internet with the use of an online questionnaire created
on and supported by the surveymonkey.com application. Communication-based approach
means that an online communication medium was used as the place where the
questionnaire could be accessed. In this case, the questionnaire was accessible through the
sampled Facebook pages (see sampling section below). Data collection was thus
asynchronous, as the respondents had the freedom to answer the self-completion
questionnaire whenever they chose to.

The choice of an online self-administered survey was motivated by several reasons. Firstly,
survey is considered the most appropriate tool to obtain data to test hypotheses (Baker,
2001). Matching the method to the research problem, and not vice versa, was the rationale
(Creswell, 2003). Surveys are particularly appropriate to the investigation of concepts,
theory testing, and analysing relationships (Klassen and Jacobs, 2001). Moreover, key
advantages of surveys, and online surveys in particular, are their ability to accommodate
large sample size at relatively low cost, whilst tapping into factors that are more rapid, and
allow increased design options (Das et al. 2011). Online surveys have no geographic
boundaries (Dillman and Bowker, 2001), which makes them particularly suited to the
investigation of such large-scale phenomena as consumer engagement on OBC.
Additionally, although self-administered surveys involve the researcher losing control over
the process, and potential lack of truthfulness of the respondent, it also implies lowering
confidentiality issues for them as they have the control, and therefore balance the
truthfulness issue.
 132

The biggest issue related to self-administered online surveys is the high, and highly
variable non-response rate, with a mean of 32 percent and a standard deviation of 19
percent (Cook et al., 2000), which are respectively much lower, and more variable than
other methods. Coverage can also be an issue when the population of interest does not have
Internet access (Couper, 2000), however this issue is eliminated from this study given the
inherent online profile of the population. Response errors can also appear due to
misunderstanding of the questions (Churchill, 1999). In general, survey methodologies are
associated with a number of sampling and non-sampling errors (Hair et al., 2006), and
sample bias is particularly common for online samples (Hewson et al., 2003). These issues
are tackled in the sampling section below.

The online survey was generated using the application ‘Surveymonkey’. Surveymonkey
was chosen because it provides high flexibility in questionnaire design (scale types, format
and layout, skip logics, filter questions, etc.), administration (custom administration link)
and data retrieval formats (Das et al., 2011).

In line with the sampling procedure detailed further, the link to the survey was
communicated to the administrators of the Facebook pages, who were in charge of posting
it on the platforms, with a word of explanation. A standard post was suggested, but
freedom was left to the page manager to post what they thought was best suited to their
community. Subsequently, the post appeared on the Timeline of the Facebook page,
allowing its members to view it whenever they clicked on the page. Moreover, the
newsfeed feature allows posts by followed page to appear on the user’s homepage
whenever they connect onto their Facebook account. Facebook uses an algorithm to show
relevant posts on the newsfeed, rendering it impossible to have control over the audience
of the post. Once the page members clicked on the survey link, they were redirected to the
survey and assumed answering it.

This way of contacting the respondents ensured that only people with prior brand page
experience were approached. Recent brand page experience was a pre-requisite to have the
required level of knowledge and memory to answer the questionnaire (Churchill, 1999;
Bryman, 2008). These criteria were ensured, thanks to the specific format of survey
administration method adopted.

Lastly, administration bias was also accounted for when collecting data on the French-
speaking sample. Methodological issues can arise in this process if there are differences in
 133

the ways participants are recruited, in the environmental administration conditions or any
ambiguity in the instructions given by the researcher (van de Vijver and Tanzer, 2004). A
careful replication of the administration procedures from the English to the French-
speaking survey ensured that such problems could not arise. The same piece of online data-
collection software was used, it was posted in the same way on the Facebook pages by the
administrators, and the same detailed instructions and incentive were given to the French-
speaking participants. Following the translation equivalence procedure detailed above, all
communication with the French-speaking page administrator was also carefully monitored
for equivalence. Lastly, temporal equivalence was also ensured as the two samples were
collected simultaneously, which is particularly relevant when dealing with consumer
phenomena subject to evolutionary change (Garnier and McDonald, 2009).

4.4.3. Pre-test and pilot

Even after careful crafting of the instrument, the researcher can still miss important issues,
which pre-test and pilot studies can help detect (Czaja and Blair, 2005). From a scale
development perspective, psychometricians advocate that once an item pool has been
judged, modified and or trimmed by experts, pilot testing on a larger sample (n = 100-200)
from a relevant population is in order (Haynes et al., 1995).

Pre-testing was first carried out on a small sample of friends and colleagues in order to
judge how long it took and identify obvious problems in wording, instructions or question
sequence (Buckingham and Saunders, 2004). The respondents were selected due to their
expertise in either marketing, linguistics, survey design, or a combination of the three, as it
is suggested that ‘expert’ should be used as pre-test respondents (Diamantopoulos et al.,
1994). One marketing student, one marketing manager, one journalist and four researchers
in total read the questionnaire, answered it, and provided written and verbal feedback to the
researcher. They helped identify complex wording and language issues in the questions
and instructions and reorganise the order of the sections into a more logic way. For
instance, the survey was initially asking questions about the community right after the
respondent had identified the brand, which they were answering about. It was pointed out
that it made more sense to keep the brand-related questions right after eliciting the brand
name, as the brand was then prominent in the respondent’s mind. The brand-related section
was thus moved forward, keeping the community-related questions at the end. Advice on
how to word a motivating foreword was also given. It is also important to point out that the
whole questionnaire was pretested, not just the newly developed items, as borrowed items
 134

also require review (Czaja and Blair, 2005). The pre-test questionnaire was administered
using the final data collection application, surveymonkey, so the participants could also
experience the real survey conditions.

The pre-test was followed by a pilot study on a larger number of respondents. The aim of a
pilot study is, not only to detect problems in the questionnaire wording or sequencing, but
also to ensure that the questionnaire is generating a range of answers, indicating variance
for the final stage (Buckingham and Saunders, 2004). Secondly, pilot studies allow making
an initial assessment of internal consistency, means, variances, inter-item correlations and
factor structure (Netemeyer et al., 2003).

There is no set answer to the appropriate number of pilot respondents: it can vary from half
a dozen to a hundred, or above (Tull and Hawkins, 1987). DeVellis (2012) as many as 300
responses depending on the number of items, while Clark and Watson (1995) suggest that
n=100 to 200 is reasonable. Within the scope of this survey, a total of 101 pilot responses
were considered sufficient given the number of items.

A second consideration is the sample composition. Authors contend that the pilot
population needs to be similar to the respondents in the actual study (Churchill, 1999), and
that convenience sample of students are acceptable (Netemeyer et al., 2003). In this study,
a sample of undergraduate and postgraduate university students was used, which qualifies
to the basic requirements of pilot sampling. Moreover, given that half of the Facebook
population is between 18 and 34 years of age (Statista, 2015a), the researcher felt that
students in this age range was an appropriate sample for the pilot and broadly
representative of the final population.

Regarding administration of the questionnaire, it is appreciated that the pilot study should
mimic the survey conditions as much as possible, i.e. use the same instrument and
administration method (Czaja and Blair, 2005). In order to do so, the survey was promoted
online, using the same surveymonkey tool as the final questionnaire. It was however not
posted on a Facebook page. Given the nature of the pilot population, the best way to reach
them was by posting the survey on the university’s online learning platform, with a short
text inviting them to take part in the study, on a completely voluntary basis, yet entirely
unrelated to their university assessment or performance.
 135

As a result, the pre-test and pilot highlighted a number of modifications to be made to the
questionnaire prior to its large-scale dissemination. Firstly, it provided indications as per
the sampling of Facebook pages. Liberty was given to the respondent to choose the brand
and type of Facebook group they wanted to refer to (official pages or consumer groups).
Firstly, it became apparent that Fashion and beauty was to be included in the list of brand
categories to target. Moreover, it appeared that large-size pages were top-of-mind for the
respondents, highlighting the need to represent them in the pages targeted. Moreover,
confirmation of the theoretical focus on company-managed pages was given because
respondents selected them in majority over consumer-governed groups.

Secondly, the pilot and pre-test gave insight into the wording of the items, particularly for
the newly-developed consumer engagement items. Redundancies were pointed out as well
as unclear meaning of some words. It also appeared that if the OBC engagement and online
brand engagement were placed back to back, the respondents could not differentiate them;
they were therefore separated in the final questionnaire.

Lastly, a simple statistical analysis of the composition of answers revealed that all items
and constructs exhibited good internal consistency scores and, overall, normality.
Purification of the consumer engagement item also appeared as necessary in future stages,
as some items exhibited skew or high construct-level Cronbach’s Alpha, which can be
understood as an indication of item redundancy. These are dealt with in the scale
development section.

Following both the pre-test and pilot phases, the instrument for the large-scale survey was
finalised. It can be found in Appendix 2 in English and in Appendix 3 in French.

4.5. Sample design

4.5.1. Sampling approach

A number of sampling issues are inherent to OBC research and the application of
traditional data collection methods like surveys to OBC settings represent certain
challenges. This is particularly true for forms of OBC like official Facebook pages. For
example, these OBC members fall into the category of ‘unique populations’ which are only
 136

accessible through dedicated channels – the OBC, and in certain ways (Wright, 2005). In a
sense, OBC members are akin to what is defined as a ‘hidden population’ because these
individuals are hard to reach through traditional means (Salganik and Heckathorn, 2004).
Although members’ identity is revealed to the page administrators, there is no publicly
disclosed list of the members of an official Facebook page, primarily for privacy concerns.

Additionally, determining the actual sample size when sampling OBC members can be
difficult. As Preece and Maloney-Krichmar (2005) explain, online community populations
have unclear boundaries, often due to content visibility settings, making sampling tricky
and prone to errors. Although official Facebook pages exhibit the number of ‘Likes’ that
they have, allowing knowing the number of registered members on the page, this does not
imply that all members are active on the page and that they will see page posts. Active
participation to the community can be extremely sporadic (Preece et al., 2004), to the point
that members never see content posted by the brand. This can be the case when a page
member never voluntarily visits the page and has unsubscribed from page updates.

In view of theses sampling issues intrinsic to OBC research, a two-stage, non-probabilistic

sampling design is used (Churchill, 1999). The primary sampling unit is not the unit of
population to be studied but groupings of those units. In this study, the cluster is the
official Facebook page in its capacity of OBC. The sampling approach used for this layer is
purposive (Bryman, 2008). The second sampling unit, which constitutes the unit of
analysis of this study, is the individual page member, or consumer. Due to the challenges
posed by the nature of OBC, sampling for this layer is a convenience, self-selected sample,
in line with existing research in this field (Casaló et al., 2010; Breitshold et al., 2015). This
section details this two-stage sample design.

4.5.2. Stage 1: Sampling of the Facebook pages

The first stage of the sampling approach consists in focusing on the Facebook pages as a
sample unit. The selection of the sampling frame is purposive to the study (Kozinets, 1999)
and follows a number of inclusion and exclusion criteria. In line with the defining
characteristics of OBCs highlighted in Chapter 2, this study focuses on official Facebook
pages managed by the brand, of diverse sizes and duration, and exhibiting specific forms of
relationship structure on the online environment, as detailed in Table 12.
 137

Table 11: Facebook pages characteristics

Criterion Characteristics
Governance In official Facebook pages, governance is always assumed by the
brand itself, or its representative (if for instance, social media accounts
are managed externally by an agency). Facebook groups, in contrast,
can be managed by consumers (Zaglia, 2013). Having a brand-
governed OBC as a context of investigation was important to ensure
that brands were present as relationship partners because, in consumer-
lead OBCs, brand might be entirely absent of the community.

Marketplace This study focuses on brand communities supporting the marketplace.

orientation Since Facebook pages are created and managed by the brand, they aim
to support rather than counter market dynamics. Although consumers
might be able to express their more negative views on Facebook pages,
these are often actively managed and dealt with by the brand.

Size Facebook supports pages of any size, from one fan at their inception to
over 5 million fans like the ‘Facebook for every phone’ page (Statista,
2015b) and this study does not discriminate pages based on their size.

Duration With the existence of their pages feature since 2007, Facebook allows
these communities to exist on a relatively stable basis. This study does
not consider OBC longevity as a selection criterion.

Relationship On Facebook pages, relationships are akin to those of pools or hubs

types (Fournier and Lee, 2009) according to a study evidencing the
differences between Facebook pages and Facebook groups (Zaglia,
2013). Members of pages feel more connected to the community as a
whole rather than to individual others and the common focus of the
page is core to this relationship (ibid., 2013).

Space Facebook pages are by nature online forms of OBCs. However, they
often promote real-life activities organised or sponsored by the brand,
therefore bridging offline and online participation.

Product category was also a selection criterion for the sampling of Facebook pages; not as
a way to restrict the type of pages included, but as a way to ensure that all relevant
categories were covered. At the time of the data collection (March-June 2014) Facebook
hosted around 30 million pages (Facebook, 2014), which constitutes the population of
interest of this first sampling stage. When creating a Facebook page, the page creator has
the choice to identify it as matching one of six different page types:

1. Local business or place 4. Artist, band or public figure

2. Company, organisation or institution 5. Entertainment
3. Brand or product 6. Cause or community
 138

When deciding to create a ‘Brand or Product’ page, a further classification based on 33

different types of products or brands is possible. Sub-categories are also available for the
other five categories of pages. Although these categories tap into a wide range of possible
brands and businesses, they are not mutually exclusive. For instance, a video game brand,
for instance, could equally belong to the ‘Entertainment’ or the ‘Brand or product’
categories.

Other classifications of Facebook pages exist, such as the one provided by Social Bakers, a
leading social media insight company. They divide Facebook brand types into 22 industry
categories, ranging from beauty to gambling or retail. Social Bakers (2014) identifies top
performing brands on Facebook in 2014 as largely belonging to the food and beverage
industry (#1 Coca-Cola, #2 McDonalds, #3 RedBull, #7 Oreo). The fashion industry is also
well represented in top performers with Converse (#4) and fastest growing pages like
Billabong. The entertainment sector also performs well (#6 Playstation, #8 Nike Football).
Based on this analysis, a first criterion to determine the sampling frame of the Facebook
pages is proposed in the form of relevant brand categories to investigate.

The following categories are selected in this study:

1. Food and beverage

2. Technology (software, telecom, computer products…)
3. Services (bank, insurance, education…)
4. Travel (airline, railways, travel agents…)
5. Fashion and beauty
6. Durable goods (automobile, electronics, home appliances)
7. Retail (stores, supermarket, e-shops)
8. Entertainment (sports, films, series, books, games…)
This categorisation is deemed representative of the variety of the Facebook pages types
and relevant to the context of investigation (Kozinets, 1999). In order to be included in the
sample frame, a brand therefore had to fit into one of these categories. Brand categories
such as celebrities or public figures were left out of the sample due to the ambiguity of
defining them as brands and the difficulty to tag them as a product that you can purchase,
which could have caused confusion for the respondents.

The reason for seeking variety in terms of brand categories rather than focusing on a
limited number of categories or brands is driven by the existing nature of OBC on the one
hand, and consumer engagement research on the other hand. So far, OBC research has
 139

tended to focus mainly on a limited amount of brand types, often researching high
involvement goods such as cars and motorbikes (Algesheimer et al., 2005; Füller et al.,
2008; Ewing et al., 2013), technology products (Wiertz and de Ruyter, 2007; Ganley and
Lampe, 2009; Faraj and Johnson, 2011), luxury good (Kim and Ko, 2011) or services
(Carlson et al., 2008). This was evidenced in Table 10 at the end of Chapter 3.

The treatment of low involvement goods has been much more limited, although repeatedly
called for (Cova and Pace, 2006; Veloutsou and Moutinho, 2009). So far, only one study
has incorporated a wide range of Facebook Pages representing different product categories
with a large-scale sample. Janh and Kunz (2012) collected data from 532 respondents
spanning over 40 different pages of goods, service, organisation, or celebrity brands. They
call for further research across Facebook brand categories.

The consumer engagement literature on the other hand, has placed much focus on the study
of services such as sport centres or public transportation services (e.g. Brodie et al., 2013;
Jakkola and Alexander, 2014). Consumer engagement scholars however acknowledge that
it is at this stage highly unclear which type of products or services are more likely to be
generative of high levels of engagement (Vivek et al., 2014). Furthermore, sampling a
variety of brand categories also ensures variance in terms of product involvement, which is
hypothesised as a key driver of consumer engagement. For all these reasons, and similarly
to Schau et al. (2009), this study is hence inclusive of a wide array of different brand
categories.

Another criterion for considering a page adequate is its commercial nature. As this study
aims to investigate brand loyalty in the form of purchase behaviours as the outcome of
consumer engagement, it was a requirement for the brands to be commercial rather than
non-profit oriented. Cause and communities were therefore excluded of the sampling frame
as well.

Industry ranking were also used for the inclusion of pages in the sample frame, focusing on
top Facebook pages in terms of engagement rate, as defined by the industry. Rankings such
as Social Baker’s (2014) top 12 pages for engagement rate were used.

Based on this list of criteria, a total of 326 Facebook pages were listed and contacted
between February and June 2014, using the data collection procedure detailed earlier in
this chapter. These pages represent a total of over 181,000,000 members. After sampling
 140

the Facebook pages, their administrators were contacted, asking them to post the link to the
survey on their page to allow their members to take part in it.

The full list of pages that participated in the survey can be consulted in Appendix 4, as well
as some screenshots of some of these participating pages in Appendix 5. This leads to the
consideration of the second level of sampling of this study: the sampling of individual page
members.

4.5.3. Stage 2: Sampling of the consumers

Theoretically, the population of interest, or consumers, as defined in this study, is

constituted by the totality of the page members. In March 2014, Facebook counted about 1
billion active monthly users (Facebook, 2014), 40 percent of which did not follow any
pages (Smith, 2014), which suggested a total number for page members of 600 million.
Based on the first sampling stage of the study, the population of interest is however
restricted to the total population of the pages that accepted to post the link to the survey on
their page, which amounts to 181,000,000 individuals at the time of the survey.

The application of random sampling relies on the existence of a sampling frame and
determining a sampling frame seems impossible due to the nature of the Facebook pages
(Bryman, 2008). As Wright (2005) points out, attempting to establish a sample frame by
counting the number of participants in an OBC, or the published number of members tends
to be highly inaccurate due to the ebb and flow of communication in OBC. Participation in
an OBC can be sporadic at best, depending on the characteristics of the groups and its
members. Some people are ‘regulars,’ who may make daily contributions to discussions,
while others only participate intermittently. Furthermore, ‘lurkers,’ or individuals who read
posts but do not post content, may complete an online survey even though they are not
visible to the rest of the community (Preece et al., 2004). Moreover, due Facebook
visibility algorithms and individual setup preferences, once the survey was out in the open
(i.e. posted by the page manager on the page), it was hard to control who came across it
and how, leading to a self-selected, voluntary sampling design (Bryman, 2008). Facebook
(2015) estimates that the average organic reach of a post on a brand page is of 16 percent,
meaning that 16 percent of the page population will see a post. Furthermore, the average
click-through-rate of a Facebook post is known to be 2 percent (of the people who see it),
according to Salesforce.com (2013). On this basis, an expected rate of persons who started
the questionnaire of 0.3 percent can be expected. However, this concerns content that the
 141

brand would normally post as part of their content strategy, not third-party consumer
surveys. With a number of 1690 started questionnaires, it can be assumed that 0.5 percent
of the people who saw the questionnaire started replying to it.

Given the lack of sampling frame, an adequate sample size could not be determined on this
basis. However, the rule of thumb to have at a participant to item ratio of 5:1 served as a
basis to determine the required amount of responses (Nunnally and Bernstein, 1967). The
total survey, before psychometric analyses begin, measures 84 items, excluding
demographics and other non-construct items, therefore a validation sample of 420
questionnaires was required. As for the calibration sample, based on a pool of 35 items of
OBC engagement, mirrored with a pool of 35 items of online brand engagement, a
calibration sample of 350 (35x2x5) responses is required. However, Nunnally and
Bernstein (1967) suggest that a sample of 300 is an adequate number for scale calibrations.
Additionally, the set of 70 items is technically based on a replication of an initial set of 35
items, suggesting that a sample of 175 respondents would be adequate. The calibration
sample amounted to 224 respondents and the validation sample 497 (224 English and 273
French). The procedure to reach this final amount of answers, as well as the issue of non-
response bias are addressed in the next section.

4.6. Sample treatment and characteristics

Prior to analysing the sample characteristics and following-up with data analysis in the
next chapter, missing data and non-response need to be dealt with (Diamantopoulos and
Siguaw, 2000). By the closing date of the survey, 1690 responses had been received,
including 989 in English and 701 in French. A large amount of missing data was detected
in these replies, which called for a missing data analysis and treatment.

4.6.1. Missing data analysis

Understanding the nature and patterns of missing data is the first step in dealing with them.
Data can be MCAR (missing completely at random), MAR (missing at random) and
NMAR (not missing at random). Using SPSS, univariate statistics, separate variance t-tests
and patterns of missing data are computed according to Tabachnik and Fidell (2000).
Little’s test is also performed in order to assess the type of missing data at hand, providing
 142

a p-value of 0. This low p-value indicates an ability to reject the null hypothesis, meaning
that the data are NMCAR; in other words, there is a pattern in the missing data. Analysing
the frequency of missing data and tabulated cases showed that most of the missing data is
grouped toward the end of the questionnaire but no other pattern of missingness is
detected. The frequency of missing data for each item is presented in Appendix 6. As
evidenced by the amount of missing cases on the last few items of the table (which
represents the questionnaire question order), missing data went up to 58 percent of the
responses for some items situated at the end.

Several factors can explain the structure and high levels of data missingness in the sample.
For example, it is well accepted that self-administered online surveys tend to provide
higher dropout rates than face-to-face surveys, as the researcher cannot ensure the survey
completion (Cook et al., 2000). Moreover, the length of the questionnaire is likely to
induce high levels of respondent fatigue, increasing the dropout rate. Lastly, the sensitive
nature of the last set of demographic questions at the end of the questionnaire has probably
led some respondents to avoid answering them altogether.

On the basis of this structure of missingness, dealing with missing data is done in a two-
step basis. Listwise deletion is first applied to the data. A cut-off percentage of allowed
missing data per case is fixed at 15 percent, which is higher than the 10 percent advocated
by Hair et al. (1998), but driven by the fact that a large amount of data is missing on
demographic variables, which are not involved in hypothesis testing or scale development.

Deletion of these cases resulted in a remaining 721 cases (448 English and 273 French).
These remaining cases are further dealt with through the EM (Expectation Maximisation)
method performed through the SPSS Missing Value Analysis function. Although
imputation can be complex with NMCAR data, this method is appropriate because: 1) it
permits specification of distributions other than normal (Tabachnik and Fidell, 2000),
which is potentially an issue with some of the variables; 2) it is the method that produces
the less bias on NMCAR data (Little and Rubin, 1989) and 3) it performs better than
methods like series mean or regression imputation which tend to reduce the variance of the
data (Byrne, 2010). Lastly, some of the sensitive data that could not be imputed due to
their nominal nature (e.g. nationality and country of residence) are marked as N/A.

The final sample of 721 valid responses after missing data analysis evidences an overall
dropout rate of 58 percent. This figure is 55 percent for the English sample and 61 percent
 143

for the French sample. Although these values are high, they are not surprising given the
online context of survey administration (see above in the data administration section) the
self-selected, voluntary nature of the sample, and the length of the questionnaire.

4.6.2. Non response bias

Dealing with non-response bias is particularly important in a context where the sample
frame could not be determined and evidencing the lack of bias adds robustness to the
sampling procedure. There are different ways to deal with non-response bias (Armstrong
and Overton, 1977). The first one is the extrapolation procedure, which assumes that late
respondents are similar to theoretical non-respondents. To compare early and late
respondents, a cut-off date of May 1st 2014 was selected to separate the respondents in
early and late responses. This was chosen because most of the data were collected before
May, when a second wave of sampling started (580 answers before May, 141 in May and
June). T-tests and chi-square tests were used to compare early and late respondents on their
characteristics and no significant differences were found between the two groups on the
sample characteristics measures. Caution in the interpretation of these results is however
required: given the method of delivery, a gap could occur between the moment when the
survey was posted and when respondents answered it.

4.6.3. Common method bias

The study used a single informant (i.e. the consumer) to measure both independent and
dependent variables and therefore controlled for common method variance bias using a
range of procedures (Campbell and Fiske, 1959). First, the items were formulated as
clearly, concisely, and specifically as possible. They were either based on previously
validated scales or the output of a strict scale development procedure. Second, a computer-
administered questionnaire was developped, with the aim to reduce social desirability
biases (Podsakoff et al., 2003). In addition, the questionnaire introduction indicated that the
questionnaire was specifically about the consumer’s own experience on OBC, suggesting
that there was no right or wrong answers. Third, the design of the web-based survey
instrument made it impossible for respondents to retrieve their answers to earlier questions.
Therefore, it was more difficult for them to maintain artificial consistency between
answers or search for patterns in the questions, which helped control both for the
consistency motif and social desirability biases (Podsakoff et al., 2003). Additionally,
although Likert scales were majorly used, a semantic differential scale was also present as
 144

well as some multiple choice-question, to reduce common scale format issues. Lastly, we
tested for common method bias using Harman’s one-factor test (Podsakoff and Organ,
1986) with unrotated factor solution to determine the amount of factor necessary to
account for the variance in the variables. A principal component factor analysis of the
dependent and independent variables yielded six factors with Eigenvalues higher than 1.0,
and the first factor explained 42% of the total variance. The test showed the absence of one
major factor (Podsakoff and Organ, 1986); thus common method bias is not a serious
problem in our data.

4.6.4. Sample characteristics

As stated above, the total usable sample is composed of 721 responses, which represent
273 French-speaking respondents and 448 English-speaking respondents.

French (FR) and English (EN) survey respondents’ ages range between 18 to 82 years old,
with a mode of 27 years old and a median of 29 years old, based on the reported birth
years. This is congruent with the age split of Facebook users in 2014, with 19 percent of
users reported to be between 25 and 34 years old, and 9 percent of them above 65 years old
(Statista, 2014). The overall gender split is 49 percent of male and 51 percent of female.
The sample is skewed in favour of male respondents compared to the actual Facebook
population, which reports slightly more female users (Business Insider, 2014).

The sample is overall highly educated, with about half of all respondents across languages
having a postgraduate degree. These results are in line with the educational level of the
Facebook population as reported by the Pew Research Center (2015), which shows that a
majority of the Internet users who are on Facebook have a college degree or above.

Respondents of the survey span a total of 75 countries across French and English samples.
It is also important to specify that the difference between the two cultures lies in the
language, rather than the countries of origin of the respondents, although there is a link
between the two. English-speaking respondents are largely residing in the UK (28 percent),
US (9 percent) and Ireland (8 percent), whereas respondents of the French survey reside
primarily in Belgium (89 percent) and some of them in France (3 percent). However, for
both languages, other countries or residence are represented as well, since a French-
speaker might very well live in Spain, for instance. This feature is inherent to the nature of
 145

the targeted Facebook pages, which are set up in a specific language, but do not restrict
member access to specific countries.

Overall, the results in Appendix 7 show that sampling equivalence exists between the
French and English-speaking samples, which evidences their compatibility. Indeed,
although it is impossible to fully eliminate them, ‘nuisance factors’ need to be controlled
for; hence a required similarity across samples in terms of their characteristics, other than
the target constructs (van de Vijver and Tanzer, 2004).

A wide range of brand categories are represented in the study, with Food and Beverage
achieving the highest overall score (33 percent across English and French studies),
followed by Travel (21 percent), Fashion and Beauty (14 percent) and Entertainment (12
percent). These brand categories match Facebook’s current best performing categories.
Between 85 percent (EN) and 82 percent (FR) of the respondents are purchasing clients of
the brand they follow at the time of the study, evidencing that online brand usage and
membership does not imply prior buying behaviour (Ruiz-Mafé and Blas, 2006). This is in
line with the theoretical stance taken in this study regarding the definition of consumers.

Facebook activity is assessed using three metrics: membership duration, number of

platform visits per day and time spent on the platform per day. These metrics are replicated
at the brand page level, focusing on page membership duration, number of active visits of
the page per week and average time spent on the page per week. Most of the respondents
joined Facebook between 2007 (EN: 20 percent; FR: 19 percent), 2008 (EN: 23 percent;
FR: 48 percent) and 2009 (EN: 17 percent; FR: 12 percent). They are highly connected to
Facebook, as on average, 35 percent of them (EN: 34; FR: 36) receive notification from
Facebook on their phone in real time, while others log on from one to above 6 times a day,
leaving only 4 percent of the sample to connect less than once a day. In terms of time spent
on Facebook, while a small seven percent reports spending less than 10 minutes on
Facebook per day (EN: 8; FR: 5); the split between people spending 10 to 30 minutes, 31
to 60 minutes or over 60 minutes on Facebook per day is relatively equal.

Page membership duration is reported to be 1 to 5 years for over half of the respondents
(Total: 56 percent, EN: 63, FR: 46). Only 3 percent of the respondents across samples have
been following their brand page for more than 5 years, which is congruent with the fact
that most respondents joined Facebook around 2008. The frequency of active page visits
exhibits a high spread among respondents, with relatively similar frequencies levels for
 146

‘never’ to ‘more than once a week’. This indicates that respondent’s need to actively click
on the page is highly variable. Despite the spread in visit frequencies, time spent on the
page is clearly short for most respondents, with half of the respondents across samples
spending less than 2 minutes per week, and only 5 percent of them spending more than 15
minutes on the page per week. Regarding the perceived page size, results vary vastly but
remain consistent across languages. For both French and English surveys, although
medium-sized (38 percent) and fairly big pages (29 percent) are more represented, other
sizes are present as well.

4.7. Data analysis

The main data analysis method of this study is Structural Equation Modelling (SEM). SEM
was deemed the most appropriate method, given the type of the research questions, and the
nature of the data. SEM is suited to the testing of causal relationships for multivariate data
sets, which pertain to research questions 2 and 3. As Hair et al. (1998) explain, SEM is the
appropriate analysis technique when multiple relationships of dependent and independent
variables are investigated. In addition, it allows the development and validation of the
consumer engagement scale deriving from research question 1. Indeed, SEM allows not
only testing at the structural level, but also at the measurement level, thus encompassing
confirmatory factor analyses techniques, which are suited for tests of interdependence
between factors (Hair et al., 1998). Due to its ability to produce models with complex
interdependencies and its separate consideration of measurement and structural models,
SEM is an attractive option for post-positivist studies (Healy and Perry, 2000).

The measurement models are first assessed using a Confirmatory Factor Analysis (CFA).
This approach is used first for the consumer engagement scale development, as well as to
assess the whole measurement model prior to hypothesis testing. CFA is a method that
aims to test goodness of fit of a model and ensure unidimensionality of each hypothesised
factor, following Gerbing and Anderson’s (1988) guidelines. CFA is concerned with the
structure of data, and confirms how well the items and factors are related to one another.
Its strength lies in the ability to formally test unidimensionality of a scale initially
developed theoretically (Anderson and Gerbing, 1982), which particularly suits this study’s
approach.
 147

An analysis of the correlation matrix between factors is then conducted as a way to detect
singularity or multicollinearity between factors. Goodness of fit then is assessed using the
chi-square statistic, in combination with the CFI, TLI and RMSEA indices. These indices
are widely used to evaluate factor structures in online community, branding and cross-
cultural research. Moreover, they are less sensitive to sample size than the chi-square
(Bagozzi et al., 1991) and allow model complexity. The Comparative Fit Index (CFI) is
considered the index of choice. It compares the hypothesised model with a null (or
independence) model and takes into account sample size. The Tucker Lewis Index (TLI)
is also considered a choice index to report (Tucker and Lewis, 1973). Both these indexes
range from 0.00 to 1.00. Values above 0.90 are indicative of good fit (Bentler, 1992) and
any value above 0.95 is even more desirable (Hu and Bentler, 1999). The root mean square
error of approximation (RMSEA) is reported as it is considered the most informative
criteria in covariance structure modelling (Byrne, 2010). Values below 0.08 represent good
fit (Hair et al., 2006). Lastly, parameters estimates and standard errors are reported. This
procedure is adopted for the calibration of the consumer engagement scale, for its
validation, as well as for the testing of the measurement model including all the study’s
constructs.

A SEM with Maximum Likelihood estimation technique is then applied to test the
hypothesis. The same goodness of fit indicators and adequate values are used than for the
CFA. The TLI, CFI, RMSEA and chi-square are reported. In line with Bagozzi (1994), for
each hypothesis, the standardised path estimate, path estimate, critical ratio (CR) and
significance level are reported. The squared multiple correlations for (R2) is given for each
dependent factor, indicating how well a given variable can be predicted using a linear
function of a set of other variables.

Lastly, a multigroup analysis is used to test the invariance of model fit between the French
and the English-speaking samples, whereby the CFI delta between two models acts as the
indicator, needing to assume a value below 0.01 to verify invariance. All these analyses are
computed using the computer software AMOS (Byrne, 2010).

Prior to presenting the data analysis related to hypothesis testing, the first section of the
data analysis chapter is dedicated to the development of the consumer engagement scales,
which required a particular methodological and analytical approach, in line with the
 148

measurement literature. For clarity purposes, this section embeds both methodological
considerations along with actual data analysis.

4.8. Summary

Based on a post-positivist paradigmatic stance, this chapter has presented this study’s
methodology. Using the context of Facebook pages, data was collected using an online
instrument, which was designed and administered following strict guidelines, pre-tested
and piloted. The two-phase sampling procedure of Facebook pages and page members has
then been explained, followed by the sample characteristics, which is composed of a total
of 721 French and English-speaking page members. The chapter then presented the sample
treatment and characteristics and concluded with the techniques which have been chosen to
analyse the data (SEM and CFA), and the guidelines according to which they have been
conducted. Throughout the chapter, elements of method equivalence and bias related to
validation of the study with a French-speaking sample have been addressed.
 149

Chapter 5: Measurement

5.1. Introduction

This chapter develops and validates the scales used to measure the constructs of interest of
the study. The chapter is divided into two main sections. The first part outlines the
development of two new measurement scales for the study’s focal construct, consumer
engagement. To this end, the scale development process is presented, starting with the
collection and analysis of qualitative data to clarify the dimensionality and content validity
of consumer engagement. The section details how items were generated and purified.

The quantitative data are then analysed. A calibration sample (n = 224) is first used to test
for item reliability, purify these items and thereafter check item unidimensionality. The
validation sample (n = 224) is then used to verify these psychometric measures and ensure
reliability and validity of the data. Since consumer engagement is composed of dimensions
and sub-dimensions, two levels of confirmatory factor analyses are used, prior to creating
an aggregate measure of the different dimensions to simplify the scale. Lastly, it must be
noted that, in line with the conceptualisation of consumer engagement in OBC, two
mirrored scales are created, reflecting the two engagement partners present in this context.

The second part of this chapter presents the scales chosen from past studies to measure the
other variables of interest of the study. These choices are supported by evidence of
conceptual and contextual fit as well as strong psychometric properties of the scales.

5.2. Measurement of consumer engagement

Unlike other constructs, the focal concept of consumer engagement necessitated the
creation of a dedicated scale. At the inception of this study, no conceptually adequate and
valid scale of consumer engagement had been published, which could have been used or
adapted. The following section details the whole scale development process applied to
generate a valid and reliable scale of consumer engagement.
 150

5.2.1. Scale development process

To develop and validate the scale of consumer engagement, a relevant scale development
paradigm was used, starting with Churchill’s (1979) seminal work, and enriched with the
views of Anderson and Gerbing (1988), and Gerbing and Anderson (1988) to include
factor analysis. Figure 15 presents the stages involved in building the scales. It details the
methods used as well as psychometric properties tested at each point.

Figure 15: The scale development process

Sources: Churchill, 1979; Anderson and Gerbing, 1988; Gerbing and Anderson, 1988;
DeVellis, 2012; Netemeyer et al., 2003, Nunally and Bernstein, 1994.
 151

The reason for following this approach is twofold. Firstly, this paradigm is widely used in
the marketing literature, leading to the production of a vast amount of scales still in usage,
whether in the areas of relationship marketing, branding, or services marketing (e.g.
Parasuraman et al., 1988; Aaker, 1997; Christodoulides et al., 2006; Brakus et al., 2009).
Secondly, Churchill’s (1979) approach is endorsed in the scaling and psychometrics
literature outside of marketing (Nunally and Bernstein, 1967; Netemeyer et al., 2003;
DeVellis, 2012).

5.2.2. Domain specification and item generation

The first step of Churchill’s (1979) paradigm requires the specification of the domain of
the construct. This deductive approach to scale development is found in a large majority of
management studies, sometimes in combination or complemented by an inductive
approach (Hinkin, 1995), which adds to the iterative nature of scale development. In this
stage, ‘the researcher must be exacting in delineating what is included in the definition and
what is excluded’ (Churchill, 1979, p. 67).

As detailed in Chapter 3, this study adopts a definition of consumer engagement positioned

in the extended domain of relationship marketing, and it clarifies its dimensions as being
affective, behavioural and cognitive (Brodie et al., 2011; Hollebeek, 2011a and 2011b).
The literature review allowed building a solid theoretical definition of the construct and
provided dimensionality guidance (Netemeyer et al., 2003; DeVellis, 2012). Seven sub-
dimensions were further identified: attention, absorption, enthusiasm, enjoyment, sharing,
learning and endorsing. The duality of engagement partners in the context of OBC was
also evidenced; leading to the realisation that consumer engagement in OBC should be
measured using two mirrored scales rather than one, to reflect online consumer
engagement with both the OBC and the brand.

After specifying the domain of reference of consumer engagement in OBC, the next step is
to develop a pool of items to measure each of the sub-dimensions. These sub-dimensions
are what the literature refers to as latent variables, which are abstractions that cannot be
directly observed or measured (Netemeyer et al., 2003; DeVellis, 2012). In order to
operationalise these latent variables, items are required to tap into the domain of the
construct (Netemeyer et al., 2003).
 152

The approach taken in item generation is that of multi-item scales, as no single item is
likely to be enough to measure a construct (Churchill, 1979; Clark and Watson, 1995;
Netemeyer et al., 2003; DeVellis, 2012). In general, it is difficult to assess the
psychometric properties of single item measures. Multi-item scale enhances scale
reliability in particular, as measurement errors cancel each other out (Peter, 1979).

Items are thus generated for the purpose of this study, following the sub-dimensions
definitions and related key words. An iterative procedure was used at first, going back and
forth between the literature on consumer engagement and the qualitative interviews
presented below. Relevant items from the literature were researched, and they served,
when possible, as a basis for the development of an initial pool of measurement items.
However, given the paucity of engagement measurement in organisational sciences, few
items could be culled directly from existing scales and the qualitative phase proved
particularly helpful.

The following sections expand on the role of each scale development phase in item
generation and trimming. Since the conceptual premises are already presented in chapter 3,
this section starts by detailing the next steps involved in the item generation process, i.e.
the qualitative interviews with consumers and industry experts.

5.2.2. Qualitative data: Interviews of consumers and industry experts

Qualitative data are often used as part of a scale development process, and it is not rare to
include elements of inductivity in scaling processes, particularly when there is little
specialist theory on a construct (Hinkin, 1995). Procedures using a combination of
literature review and interviews typically produce scales with higher reliability than any
single-method approach (Churchill and Peter, 1984). This practice is also vastly used in
the past scale development literature (Lytle et al., 1998; Christodoulides et al., 2006;
Walsh and Beaty, 2007; Brakus et al., 2009; Freling et al., 2010, Sharma and Chan, 2011).

The reason behind using qualitative data to develop a scale is twofold. Firstly, it allows
validation of the conceptual dimensionality and domain specification (Churchill, 1979;
Hinkin, 1995) of consumer engagement. Secondly, and subsequently, it works toward the
generation of a more relevant pool of initial items. As Churchill (1979) points out, a
carefully selected sample of experienced consumers can offer precise insight into the
phenomenon, thereby allowing the researcher to tap into each of the identified dimensions
 153

with slightly different shades of meaning. Moreover, a number of previous studies in the
branding literature have used in-depth semi-structured interviews for the purpose of item
generation (e.g. Seiders et al., 2007; Brakus et al., 2009; Freling et al., 2010).

The use of qualitative phases for scale development purposes has been subject to criticism,
and not all scale development approaches adopt it (e.g. Netemeyer et al., 2003). However,
skipping this stage is best applicable in contexts where a vast amount of scales already
exist for the construct and ‘culling’3 is possible. Given the embryonic nature of the scale
literature on customer engagement, exploratory qualitative efforts were required to
determine the domain of observables (Hair et al., 2006).

Participants’ selection and profile

A set of 20 consumers and 5 managers, experts in social media, was interviewed.

Convenience sampling and snowballing techniques were used to recruit respondents
(Bryman, 2008). A number of respondents was known to the researcher, and the remainder
was acquired through recommendations of said respondents. As proposed by Churchill
(1979), experienced consumers can provide detailed insight into a construct’s domain and
dimensionality. The inclusion of industry experts can provide complementary insight to
consumers’, and it is not uncommon in the scaling literature (Vivek et al., 2014).
Respondents were recruited until saturation was reached in the analysis, as advocated by
Creswell (2007), who recommends 20 to 30 interviews until this point is attained.

Consumer informants were recruited based on two main criteria: their social media
involvement, and their socio-demographic profile. Given the nature of the research, it was
important to recruit experienced social media users, who exhibited levels of social media
activity, which can be qualified of ‘highly engaged’ thus ensuring high exposure and
participation in OBC. Selection of highly engaged individuals is a common practice in
brand community research (e.g. Muñiz and Schau, 2005; Cova et al., 2007). Participants’
level of engagement was assessed through a four week-long observation of their behaviour
on relevant platforms prior to their recruitment. In line with previous research in OBC (e.g.
Healy and McDonagh, 2013), this was done within the researcher’s extended network to
speed-up access. The observation included an assessment of the frequency and volume of
community interactions and duration of OBC membership (Kozinets, 2002). A high

3
Adaptation of items from existing scales
 154

frequency of posts, likes and comments was indicative of high levels of engagement
(Gummerus et al., 2012).

Secondly, diversity was sought in terms of the interviewee’s gender, occupation,

nationality and location, which seemed necessary to tap into the complexity of online
consumer engagement from different perspectives. Appendix 8 presents the profiles of
consumer respondents whereas Appendix 9 presents the industry experts’ features.
Industry experts were recruited based on their practical and strategic experience in social
media management, covering different function, including social media consultant,
marketing manager and community manager. Expert informants resided in Europe, India
and North America.

Interview modalities

Participants were contacted through private message on social media or via email.
Interviews lasted between 35 and 140 minutes and were conducted in English or in French.
All respondents were given the choice of the language, the researcher being bilingual and
richness in data being the goal. Translation equivalence was sought when translating the
interview guide (see below). The interviews were conducted face-to-face for respondents
based in the UK, and using Skype calls for non-UK residents.

The interview guide and interview process

Semi-structured qualitative interviews were used in accordance with Kvale’s (1996)

approach, advocating the use of a series of themes to be covered, as well as suggested
questions, but leaving room for unexpected topics. An open-ended elicitation procedure
(Netemeyer et al., 2003) was used to ensure that the author-generated construct definition
was consistent with the views of consumer engagement by typical consumers, and
managers, similarly to Walsh and Beatty(2007).The interview guide, which can be found
in Appendix 10, included broad question areas, each of which started with a grand tour
question followed by planned prompts (e.g. ‘Can you recall an instance in which you
engaged with this brand on social media?’) and floating prompts (e.g. ‘What do you mean
by…’) McCracken (1988). Participants were asked to first talk about their general online
behaviour, then their social media behaviour. They were subsequently asked to focus on
their participation in one, or several, OBC of their choice, with which they felt particularly
engaged. No predetermined social media or brand categories were suggested. Questions
then tapped into the emotional, behavioural and cognitive aspects of online consumer
 155

engagement, and concerned the relationship with the community on the one hand, and the
brand on the other hand. Interviewees were also asked to articulate their own definition of
consumer engagement. Overall, the guide was designed to go from the general to the
particular and make sure that all areas of interest were covered while remaining open to
unplanned information (Kvale, 1996). It concluded with profile questions.

Recoding and transcription

The entire set of interviews was recorded in order to ensure that the data was traceable, and
to create a consistent source format for the qualitative data. Interviews were then
transcribed for analysis, resulting in 192 single-space A4 pages of text. The researcher and
interviewer performed a full transcription of the interviews. This allowed maintaining the
full richness of the data and a first immersion in the interview content prior to analysis.

Interview data analysis

The analysis of the qualitative data issued from interviews with 20 consumer and 5
industry experts serves several purposes and is in line with conceptual developments in
chapter 3, and the study’s first research question.

1) Confirm the existence of the three key dimensions of consumer engagement;

2) Clarify and deepen the meaning of these three broad dimensions, seeking evidence
of the 7 conceptual sub-dimensions;
3) Generate measurement items based on the dimensions and sub-dimensions.

To achieve these goals, the interview data were coded and analysed manually, the aim of
the coding being to bring together all extracts of data (sentences, words, expressions,
paragraphs) pertinent to a particular theoretical theme, or topic. A line-by-line content
analysis was performed (Miles and Huberman, 1994). Expressions or phrases that offered
insight into consumer engagement in OBC were sought to generate a pool of measurement
items. Coding was done using a directed data analysis approach (Hsieh and Shannon,
2005): codes were first created based on theoretical themes (the expected theoretical
subdimensions) and keywords were identified based on the data. A constant iteration
between data and theory also underpinned the analysis, as codes were progressively
integrated and adapted based on the data, and in line with the customer engagement
literature.
 156

This section details the existence and nature of each of the three dimensions and seven sub-
dimensions identified in the literature. It provides a deeper meaning to these aspects of
consumer engagement with the support of interview quotes and analysis. In order to
respect the anonymity of all respondents, fictitious names are used throughout the analysis.
The theoretical codes and keywords are presented in Appendix 12. For each of these, a
summary of three representative quotes from the data is provided.

Theme 1: Dimensions of engagement

Based on an iterative procedure of data analysis and theoretical framing, the results of the
interviews with social media users are not only aligned with the literature on the
dimensionality of engagement, but they also complement it and deepen the understanding
of these dimensions. The three main aspects of consumer engagement in OBC have
theoretically been categorised as: affective cognitive and behavioural, in line with most
conceptualisations of the concept. However the conceptual fog surrounding these
dimensions prompts empirical support.

Analysis of the data allows breaking down the three dimensions into seven sub-
dimensions, as per the conceptual framework: (1) Enthusiasm, (2) Enjoyment, (3)
Attention, (4) Absorption, (5) Sharing, (6) Learning and (7) Endorsing.

Affective engagement

The existence of an affective aspect of engagement in the data first became apparent
through the repeat usage of keywords such as ‘bond’, ‘care’, ‘emotions’, ‘love’, ‘hate’ or
‘like’ by respondents. This semantic particularity was further explored, and revealed a
strong affective element to respondents’ experiences in OBC.

The questionnaire started by asking interviewees about their general experience on OBC
embedded on social media, leaving them free to explore the topic as they felt. A lot of
them expressed that they were part of an OBC on Facebook for emotional reasons, such as
simply liking the brand, or finding that the content they post is ‘nice’, ‘entertaining’ or
‘fun’. Others explicitly stated that they felt they had a bond with the brand, that they shared
something, or had the same values. This represents a strong form of emotional bonding
close to what is expressed by the construct of relationship quality (Hollebeek, 2011) Others
acknowledged that seeing the branded content online brought them happiness and made
them feel good, through the aesthetic of a picture, on the content of a text.
 157

When prompted to define what engagement represented, respondents used expression such
as ‘it is being excited about something, at least a little bit’ (Derek), or ‘it is enjoying doing
something’ (Claire). Brand engagement in particular was associated with ideas of ‘liking’
the brand or ‘caring’ about it. Depth was brought into the discussion by some interviewees
who went as far as recognising that engagement, although relying on affect, did not
necessarily have to be positive, and could be linked to elements of complaint, loss of trust,
or negative emotions in general (Hollebeek et al., 2014). ‘Even if you hate it, like anti-
Walmart or any organisation, it’s still engagement I would say’ (Nigel). Aubin, an expert
informant from the company Agentia, even nuanced the positive/negative affect
dimensions by explaining that too much emotional engagement can also be detrimental in
the sense that it disturbs consumers’ ability to make sound consumption decisions.

Overall, the affective dimension of engagement captures the summative and enduring level
of emotions experienced by a consumer with respect to his/her engagement focus (Calder
et al., 2013). These emotions can be of different sorts and transpires through long-lasting
and recurrent feelings. In expressing their feelings about the OBC they are members of,
and the brands they represent, respondents allow restricting the affective dimension to two
sub-dimensions: enthusiasm and enjoyment.

Enthusiasm reflects a consumer’s intrinsic level of excitement and interest regarding the
OBC or brand. Enthusiasm is explicit when people show genuine and active interest in the
brand or community activities, or derive such enthusiasm form their own repeat
participation and interactions. Anthony and Maria show that being enthusiastic about a
brand can go as far as putting oneself at risk for it, or liking it so much that you want to
work for the company.

‘Being engaged with (…) is to tell oneself that you are so interested in it, that you spend
time and effort on it, and that to some extent, you even put yourself at risk for it.’ Anthony.

‘I was really thinking of applying for a job there, because I like this company because of
Facebook! I don’t know, they have a nice face, they are very kind, and helpful and
enthusiastic about their products and I relate to that, you know’ Maria.

Liam on the other hand, expresses true enthusiasm for community participation and relates
these feeling to repeat interactions with other community members.
 158

‘If I like a brand post, or comment on a brand status, my friends in turn are likely to
comment on it as well. And in these cases, I am so excited that I am quite happy to keep the
conversation going and talk more’ Liam.

Enjoyment is indicative of the consumer’s pleasure and happiness derived from

discoveries and interactions triggered by the brand or OBC. As expressed by Sam about his
participation on the Pakistani National Cricket team Facebook page: ‘Exchanging ideas
about it makes me happy’. In contrast to enthusiasm, enjoyment is a more contemplative
form of affect. It is less active and motivational, but taps into deep and repeated feelings of
joy and pleasure. Anthony hereunder expresses the satisfaction he receives from getting
comments on his posts, despite his clarification that it is not the number one goal.

‘It’s not necessarily important to have comments on what you posted, but it’s a pleasure,
it’s a nice added value’ Anthony.

Sabrina, on the other hand, shows that one can derive pleasure simply from the content
posted by the brand.

‘They always have something fun to tell on their page, something that is really ‘Nutella’,
something that is really about gourmandise, fun...so I really like this page because it
represents me and it represents what I enjoy in life’ Sabrina.

The affective dimension of engagement relates to various forms of content and

interactions. Respondents express pleasure in seeing comments on their own posts and
sustaining the conversation by replying to these. They enjoy seeing and replying to the
brand or community members posts. At other times, they simply enjoy reading fun and
relevant posts by the brand.

Consumer’s views on the affective dimensionality of engagement are complemented by the

views of the industry experts. Although they do not go as far as pinpointing the sub-
dimensions of affective engagement, they all recognise that consumers need to be engaged
by companies at an affective level, reflecting the depth of the relationship they have.
Considering the interactive nature of engagement, they explain that consumer are more
likely to be emotionally engaged if they see that the brand cares for them as well.

The interview data are therefore is accordance with the literature, which indicated, mainly
on theoretical grounds, that affect is one of the core three dimensions of consumer
 159

engagement and that it can be broken down into enthusiasm (Schaufeli and Bakker, 2004)
and enjoyment (Calder et al., 2013).

Cognitive engagement

The cognitive aspect of consumer engagement was prominent in the data. In an

engagement context, cognition has been defined a set of enduring and active mental states
that a consumer experiences with respect to the focal object of his/her engagement (Mollen
and Wilson, 2010; Hollebeek, 2011a). This means that, when engaged, consumers must be
actively mentally involved. When discussing their activity on OBC, informants often point
out that they spend a vast amount of time on social media, a part of which is dedicated to
reading, interacting with, or searching brand-related content. When talking about a
community she engages with, Flora expresses that she consciously makes time to think
about it.

‘It [engagement] just depends on how much time you are willing to sacrifice for the
group…how much time you spend thinking about it’ Flora.

Further exploration of the cognitive aspect of informants’ OBC experience reveals that this
dimension can be broken down into two complementary sub-dimensions: attention and
absorption.

Attention is the cognitive availability voluntarily dedicated to interacting with the OBC, as
Flora’s quote signalled. Sophia furthers this comment by vividly expressing her view of
online communities participation: ‘It is an engagement of the mind!’ Interviewees clearly
exhibit consciousness that time spent on OBC requires some mind space, which keep them
from performing other tasks. Despite the fact that attention span is relatively short on
social media, if brands provide enough interesting content at a relatively high frequency,
users’ attention can be grabbed and sustained, as Derek explains. This consumer view is
complemented by industry experts, as both George from Ironvalley and the consultant
team from SmartForest agree that in order to gain consumer’s attention on social media,
brands must post interesting and relevant content.

Absorption is indicative of the inability to detach oneself when interacting with the OBC
(Schaufeli and Bakker, 2004), and therefore goes a step further than dedicating attention to
something. This aspect is particularly well described by Judith and her seemingly extreme
relationship to Facebook and Pinterest participation.
 160

‘So when I go on Facebook I essentially read the newsfeed, but I really read every single
thing, even if it takes me hours. It’s really stupid and every time I get mad at myself
because I am wasting time and I don’t really see the point, but I always end up finding
something interesting so I tell myself that I do get something out of the two hours I just
spent on it. As for Pinterest, I try not to go too often, but when I do go, it takes a lot of time.
It’s like Facebook but it is even worse…Facebook I am more able to turn it off when I
really want to. Pinterest it’s impossible!’ Judith.

Although this might seem like an extreme level of absorption, number of interviewees
exhibits a considerable difficulty to detach themselves from their favourite social media.
They reflect that this absorption is largely due to the interesting and entertaining content
posted by brands on such platforms (Chauhan and Pillai, 2013).

High levels of absorption can also be related to engagement in a specific brand-related

event, as the case of Sabrina portrays. She explains taking part in a photo-posting contest
launched by one of her favourite Facebook pages. She recalls that she spent ‘three whole
days’ on the page to promote her participation in the contest and try and win. She states
that she was ‘really engaged with the page at the time of the contest’, which indicates a
complete dedication and absorption (Patterson et al., 2006). This absorption was triggered
by the brand campaign, and sustained by collaboration and help from other community
members, showing the importance of all OBC actors in the mental activation process.

Behavioural engagement

When prompted to define consumer engagement, a number of respondents immediately

pointed out the behavioural aspect of the construct, in line with studies that put behaviours
at the centre of engagement practice (e.g. van Doorn et al., 2010; Verhoef et al., 2010). In
particular, Nigel speaks of being ‘proactively and physically involved in an activity’,
Sandra of ‘going out of a passive situation and entering an active situation’, Sophia of
‘constantly interacting with a thing’. Put simply, being engaged signifies doing something.
A deeper exploration of informants’ OBC participation allowed the classification of
engagement behaviour in three interrelated, yet distinct performative acts.

Sharing is a highly recurring theme in the participants’ stories. Indeed, social media
environments and brand communities lend themselves particularly well to the development
of sharing behaviours, as they are based on usage and content. On Facebook, sharing
manifests itself through ‘shares’, ‘comments’, ‘posts’, ‘likes’ or ‘replies’. When asked
 161

about what they do on OBC and what they used them for, Maria and James, among others,
speak of sharing with the community, either on a one-off or regular basis.

‘[I use the Facebook group]…to exchange experiences about visits. If there is a place
where we have been, we can inform other people who are interested in visiting!’ Maria.

‘Well yes, for instance, if somebody asks a question about a football game (i.e. ‘Did you
see what just happened?’) I would very quickly answer.’ James.

These quotes demonstrate that the act of sharing is a collaborative and interactive
exchange, driven by the motivation to provide resources. In line with Brodie et al.’s (2013)
findings, sharing is strongly present on social media and a way for OBC members to
exchange experience, ideas or just interesting content. In accordance with seminal research
on co-creation (Vargo and Lush, 2004) OBC engagement relies heavily on the exchange of
experiences (Vivek et al., 2012), content and information.

In contrast with the sub-dimension of endorsement detailed further in this section, sharing
is not necessarily based on positive, supporting intentions. While some respondents
express the need to share their views on OBC, whether good or bad, such as Sam, others
like Helen go one step further by saying that this openness and authenticity is exactly what
they seek on OBC.

‘I would give my views, my very hard views even. When I don’t like something I say it, I
don’t mind, I have to raise my voice sometimes, which is something that people don’t often
do. ‘ Sam.

‘You know, I love these pages on Facebook more than the company websites, because it’s
more objective, you can get different opinions! Ok, if you want to buy something you can
go to the website, but this kind of things (the varied consumer views), you cannot find it
easily.’ Helen.

Learning. As much as respondents expressed an urge to share resources online, they also
use OBC to seek help, ideas, resources and information from the company or other
consumers (Hennig-Thurau et al., 2010). Learning thus represents the other side of the
coin. Like sharing, learning is an active conduct and an important facet of consumer
engagement (Brodie et al., 2013), as shown by the increased focus on content strategies.
 162

‘If you post something on Twitter people will help you and give you suggestions, which I
think is quite nice because it helps. Like, I got a pen burst out in one of my favourite bags
and I tweeted about it and asked if anybody had any ‘at-home’ remedies for what to do,
and I got loads back’ Claire.

‘I follow them just to make sure I know which products they are launching, what is in their
new summer collection. I want to know what is new at the moment ‘ Sophia.

Enhancement of brand usage, problem solving or information updates are a key aspect of
OBC participation (Dholakia et al., 2004; Wiertz and de Ruyter, 2007) and social media
are particularly suited to this goal. They allow users to post their questions freely and
receive feedback from other knowledgeable members, or the brand itself (Zaglia, 2013).

Learning represents a complex sub-dimension of engagement, as it could be considered to

be a passive rather than an active endeavour. Despite the fact that the information-
processing aspect of learning intuitively lends itself to a cognitive categorisation, excerpts
from the interview suggest its underlying behavioural nature, as they stress the active and
committed act of looking for information. Sandra makes this active/passive distinction
particularly clear:

‘Yes, Facebook helps me connect with the brand, because I can see maybe a campaign on
TV or hear something on the radio, or see a banner, but on Facebook, when you become a
fan of the page, you always see information about the brand. Maybe you are not going to
comment or like every time, but you will see what they are talking about, so unconsciously,
you will be updated about what the brand is doing, whereas watching a TV campaign is
really…you just sit down and watch and don’t do anything, it’s passive. On Facebook, you
are active, so it drives engagement.’ Sandra.

Consumers show engagement by searching to improve their experience, learn more, or fix
issues. Passively receiving information, like watching a TV ad, is not indicative of
information search and is rather found in a disengaged consumer (Lee et al., 2009).
Participants suggest that learning behaviour is done for instance by using the community as
a source for the latest news and trends or by seeking help for a specific consumption-
related issue.

According to the industry experts, allowing consumer to learn is key to a good social
media content and engagement strategy. By educating consumers with information that
 163

they are actively or unconsciously searching for, experts Aubin and George explain that
higher levels of engagement can be fostered.

Endorsing is a third behavioural manifestation that came through interview analysis. It is

considered to be different from the act of sharing, as it carries stronger elements of
purpose, and is exclusively positive in valence. Endorsing is done on social media in a
variety of ways (Gummerus et al., 2012), which our respondents seem to exploit fully, or
wish to do more than they can afford.

‘I’m liking things a lot, I’m the kind of person that sees something and then, hop, I like it’
Judith.

‘I took part in the vote (launched by a design brand) and then promoted it on Facebook.
It’s not only because I want people to buy their product, but because they are really nice
and really good.’ Laura.

‘If days were twice as long, I would love to spend much more time on social media and
write product reviews for the artists I like (…) I know that I have a form of power when I
say things. When I give conferences, or when I post something, people forward or report
it.’ Anthony.

Interestingly, this motivation to share can go as far as developing a conscious self-

positioning as an expert, or at least recognising one’s influencing potential.

In addition to this, endorsement can also go beyond the community boundaries rather than
being limited to close community settings, when people want to get others to discover
something they like (van Doorn et al., 2010). Schau et al. (2009) refer to external
endorsement as ‘impression management’. This is supported by Brodie et al.’s (2013) sub-
dimension of ‘advocating’. This occurs ‘when consumers actively recommend specific
brands, products/services, organisations and/or ways of using products or brands’

‘I suggested my friends to follow them on Facebook, and see more designs. I didn’t just say
that I bought my shoes from a store and gave them the address; I told them ‘No, you have
to go on the page and see more design and ask them more questions, they are very kind.’
Mary.

The endorsement sub-dimension is also expressed as key to engagement by industry

experts with Benjamin from SmartForest stating that consumer engagement results in
 164

publicly showing the connection that one has with a brand. Exhibiting this positive
connection and promoting the brand in place of the brand is here again understood as
expressive of consumer engagement. As Daniel explains, spreading word-of-mouth is a
key component of engagement.

As this analysis shows, the three theoretical dimensions and seven sub-dimensions of
consumer engagement are reflected in the data. Participants’ experiences not only
corroborate existing conceptualisations but, most importantly, deepen and clarify them
through rich social media user insight.

An interesting take from experts’ interview which support the multidimensionality of

engagement is that all of them agree that consumer engagement on social media cannot be
measured only with site metrics such as ‘likes’ and ‘comments’ and that consumer
sentiment and original content need to be taken into account. According to SmartForest,
the traditional measure of consumer engagement on Facebook measured by
(likes+shares+comments)/number of persons who see the post, is not reflective of the full
picture. This comment is clearly aligned with the multidimensional stance of consumer
engagement taken in this thesis.

Based on the consumer interviews and the consumer engagement domain specification
performed based on the literature, a set of items was created. The first iteration resulted in
64x2 items, subsequently revised and narrowed-down to 47x2 items, which were them
submitted to a panel of academic experts to ensure content validity.

Panel of academic experts

Following the analysis of the interviews with the consumers and industry experts, and
having generated a preliminary pool of items resulting from these interviews and the
literature, a panel of academic experts was approached. Exposing experts to the list of
preliminary items is a common method in marketing scale development (e.g. Veloutsou et
al., 2013).

The goal in seeking experts’ opinion is to ensure content validity. Content validity is
defined as ‘the degree to which elements of an assessment are relevant and representative
of the targeted construct for a particular assessment purpose’ (Haynes et al., 1995, p.238).
Experts’ panels are particularly helpful in confirming or invalidating the definition of the
 165

construct and rating the relevance of each item with respect to what it is supposed to
measure. They also help ensure item’s clarity and conciseness and can point out any
missing items (DeVellis, 2012).

Experts were selected based on their experience and publication record in customer
engagement and branding. A number of them also had extensive experience in scale
development methods. They came from institutions in the United Kingdom, continental
Europe, the United-States and Oceania. In total, 12 academics were approached, and 9 of
them completed the questionnaire, some of them even following up with further
considerations by email. These 9 answers were gathered over the course of 3 weeks. The
number of respondents is in line with Haynes et al.’s (1995) guidelines, who commend the
use of over 5 judges.

Experts’ point of view was gathered following thoroughly Netemeyers et al. (2003) and
DeVellis’ (2012) recommendations. They were contacted by email using an introductory
text, which was presenting the context and purpose of the study. This can be found in 11.
They were then redirected to an online questionnaire, which contained all 47 initial items.
The questionnaire was designed as follows: the concept of consumer engagement was first
defined. Then, each category (affective, cognitive and behavioural) was addressed. For
each of them, a definition was provided, and the sub-dimensions identified and defined as
well. The items were then presented and experts were asked to rate the representativeness
of each item to its dimension on a 5-point Likert scale (1 = completely agree; 5 =
completely disagree) in accordance with Haynes et al. (1995). This was done for OBC
engagement on the one hand, and brand engagement on the other hand. For each sub-
dimension, the experts were encouraged to leave comments regarding the clarity and
conciseness of items, or provide any other relevant comment. More general comments
could be left at the end of the questionnaire. This approach was chosen as it is deemed that
a combination of quantitative and qualitative insight is preferred to identify elements of the
items that need to be changed, refined or deleted (Netemeyer et al., 2003).

It is recommended to analyse data from the experts’ panel carefully, taking into account
their own field and methodological biases. As no statistical analyses can be significantly
computed on such a small sample, the answers for item representativeness were analysed
one by one. Items with particularly low representativeness scores from most experts were
considered for deletion. Advice on the redundancy or semantic similarity of certain
 166

dimensions and words was also taken on board. A constant re-framing of the comments in
line with the literature was made, making sure to cover all dimensions of customer
engagement, whilst also taking into account experts’ ideas that were initially unthought-of.
As a result from the experts’ panel advice, 8 of the 47 items were deleted due to
redundancy or failure to appropriately tap into the latent dimensions. 14 items were also
edited to improve wording clarity and conciseness, and avoid redundancy in meaning.

Final pool of items

To summarise, the first stage of item generation relied on three successive iterations,
balancing views between the data and literature (Clark and Watson, 1995), with the aim of
coming up with a large pool of items (64 x 2 for the initial iteration), which were
subsequently narrowed down in two more iterations at the end of the interview analysis.

Following these first iterations, 47 x 2 items were presented to a group of academic

experts, which resulted in further trimming, leading to 2 x39 items. The final item
trimming was produced after the pre-test and pilot phases. The final pool resulted in a total
of 35 x 2 items to measure OBC engagement and brand engagement with mirrored items to
be included in the online survey. Figure 16 shows the item evolution from its early
iterations to the pool used in the large-scale consumer survey, with a dimension
breakdown.

Figure 16: Evolution of the number of items

20
19 17 22
13 22 18
10
25 17 15 8 8
9 9

Affective Cognitive Behavioural

 167

Table 12 presents the 47x2 items generated after the third iteration resulting from the data
analysis and literature review. The main source of item generation is also presented. A
reference indicates that the item is issued from the literature, and the word ‘interviews’
indicates that the item came through mainly in the consumers and industry experts’
interviews. This pool of item was first reviewed by academic experts and was subsequently
the subject of pretest and pilot studies. The last column of the table shows which items
were removed or edited after the academic expert’s feedback and pretest/pilot phases.
 168

Table 12: Consumer engagement items evolution

Dimensions OBC/online brand engagement Source Reason for deletion or edit
Enthusiasm
I feel excited about the group/brand Vivek, 2009 DELETED - Experts
I feel excited about what the group/brand does Vivek, 2009 DELETED - Experts
The group/brand generates in me a feeling of excitement Vivek, 2009 DELETED - Experts
I feel enthusiastic about the group/brand Interviews
The group/brand makes me enthusiastic Interviews
I am heavily into this group/brand Vivek, 2009
I am interested in anything about this group/brand Interviews
Affective I find this group/brand interesting Interviews
This group/brand drives my interest Interviews DELETED – Pretest
Enjoyment
I enjoy being part of the group/a fan of the brand Interviews DELETE - Experts
I enjoy interacting with the group members/brand Interviews
When interacting with the group/brand, I feel happy Schaufeli et al., 2002
The group/brand makes me feel good Interviews DELETE - Experts
I get pleasure from group participation/interacting with the brand Interviews
Participating in the group/interacting with the brand is like a treat for me Calder et al., 2013
Attention
I pay a lot of attention to the group/brand Rothbard , 2001 EDIT - Experts
Things related to this group/brand grab my attention Vivek, 2009 EDIT - Experts
I spend a lot of time thinking about the group/brand Rothbard , 2001
I make time to think about the group/brand Interviews
Absorption
Cognitive
I concentrate a lot on this group/brand Rothbard , 2001 DELETE - Experts
When I interact with this group/brand, I forget everything else around me Schaufeli et al., 2002
Time flies when I am interacting with this group/brand Schaufeli et al., 2002
When I am interacting with this group/brand, I get carried away Schaufeli et al., 2002
When interacting with the group/brand, it is difficult to detach myself Schaufeli et al., 2002
In my interaction with the group/brand, I am fully concentrated Schaufeli et al., 2002 DELETE - Experts
 169

Dimensions OBC/online brand engagement Source Reason for deletion or edit

Sharing
I reply to the questions of other group members/of the brand managers (through
Gummerus et al., 2012 EDIT - Experts
commenting, sharing, posting, replying, etc.)
I share my opinion with the members of the group/brand managers Interviews EDIT - Experts
I express my opinion to the group/brand managers Interviews DELETE - Pretest
I share my experiences with the group/brand managers Interviews EDIT - Experts
I share ideas with the members of the group/brand managers Interviews EDIT - Experts
I provide ideas to make the group/brand better Interviews DELETE - Pretest
I share interesting content with the other group members/brand managers Interviews EDIT - Experts
I provide help to other group members/brand managers Interviews EDIT - Experts
Learning
I ask questions to the other group members/brand managers Interviews EDIT - Experts
I seek ideas, or information from other members of the group/the brand managers Interviews EDIT - Experts
Behavioural I seek help from group members/the brand managers Interviews EDIT - Experts
I seek information from other members of the group/the brand managers Interviews EDIT - Experts
I learn from the content provided by other group members/the brand managers Gummerus et al., 2012 EDIT - Experts
Endorsing
I show support to what the members of the group/the brand say/s or do/es (by liking,
Interviews
sharing, commenting)
I approve the group members'/brand’s behaviour Interviews DELETE - Experts
I approve the group members'/brand’s ideas Interviews EDIT - Experts
I endorse the group/brand Interviews DELETE - Pretest
I share content from the group/brand to my wider network Interviews
I promote the group/brand Interviews
I try to get others interested in the group/brand Interviews
I actively defend the group/brand from critics Interviews
I say positive things about the group/brand to other people Lee et al., 2011
 170

Based on the final pool of items derived from the exploratory phase, the large-scale survey
as described in the methodology chapter was completed, allowing the use of this data to
build two reliable and valid scales of consumer engagement based on psychometric testing
described below. The following section explains the statistical development of the OBC
engagement and online brand engagement scales. The two constructs are analysed and
modelled separately throughout the analysis, yet always in a mirrored fashion to ensure
perfect replication of the content and structure of the scales.

5.2.3. Data screening

Data screening is an essential step in the preparation of the data for analysis for the
researcher to become familiar with it and detect potential issues. Graphical and numeric
data outputs are explored in this process, which involves normality checks and sample
treatment. As all survey data collected for the purpose of this study need to exhibit the
same standards of cleanliness, data screening is performed on the aggregate French and
English sample (n=721). Descriptive statistics for all measures (mean, median, range, SD)
can be found in Appendix 6.

Normality

Prior to conducting the engagement scales development and hypotheses testing, an

important step of data screening regards the assumption of normality. The Kolmogorov-
Smirnov and Shapiro-Wilk tests are performed, and all variables display significance
levels of p<.05, which is indicative of non-normality. However, Field (2009) and
Tabachnik and Fidell (2000) warn researchers that these tests are usually irrelevant for
large samples. Skewness and Kurtosis indices are thus computed, measuring respectively
the symmetry and peakedness of the distribution.

Any value above |1| signals a departure from normality. The analysis revealed a number of
issues, as highlighted in Appendix 6. Items measuring absorption with the OBC items
display a slight positive skew and the first brand trust item has a slight negative skew.
Most abnormal values however reflect Kurtosis issues: the fourth item of OBC absorption
also displays positive Kurtosis, while a number of items exhibit slight negative Kurtosis,
and are largely centred mainly on some of the consumer engagement constructs.

Despite these slightly above average values, no manipulation of the data is judged
necessary, for the following reasons. Firstly, although the modified solution might be a bit
 171

better (Tabachnik and Fidell, 2000), normality is not an essential criterion to proceed with
analysis, and it is not required to process with SEM and CFA. Secondly, bar a few
exceptions, departure from normality is marginal, as most values remain below |1.1|, and
significant skewness and Kurtosis values can arise even from small deviations from
normality (Field, 2009). Moreover, typically, data from 7-point scales are not normally
distributed (Malthouse, 2001). Lastly, after transformation, a different construct than the
initial one is measured, and consequences of applying the wrong transformation might hurt
the data (Field, 2009). The assumption of normality is further checked by examining the
frequency histograms and their distribution curves, which are all mound shaped, implying
that deviations of normality are not so severe as to be worrisome.

Lastly, the consideration of possible outliers that might affect normality is considered.
Common causes for outliers (Tabachnick and Fidell, 2000) are not applicable to this study
in line with the strict data collection and screening procedures use: incorrect data entry,
failure to specify missing value code or respondent outside the population are all controlled
for and should result in outlier-free data. To verify this assumption, an univariate outlier
detection method is applied as proposed by Tabachnick and Fidell (2000) and z-scores
computed for all items. Values consistently comprised between -3.00 and 3.00 indicate an
absence of univariate outliers.

Sample split

At this stage in the data treatment, the full sample is split. Firstly, French and English
samples are used separately: the consumer engagement scale validation requires testing in
one cultural context first, prior to being applied to other cultures. Secondly, one cannot
assume group invariance between two languages without testing it. As the survey is
initially crafted for the English sample, this sample is used for the scales development and
hypotheses testing, leaving late-stage group invariance testing to account for the
equivalence between the two languages. This procedure is in line with previous studies
developing scales using cross-cultural samples, such as Cadogan et al. (1999).

Thereafter, the English sample is kept and further split into a calibration and a validation
sample for the purpose of the consumer engagement scales development, rather than
collecting new data for measure validation (Churchill and Peter, 1984). The split sample
technique is common in the marketing literature on scale development (e.g.
Christodoulides et al., 2006). This approach is chosen due to its practical time-saving
 172

aspect, as well as methodological advantages. Some methods of measure development,

such as confirmatory factor analysis, tend to become too sensitive with samples exceeding
400 cases. Goodness of fit measures also tend to indicate poor fit with larger samples (Hair
et al., 1998). To avoid this problem, it is recommended to use samples of about 200-300
cases (Diamantopoulos and Siguaw, 2000). In this instance, the English sample has been
randomly split into two groups, using the SPSS random sample function, and producing
two samples of 224 cases each. The first half (calibration sample, n1=224) has been used
to develop the scale, whilst the second half (validation sample, n2=224) has been used to
validate the results of the scale and subsequently test the research hypotheses with
structural equation modelling (Churchill, 1979; Flynn and Pearcy, 2001).

5.2.4. Item purification

The first step in statistically purifying the instrument is to compute the coefficient alpha
and inter-item correlations to delete ‘garbage items’ (Churchill, 1979). Cronbach’s Alpha
assesses unidimensionality and internal reliability (Churchill, 1979). The seven theoretical
dimensions are used to pre-group items, to accommodate the factor analysis requirement
that variables that are related should be separated prior to running factor analysis (Sharma,
1996). No Cronbach’s Alpha achieves a value below the advocated cut-off point of 0.70
(Nunnally and Berstein, 1967), and all values are above 0.84, exhibiting excellent
reliability, as shown in table 13.

Table 13: Cronbach's Alpha

Online brand engagement OBC engagement

Factor Cronbach’s Alpha Factor Cronbach’s Alpha
Enjoyment 0.95 Enjoyment 0.93
Enthusiasm 0.94 Enthusiasm 0.95
Attention 0.88 Attention 0.84
Absorption 0.95 Absorption 0.96
Sharing 0.94 Sharing 0.96
Learning 0.89 Learning 0.93
Endorsing 0.92 Endorsing 0.95

Inter-item correlations within the theoretical dimensions are then checked to detect signs of
singularity (item does not correlate with others) or multicollinearity (item correlates too
much with others), as well as verify the item to total correlations. Again, theoretical
dimensions are kept for item grouping. An analysis of the item-to-total matrixes indicates
 173

that the removal of the fourth ‘learning’ item (‘I learn from the content they provide’) in
both scales significantly increases the respective Cronbach’s Alphas to 0.89 and 0.93. A
poorer correlation between the fourth and fifth learning items is a further indication of the
need to remove this item in both scales. Lastly, the second ‘attention’ item also exhibits
low correlation with the other items of the same factor, in both scales. Because the deletion
of this item has no significant impact on the Cronbach’s Alpha, it has been decided not to
remove it at this stage. No other multicollineraity or singularity issue transpires from the
inter-item correlations.

5.2.5. Confirmatory Factor Analysis

A CFA has been used to test the dimensionality of the engagement scale. CFA is preferred
to EFA when measurement models have a strong and well-developed underlying theory for
hypothesised patterns of loadings and a detailed model has been specified prior to data
collection (Hurley et al., 1997). CFA estimates the regression coefficients between the
items and the latent constructs (Bagozzi, 1994). To achieve convergent validity, the
regression coefficients must be statistically significant and the model must show good fit
values.

A measurement model has thus been specified separately for OBC engagement and online
brand engagement, retaining for each of them the factors and items specified by the theory
and content validation stage. The procedure has been performed separately for online
brand engagement and OBC engagement items, although in parallel to maintain similarity
across scales. Given that they measure the same concept for two different objects, the two
scales should be mirrored in terms of dimensions, although they must be modelled
separately given their different engagement object.

A five-step approach has been used to perform the CFA:

1) Specification of the measurement model;

2) Model identification;
3) Model estimation;
4) Goodness-of-fit evaluation;
5) Check of the parameter estimates.

Model identification is the first step whereby parameter values are either left to be
estimated or constrained to 1. All independent variable variances are constrained to one,
 174

including factor and error terms’ variances in the aim to create an over-identified model.
Maximum Likelihood estimation is used as a mode of estimation, as it constitutes a widely
used and robust method that can also account for normality discrepancies (Byrne, 2010).

An analysis of the correlation matrix between factors has then been then conducted as a
way to detect singularity or multicollinearity between factors. Goodness of fit then has
then been assessed using the chi-square statistic, in combination with the CFI, TLI and
RMSEA indices. These indices are widely used to evaluate factor structures in online
community, branding and cross-cultural research. Moreover, they are less sensitive to
sample size than the chi-square (Bagozzi et al., 1991) and allow model complexity. The
Comparative Fir Index (CFI) has been considered the index of choice. It compares the
hypothesised model with a null (or independence) model and takes into account sample
size. The Tucker Lewis Index (TLI) has also been considered a choice index to report
(Tucker and Lewis, 1973). Both these indexes range from 0.00 to 1.00. Values above 0.90
are indicative of good fit (Bentler, 1992) and any value above 0.95 is even more desirable
(Hu and Bentler, 1999).

Lastly, the root mean square error of approximation (RMSEA) is reported as it has been
considered the most informative criteria in covariance structure modelling (Byrne, 2010).
Values below 0.08 represent good fit (Hair et al., 2006). Lastly, parameters estimates and
standard errors are reported. The procedure is then repeated in the following section with
the validation sample and an invariance test computed across the French and English
sample to cross-validate the scales.

Given the hypothesised dimensions and sub-dimensions of consumer engagement a 2-step

procedure has been applied. CFAs are first computed at the sub-dimension level. The sub-
dimensions of the same dimension are grouped together and correlated, and thus three
first-order CFAs are performed for each scale. Subsequently, second-order CFAs with the
three engagement dimensions (affective, cognitive and behavioural) are performed for each
scale. Each item is prescribed to load on only one factor, or sub-dimension, as theoretically
hypothesised. Each of the two levels of CFA is performed using the English calibration
sample and the statistical software AMOS. The two CFA levels can be visualised on
Figure 17 below.
 175

Figure 17: The two-level CFA

 176

CFAs on calibration sample

The sub-dimensions belonging to the same dimension are grouped for the first-order CFA.
Specifically, enjoyment and enthusiasm are correlated for the affective dimension,
absorption and attention for the cognitive dimensions, and learning, sharing and endorsing
are grouped to represent the behavioural dimension (see Figure 17). Some of the models
initially exhibiting poor fit, model respecifications are performed based on the
modification indices. The use of modification indices to re-specificy a model is a common
practice and considered to be an exploratory adaptation of the normally confirmatory CFA.
However, Gerbing and Hamilton (1996, p71) contend that ‘most uses of confirmatory
factor analyses are, in actuality, partly exploratory and partly confirmatory in that the
resultant model is derived in part from theory and in part from a respecification based on
the analysis of model fit’. The modification indices have been used here to delete
redundant or irrelevant items. Out of the initial 33 items, 11 of them have been deleted.
After these respecifications, all first-order models exhibited good fit, as detailed Table 14.

Table 14: First order CFA results – Calibration sample

Online Brand Engagement OBC engagement

Prior to respecifications

Affective Cognitive Behav. Affective Cognitive Behav.

Chi-square 152.57 127.58 340.38 127.84 210.57 331.00
p-value 0.00 0.00 0.00 0.00 0.00 0.00
df 27.00 14.00 876.00 27.00 19.00 87.00
CFI 0.94 0.93 0.91 0.95 0.90 0.94
TLI 0.93 0.84 0.90 0.93 0.85 0.93
RSMEA 1.20 1.19 1.14 1.33 2.13 1.11
After respecifications
Chi-square 18.94 11.85 56.57 10.37 6.92 71.28
p-value 0.01 0.07 0.01 0.25 0.32 0.00
df 7.00 6.00 31.00 8.00 6.00 32.00
CFI 0.99 0.99 0.98 0.99 0.99 0.98
TLI 0.98 0.99 0.98 0.99 0.99 0.98
RSMEA 0.08 0.06 0.06 0.04 0.03 0.07

For the second-order CFA with the calibration sample, modifications based on the first-
order models are kept; however the dimensions (affective, cognitive and behavioural) have
been added as second-order factors and correlated with each other. No further
 177

modifications have been made to these models, which had exhibited good fit from the
onset. The online brand engagement second-order model shows adequate fit with a χ2 =
473.17 (p = 0.00) and 195 degrees of freedom. The RMSEA is 0.08, which denotes
satisfactory fit, the CFI equals 0.95 and TLI 0.94, which are indications of good fit. The
OBC engagement model does not perform as well in terms of fit, with a χ2 = 744.65 (p =
0.00) and 204 degrees of freedom. The RMSEA is 0.10, which is at the limit of the
advocated guidelines, and the CFI and TLI respectively reaching 0.91 and 0.90, which
indicate only acceptable fit. Furthermore, the AMOS results indicate that there is an issue
of high collinearity between the affective dimension and the two other dimensions. In
order to maintain the equivalence between the two consumer engagement scales, a decision
is made not to modify the OBC scale. Rather, further tests to assess the validity and
reliability of the scale are performed on the validation sample to decide whether this
performance of the scale is worrisome.

CFAs on validation sample

A second set of CFAs has been performed on the validation sample (n2 = 224) to provide
further support for the models developed with the CFA on the calibration sample.
Following the approach in the previous section, the first-order sub-dimensions are first
assessed in terms of reliability and validity. Once these criteria are secured at the first-
order level, a second-order analysis is performed. The online brand engagement model’s
χ2 is 326.10 (p =0.00) with 184 degrees of freedom. RMSEA is 0.06, which shows
adequate fit, CFI is 0.97 and TLI is 0.96, again displaying good fit of the model. The OBC
engagement model’s χ2 is 438.04 (p =0.00) with 186 degrees of freedom. RMSEA is 0.07;
CFI 0.96 and TLI 0.95, which denotes good fit and temper the issue associated with the
calibration sample’s borderline values. The item loadings to their constructs on the
validation sample range from 0.80 to 0.99, as shown in Table 15 below.
 178

Table 15: First order CFA results – Validation sample

Online brand engagement OBC engagement

Latent factors and items
St Loading t-value St Loading t-value

Alpha= 0.94, AVE= 0.79, Alpha= 0.93, AVE= 0.83,

Enthusiasm
CR= 0.92 CR= 0.94
I feel enthusiastic about the brand 0.88 17.78 0.93 19.85
The brand makes me enthusiastic 0.89 18.15 0.93 20.98
I am heavily into this brand Deleted CFA
I am interested in anything about this
0.90 17.08 0.87 21.78
brand
I find this brand interesting Deleted CFA
Alpha= 0.95, AVE= 0.88, Alpha= 0.94, AVE= 0.85,
Enjoyment
CR= 0.96 CR= 0.94
I enjoy interacting with the brand Deleted CFA
When interacting with the brand, I feel
0.97 22.6 0.91 24.04
happy
I get pleasure from interacting with the
0.99 23.69 0.92 24.8
brand
Interacting with the brand is like a treat
0.86 23.04 0.93 24.52
for me
Alpha= 0.93, AVE= 0.87, Alpha= 0.97, AVE= 0.94,
Attention
CR= 0.93 CR= 0.97
I pay a lot of attention to the brand Deleted CFA
Things related to the brand grab my
Deleted CFA
attention
I spend a lot of time thinking about the
0.92 23.01 0.97 35.54
brand
I make time to think about the brand 0.94 24.85 0.97 32.64
Alpha= 0.96, AVE= 0.87, Alpha= 0.98, AVE= 0.88,
Absorption
CR= 0.96 CR= 0.96
When interacting with this brand, I
0.94 23.86 0.94 29.9
forget everything else around me
Time flies when I am interacting with
0.96 25.01 0.96 33.08
this brand
When I am interacting with this brand, I
0.92 27.14 0.94 42.46
get carried away
When interacting with this brand, it is
0.90 25.16 0.95 37.18
difficult to detach myself
Alpha= 0.94, AVE= 0.83, Alpha= 0.95, AVE= 0.88,
Sharing
CR= 0.94 CR= 0.96
I reply to the questions of the brand page
Deleted CFA
managers
I share my opinion with them Deleted CFA
I share my experiences with them Deleted CFA
I share my ideas with them 0.90 20.44 0.92 23.97
I share interesting content with them 0.93 22.95 0.97 28.02
I help them 0.90 19.56 0.92 29.56
 179

Online brand engagement OBC engagement

Latent factors and items
St Loading t-value St Loading t-value
Alpha= 0.90, AVE= 0.72, Alpha= 0.90, AVE= 0.76,
Learning
CR= 0.88 CR= 0.90
I ask them questions 0.89 13.83 0.85 16.36
I seek ideas, or information from them 0.84 16.02 0.90 18.09
I seek help from them 0.81 18.52 0.87 17.52
I learn from the content they provide Deleted Item Purification
Alpha= 0.92, AVE= 0.74, Alpha= 0.95, AVE= 0.82,
Endorsing
CR= 0.92 CR= 0.95
I show support to what they say or do Deleted CFA
I share their content to my wider
Deleted CFA
network
I promote the brand 0.88 15.59 0.93 22.16
I try to get other interested in the brand 0.89 15.74 0.93 22.75
I actively defend the brand from its
0.86 15.15 0.87 19.11
critics
I say positive things about the brand to
0.80 16.2 0.89 21.05
other people

Construct reliability has first been assessed Cronbach’s Alpha, similarly to the procedure
applied to the calibration sample. Similar to Cronbach’s Alpha and using the same
benchmark of 0.70, the composite reliability (CR) measures the constructs’ internal
consistency (Fornell and Larcker 1981), and is computed using the following formula:

(Sum of standardised loadings)2

CR=
(Sum of standardised loadings)2+Sum of indicators'measurement error

The online brand engagement scale exhibits composite reliability values of 0.92, 0.96,
0.93, 0.96, 0.94, 0.88 and 0.92 respectively for the enthusiasm, enjoyment, attention,
absorption, sharing, learning and endorsing sub-dimensions, or first-order factors. The
OBC model shows equality good composite reliability with 0.94, 0.94, 0.97, 0.97, 0.96,
0.90 and 0.95 for the enthusiasm, enjoyment, attention, absorption, sharing, learning and
endorsing sub-dimensions. Coefficient alphas and composite reliability indexes are all
reported in Table 16. Reliability being a necessary but not sufficient condition, the validity
of the scale is also assessed, focusing on all types of validity.

Construct validity is defined as the extent to which an operationalisation measures the

concept it is supposed to measure (Cook and Campbell, 1979). It identifies whether the
variable is the underlying cause of item covariation (DeVellis, 2012) and is one of the
 180

main indicators of the overall quality of a study. Construct validity can be broken down
into four sub-categories: (1) Content validity, (2) Criterion-related validity, (3) Construct
validity and (4) Nomological validity.

Content validity is ensured with a strong theoretical foundation, certifying that the items
reflect the content encompassed by the target construct. In addition to a precise
conceptualisation (see chapter 3), following Netemeyer et al. (2003), content validity is
further supported by the experts’ panel review, pre-test and pilot phases detailed in this
chapter. The scale is therefore considered content valid as the items are representative of
the construct’s domain and reflective of its difference facets.

Criterion-related validity concerns convergent and discriminant validity. Convergent

validity has been assessed using the average variance extracted (AVE) and computed for
each dimension and sub-dimension of the scale. AVE measures the percentage of total
variance of the data accounted for by each dimension or in other words, the average
variance that the latent variable can explain of all its indicators. Fornell and Larcker (1981)
advocate the AVE not to be lower than 0.50. The formula used to compute the AVE is:

Sum of squared standardised loadings

𝐴𝑉𝐸 =
Sum of squared standardised loadings + Sum of indicator measurement error

The online brand engagement scale shows AVE values for the sub-dimensions of 0.79
(enthusiasm), 0.88 (enjoyment), 0.87 (attention), 0.87 (absorption), 0.82 (sharing), 0.72
(learning) and 0.74 (endorsing), which are well above 0.50. The OBC engagement scale
also exhibits excellent AVE for all its sub-dimensions. They are performing well with
AVEs of 0.83 (enthusiasm), 0.85 (enjoyment), 0.94 (attention), 0.90 (absorption), 0.88
(sharing), 0.76 (learning) and 0.82 (endorsing). AVEs are reported in Table 16.

Evidence of discriminant validity among the dimensions, and sub-dimensions of the

online consumer engagement scales has been provided by three separate tests. Using three
different methods is deemed necessary due to the length and dimensional complexity of the
scales, which renders them more prone to discriminant validity issues.

The first method used to test for discriminant validity is the correlation test. By examining
correlations between all pairs of factors, one can identify if two factors are independent, if
their pairwise correlation is smaller than one (Anderson and Gerbing, 1988), which is the
case for all pairs of factors at the sub-dimension level, as per Table 16.
 181

The second method compares the AVE of each factor with each squared pairwise
correlation including that factor. For discriminant validity to exist, the AVE of each factor
must be larger than each squared pairwise correlation involving this factor (Fornell and
Larcker, 1981). This test is first done at the first-order level to verify discriminant validity
across the 7 sub-dimensions. As displayed in Table 16, AVEs are overall greater than any
of the two squared pairwise correlations associated with them.

Table 16: Validity- Validation sample

Online brand engagement

Sub-dimensions 1 2 3 4 5 6 7
1. Enthusiasm 0.79 0.67 0.50 0.35 0.19 0.12 0.59
2. Enjoyment 0.82 0.88 0.52 0.44 0.28 0.17 0.48
3. Attention 0.71 0.72 0.87 0.69 0.34 0.15 0.44
4. Absorption 0.59 0.66 0.83 0.87 0.32 0.18 0.42
5. Sharing 0.44 0.53 0.58 0.57 0.82 0.69 0.36
6. Learning 0.34 0.41 0.39 0.42 0.83 0.72 0.25
7. Endorsing 0.77 0.69 0.66 0.65 0.60 0.50 0.74
OBC engagement
Sub-dimensions 1 2 3 4 5 6 7
1. Enthusiasm 0.83 0.56 0.45 0.25 0.29 0.19 0.59
2. Enjoyment 0.75 0.85 0.61 0.67 0.49 0.32 0.48
3. Attention 0.67 0.78 0.94 0.69 0.38 0.24 0.40
4. Absorption 0.50 0.82 0.83 0.90 0.46 0.29 0.31
5. Sharing 0.54 0.70 0.62 0.68 0.88 0.58 0.45
6. Learning 0.44 0.57 0.49 0.54 0.76 0.76 0.36
7. Endorsing 0.77 0.69 0.63 0.56 0.67 0.60 0.82
NB: Bottom half = correlations; Top half = Squared correlations; Diagonal = AVE. For
the sub-dimensions, the only correlations to consider are those with sub-dimensions from
the same dimension, which are shaded.

A chi-square difference test is a third way to account for discriminant validity. Following
this method, models with fewer dimensions are compared against models with more
dimensions (Anderson and Gerbing, 1988). Model A represents the default baseline model
with all sub-dimensions. Models B, C, D, E, F and G, are created, with respectively 6, 5, 4,
3, 2 and 1 dimensions to compare with model A, in which all 7 dimensions of engagement
are kept. Dimension reduction is achieved by setting the correlation between two
dimensions to 1. For instance, in model B, the covariance between Enthusiasm and
Enjoyment is set to 1. If the chi-square for model A is significantly lower than all chi-
squares for models with fewer dimensions, discriminant validity is achieved (Bagozzi and
 182

Phillips, 1982). In the AMOS output, CMIN represents the chi-square. Since all p-values
are below .05, this indicates that the chi-square of model A is significantly lower than all
other nested models chi-squares, therefore indicating discriminant validity between sub-
dimensions. The results of this test are detailed in Table 17.

Table 17: Chi-Square test sub-dimensions – Validation sample

Online brand engagement: Assuming model A to be correct

p-
Model df CMIN NFI IFI
value
Model B: corr BENT and BENJ=1 1 479.45 0.00 0.03 0.03
Model C: corr BENT- BENJ=1; corr BAT –BABS =1 2 611.10 0.00 0.04 0.04
Model D: corr BENT- BENJ=1; corr BAT –BABS
3 744.75 0.00 0.05 0.05
=1; corr BSH-BLE=1
Model E: corr BENT- BENJ=1; corr BAT –BABS
5 1699.66 0.00 0.10 0.11
=1; corr BSH-BLE-BEND=1
Model F: corr BENT-BENJ-BAT –BABS =1; corr
9 2661.17 0.00 0.16 0.17
BSH-BLE-BEND=1
Model G: corr BENT-BENJ-BAT –BABS-BSH-
21 4591.88 0.00 0.28 0.29
BLE-BEND=1
OBC engagement: Assuming model A to be correct
p-
Model df CMIN NFI IFI
value
Model B: corr OENT and OENJ=1 1 25.50 0.00 0.00 0.00
Model C: corr OENT- OENJ=1; corr OAT –OABS
2 69.88 0.00 0.00 0.00
=1
Model D: corr OENT- OENJ=1; corr OAT–OABS=1;
3 112.47 0.00 0.01 0.01
corr OSH-BLE=1
Model E: corr OENT- OENJ=1; corr OAT–OABS
5 115.44 0.00 0.01 0.01
=1; corr OSH-OLE-OEND=1
Model F: corr OENT-OENJ-OAT–OABS =1; corr
9 279.23 0.00 0.01 0.01
OSH-OLE-OEND=1
Model G: corr OENT-OENJ-OAT–OABS-OSH-
21 561.80 0.00 0.03 0.03
OLE-OEND=1

In order to proceed to the second-order scale validation procedures, a manipulation of the

first-order scale is necessary. In order to manage the dimensional complexity of the scale
and its large number of items, sub-dimensions are computed as summated or aggregate
scores following Yoo and Donthu’s (2001) approach. Summating the raw score of the 22
items is deemed incorrect since the items do not load equally on each sub-dimensions.
Observed items’ causal paths standardised regression weights are therefore used as weights
to create an aggregate score for each sub-dimensions (e.g. all enthusiasm items are used to
 183

create an enthusiasm aggregate score). The weight of an item is calculated as the fraction
of the path estimate of that dimension over the sum of the other relevant path estimates.
For example, the weight of the first enthusiasm item on the overall enthusiasm aggregate
value for online brand engagement is 0.332, which derives from 0.88/(0.88+0.89+0.88).
Subsequently, the summated enthusiasm scale is computed as:

Aggregate value of enthusiasm items =0.332*BENT1+0.336*BENT2+0.332*BENT4

CFAs have been calculated at the dimension level where the sub-dimensions have become
items, and the dimensions of which they are reflective, first-order factors. The online brand
engagement CFA performs adequately with a χ2 of 26.78 (p =0.003) with 10 degrees of
freedom. RMSEA is 0.08, the CFI is 0.99 and TLI is 0.98. The OBC engagement model
exhibits a χ2 of 15.03 (p =0.053) with 8 degrees of freedom, and RMSEA of 0.06, a CFI
equal 0.99 and a TLI of 0.98. The item loadings to their constructs on the validation
sample ranged from 0.71 to 1.00, as shown in Table 18 and they all are significant.

Table 18: CFA second order – Validation sample

Latent factors/ Online brand engagement OBC engagement

items (aggregate
scores) St Loading t-value St Loading t-value
Alpha= 0.86, AVE= 0.76, CR= Alpha= 0.83, AVE= 0.76, CR=
AFFECTIVE 0.86 0.84
Enthusiasm 0.94 14.60 0.74 13.74
Enjoyment 0.80 15.20 0.96 15.89
Alpha= 0.88, AVE= 0.78, CR= Alpha= 0.90, AVE= 0.82, CR=
COGNITIVE 0.87 0.90
Attention 0.89 16.49 0.88 19.97
Absorption 0.87 15.28 0.93 21.54
Alpha= 0.93, AVE= 0.76, CR= Alpha= 0.93, AVE= 0.82, CR=
BEHAVIOURAL
0.91 0.93
Sharing 0.89 24.93 0.95 40.63
Learning 0.71 14.77 0.74 16.39
Endorsing 1.00 18.52 1.00 17.20

Scale reliability is achieved at the dimensions level, with Cronbach’s Alpha all largely
above the cut-off value of 0.70, and coefficients of reliability (CR) also ranging from 0.76
to 0.93. Convergent validity is ensured thanks to AVE values above 0.50 for all three
dimensions of each scale. Specifically, for the online brand engagement scale, the AVE is
0.76 for the affective dimension, 0.78 for the cognitive dimension, and 0.76 for the
 184

behavioural dimension. These values for the OBC engagement scale are namely 0.76, 0.82
and 0.82.

In terms of discriminant validity, evaluating the paired correlations for the dimensions
proved to be adequate for all bar one dimension, the affective dimension of the OBC
engagement scale, which exhibits a value of 1, and therefore a squared correlation of 1,
which is above the cut-off value advocated by Anderson and Gerbing (1988). Accordingly,
the second test for discriminant validity also yields the same results. Comparing the AVE
with the squared pairwise correlations shows that the AVEs for each factor are greater than
all the squared pairwise correlations involving this factor, except for the affective
dimension of the OBC engagement scale. Specifically, the affective dimension’s AVE
equals 0.76, which is below its squared pairwise correlation with the cognitive dimension
(1.00) and with the behavioural dimension (0.88).

Table 19: CFA second order validity – Validation sample

Online brand engagement

Dimensions 1 2 3
1. Affective 0.76 0.67 0.38
2. Cognitive 0.82 0.78 0.45
3. Behavioural 0.62 0.67 0.76
OBC engagement
Dimensions 1 2 3
1. Affective 0.76 1.00 0.88
2. Cognitive 1.00 0.82 0.72
3. Behavioural 0.94 0.85 0.82

In order to evaluate whether this violation of discriminant validity is worrisome, a further

test of discriminant validity is performed: the chi-square difference test. Because the
existence of high correlations indicates that the offending factor (the affective dimension)
should either be deleted or merged with another factor, a comparison of the baseline model
with models with fewer dimensions is performed through the chi-square difference test
whereby models with fewer dimensions are compared against models with more
dimensions (Bagozzi and Phillips, 1982; Anderson and Gerbing, 1988).

Models B, C, D and E have been created, with respectively 2, 2, 2, and 1 dimensions to

compare with model A, in which all 3 dimensions of engagement are kept and their
 185

covariances left unconstrained. Dimension reduction has been achieved by setting the
correlation between two dimensions to 1. In model B, the covariance between the affective
dimension and the cognitive dimension is set to 1, in model C, this is done between
cognitive and behavioural, in model D between affective and behavioural, and in model E
all covariances are constrained to 1. If the chi-square for model A is significantly lower
than all chi-squares for models with fewer dimensions, discriminant validity is achieved
(Bagozzi and Phillips, 1982). In the AMOS output, CMIN represents the chi-square. Since
all p-values are below 0.05, this indicates that the chi-square of model A is significantly
lower than all other nested models chi-squares, therefore indicating discriminant validity
between sub-dimensions and ensuring that all dimensions are rightfully needed to represent
the concept.

Table 20: Chi Square test – Validation sample

Online brand engagement: Assuming model A to be correct

Model df CMIN p-value NFI IFI RFI TLI
Model B: Corr Aff-Cog =1 1 40.09 0.00 0.03 0.03 0.05 0.05

Model C: Corr Cog-Behav =1 1 15.40 0.00 0.01 0.01 0.02 0.02

Model D: Corr Aff-Behav=1 1 16.42 0.00 0.01 0.01 0.02 0.02

Model E: Corr Aff-Behav-Cog = 1 3 40.36 0.00 0.03 0.03 0.04 0.04

OBC engagement: Assuming model A to be correct

Model df CMIN p-value NFI IFI RFI TLI
Model B: Corr Aff-Cog =1 1 48.17 0 0.03 0.03 0.05 0.04

Model C: Corr Cog-Behav =1 1 22.14 0 0.01 0.01 0.02 0.02

Model D: Corr Aff-Behav=1 1 37.93 0 0.02 0.02 0.04 0.04

Model E: Corr Aff-Behav-Cog = 1 3 53.88 0 0.03 0.03 0.04 0.04

Another form of validity, which needs to be verified, is construct validity. Since the study
develops two mirrored scales of engagement, assessing the correlations between the two
scales provides an indication whether they are both tapping into the same underlying
engagement concept. A correlation of 0.97 between the online brand engagement and the
OBC engagement scales is supportive of construct validity. In other words, a high
correlation between the scales indicates that they both measure the same construct, albeit
 186

for a different object. A last measure of validity of a scale concerns its relationship with
other constructs, and is called nomological validity (Netemeyer et al., 2003). This aspect of
the scale validity will be addressed thanks to the hypothesis testing section, in which the
two consumer engagement constructs are causally linked with other theoretically related
constructs.

5.2.6. Norm

The last stage of the scale development process advocated by Churchill (1979) is the
development of norms. The aim of this procedure is to allow the assessment of the position
of a relevant unit of investigation in terms of the score it achieves on the scale items. This
is particularly useful to know if the level of engagement generated by brand alpha is
higher, lower or similar to the average of all other brands. In this case, it is relevant to
categorise values according to the product categories. Indeed, Table 10 showed that few
OBC studies compared brands belonging to different product categories, and a large
amount of them focused on durable goods, especially automotive goods. Consequently,
means are computed for each item of the mirrored scales. Similarly to Lytle et al. (1998),
these means are broken down by product category, in the aim to compare the performance
of each product category with respect to the others. Table 21 offers the detail of these
figures.

The highlighted values represent the hightest item means across product categories,
whereas the values in bold represent the lowest scores. The results show that, overall,
services and durable goods achieve the highest cross-category means, whereas retail
brands have the lowest levels of engagement across items. More specifically, service
brands exhibit the highest levels of sharing and learning, both with the brand and with the
OBC. Services also have the highest values for attention and absorption with the
community. Durable goods, on the other hand, rank the highest on all the affective and
cognitive brand engagement items, as well as most of the OBC affective items. Durables
also perform particularly well on both brand and OBC engagement “endorsing” items.
Aside from a few items, the product category encompassing all the lowest engagement
means is the retail category, both for brand and OBC engagement. On average, brand
engagement scores higher values than OBC engagement. Additionally, the sub-dimensions
achieving the highest means are enthusiasm and endorsing whereas cognitive dimensions
are on the low end.
 187

Table 21: Item means by product category

Food and Fashion and Durable

Product categories Technology Services Travel Retail Entertainment Others Total
Beverage Beauty Goods
Sub
Items Brand OBC Brand OBC Brand OBC Brand OBC Brand OBC Brand OBC Brand OBC Brand OBC Brand OBC Brand OBC
dimension
1 5.49 4.27 5.00 3.63 4.60 4.28 4.73 3.89 5.22 4.18 6.11 4.90 4.08 2.67 5.31 4.52 5.41 4.12 5.25 4.22
Enthusiasm 2 5.22 3.77 4.38 2.94 4.50 4.02 4.48 3.50 4.94 3.66 6.02 4.49 3.92 2.67 5.09 4.29 4.82 4.00 5.00 3.81
3 4.69 4.00 4.00 3.00 4.22 4.08 4.50 3.73 4.34 3.43 5.86 4.32 3.50 2.42 4.39 3.91 4.88 4.03 4.62 3.84
1 3.95 2.62 3.06 2.31 3.70 3.48 3.95 2.79 3.42 2.29 5.08 3.37 2.83 1.92 4.09 3.04 4.06 2.41 3.95 2.75
Enjoyment 2 4.01 2.82 3.13 2.06 3.75 3.52 3.94 3.03 3.47 2.55 5.00 3.74 3.17 1.92 4.13 3.17 4.06 2.53 3.98 2.95
3 3.39 2.36 2.35 2.09 3.45 3.02 3.58 2.79 3.01 2.23 4.65 3.30 2.58 1.83 3.22 2.61 3.18 2.35 3.43 2.57
1 3.33 2.56 3.00 2.37 3.70 3.30 3.30 2.68 2.88 2.15 4.52 3.00 2.75 1.83 3.20 2.58 2.82 2.65 3.35 2.60
Attention
2 3.16 2.49 2.81 2.31 3.33 3.08 3.17 2.56 2.72 2.04 4.06 2.86 2.33 2.00 2.92 2.53 2.59 2.53 3.13 2.50
1 2.55 2.03 2.31 1.81 3.13 2.93 2.66 2.31 1.97 1.86 3.38 2.06 2.17 1.75 2.46 2.12 2.18 1.89 2.57 2.11
2 2.94 2.16 2.44 1.94 3.20 2.97 2.90 2.33 2.47 1.93 3.88 2.47 2.25 2.00 3.25 2.33 2.71 2.08 2.98 2.25
Absorbtion
3 2.80 2.01 2.44 1.94 3.22 2.78 2.86 2.29 2.46 1.89 3.85 2.17 2.50 1.58 3.00 2.20 2.47 2.00 2.89 2.12
4 2.42 1.92 2.12 1.75 3.17 2.66 2.63 2.17 2.22 1.73 3.29 2.13 1.83 1.75 2.62 2.03 2.06 1.82 2.56 2.01
1 3.37 2.95 2.75 1.95 3.80 3.92 3.39 3.08 2.72 2.55 3.39 3.28 1.75 1.75 3.34 3.00 2.76 2.38 3.25 2.96
Sharing 2 3.15 2.73 2.25 2.03 3.97 3.90 3.17 2.92 2.53 2.53 3.20 3.20 2.17 1.92 3.27 2.90 2.41 2.06 3.08 2.83
3 2.84 2.57 2.63 2.11 3.66 3.63 2.76 2.73 2.44 2.33 3.23 3.39 1.75 1.67 3.13 2.90 2.29 2.12 2.85 2.70
1 3.02 2.76 3.25 2.57 3.85 3.38 3.36 3.02 2.98 2.82 3.23 3.22 2.08 2.00 3.47 3.15 2.00 1.97 3.17 2.91
Learning 2 3.25 3.00 3.50 2.54 4.40 4.20 3.73 3.31 3.42 3.18 3.74 3.45 2.83 2.33 3.94 3.53 2.88 2.75 3.55 3.23
3 2.68 2.44 3.44 2.63 4.08 3.73 3.57 3.15 2.69 2.62 2.91 2.70 1.58 2.25 3.36 2.94 2.35 2.33 3.03 2.76
1 4.53 3.86 3.97 2.45 4.40 3.70 3.48 2.89 3.69 3.18 4.88 4.12 3.33 2.17 3.80 3.43 3.82 3.58 4.09 3.46
2 4.68 3.79 4.25 2.70 4.35 3.60 3.39 2.87 3.48 3.08 4.39 3.80 2.58 1.92 3.89 3.36 4.00 3.38 4.05 3.38
Endorsing
3 3.94 3.30 3.63 2.63 3.70 3.55 3.20 2.77 2.81 2.62 4.55 3.50 2.25 2.17 3.34 3.12 2.82 2.94 3.54 3.07
4 5.19 4.05 4.38 2.93 4.02 3.87 4.23 3.30 4.49 3.35 5.73 4.54 4.25 2.83 4.72 3.93 4.35 3.71 4.77 3.77
 188

5.3. Scales for existing constructs

This section deals with the scales chosen for existing constructs, which were adapted from
past studies. It provides a rationale for the specific scales chosen for each construct based
on conceptual fit, reliability, validity and evidenced generalisability of the scales. The
choice of scales for the concepts of OIP, attitude toward community participation, product
involvement, brand trust, brand commitment and brand loyalty is presented below.

5.3.1. Online interaction propensity

The OIP scale is derived from Wiertz and de Ruyter’s (2007) work on firm-hosted
commercial online communities. The scale was developed following Churchill’s (1979)
paradigm, using a mix of quantitative and qualitative data collection phases, in total 5
studies performed on online and offline samples, which establish the strong psychometric
properties of the scale (composite reliability of 0.96; average variance extracted of 0.85).
The resulting scale is a 4-item instrument, measured with a seven-point Likert type
response format, with ‘strongly disagree’ and ‘strongly agree’ as anchors.

5.3.2. Attitude toward OBC participation

Attitude toward OBC participation is measured based on Bagozzi and Dholakia’s (2006)
semantic differential scale which asks respondents to anchor their attitudes on the
following four 7-point scales: (1) foolish/wise; (2) harmful/beneficial; (3) bad/good) and
(4) punishing/rewarding. The scales achieved a reliability of 0.94 and 0.88 respectively for
the two type of brand communities involved in the project. Its wording is adapted in this
study to be understood by the Facebook brand page members.

5.3.3. Product involvement

Given the complexity and amount of attributes of the involvement concept, a large number
of ways to measure it exist, taking different conceptual angles. This section reviews the
most prominent ones. Lastovicka and Gardner (1979) developed the first scale dedicated to
this measure, focusing on normative importance (how important a product class is to an
individual’s values) commitment (the binding of an individual to his/her brand choice), and
familiarity. The 22-items scale uses a 7-point Likert scale and has the advantage of having
been developed on a sample of 14 different product categories. However, in addition to its
 189

length, a major reason for deeming it inappropriate to this study is its inclusion of
commitment, which we defined as conceptually distinct from involvement. This same
concept/measurement issue is found in Traylor and Joseph’s (1984) scale. Although it
achieves high reliability and validity, its items are strongly reminiscent of social
identification, or congruity, rather than involvement (e.g. ‘when other people see me using
this product, they form an opinion of me’).

Bloch (1981) later devised a scale intended to measure four dimensions of product
involvement, which are mainly cognitive, namely (1) knowledge, (2) opinion leadership,
(3) interest and (4) information search. This scale suffers from a strong contextual bend, as
is has been developed and later used mainly in the context of automobile (Richins and
Bloch, 1991; Richins et al.,1992). This industry specificity is also strongly reflected in the
wording of items, which creates a concern for the need of an important adaptation of the
scale to fit the purpose of this study.

One of the most widely used scale of involvement is Zaichowski’s (1994) Personal
Involvement Inventory, which is adaptable to product, advertisement and purchase
situations. The 20-item scale uses a 7-point semantic differential scale with bipolar
adjectives as anchors and was vastly used by researchers (e.g. Flynn and Goldsmith, 1993;
Warrington and Shim, 2000) due to its wider range of applicability, and high reliability and
validity. It was however later criticised for being too lengthy, and having a problem with
discriminant and construct validity (McQuarrie and Munson, 1992).

The scale adopted in this study is Laurent and Kapferer’s (1985) Consumer Involvement
Profile scale. It measures 5 dimensions of involvement: (1) the perceived importance of the
product, (2) the pleasure or hedonic value derived from the product, (3) the sign, or
symbolic value attributed to the product, (4) the risk probability associated with a potential
mispurchase, and (5) the risk importance associated with a mispurchase. The reason for
using this scale is threefold. Firstly, it has been developed using a sample of 14 different
product categories, ensuring a better generalisability than other scales, as shown by its
broad usage (e.g. Mittal and Lee, 1989; Havitz and Howard, 1995). Secondly, its
dimensions are inclusive of both enduring and situational involvement, which is in line
with the scope of this research. Thirdly, its conciseness and clarity work favourably toward
its inclusion in a web survey of Facebook page members. With Cronbach Alphas ranging
from 0.72 and 0.90 on its different dimensions, this scale also proved to be highly reliable.
 190

5.3.4. Brand trust

It is apparent from the literature on brand trust that research is scarce in the measurement
of this concept, and it has been qualified as hard to conceptualise and measure (Matzler, et
al., 2008). Organisational research has produced a number of measurements of trust in
various contexts and focused on various trust objects. Probably one of the best known
study on trust is Morgan and Hunt’s (1994), which uses an 8-item Likert scale. Garbarino
and Johnson (1999) later measure organisational trust with the intent to see how important
it is for consumers, based on their relational orientation. Trust has also been studied in
online contexts in the form of e-trust, or trust with a website, loosely building from
Garbarino and Johnson’s (1999) scale.

In the branding literature, trust has been measured by Hess (1995), focusing on 3
dimensions of trust, namely: (1) Altruism; (2) Honesty and (3) Reliability, with an 11-item
Likert scale. This scale is however not used in this study, in favour of Chaudhuri and
Holbrook’s (2001) scale of brand trust, using a more parsimonious 4-item, 7-point Likert
scale with anchor 1= very strongly disagree, and 7= very strongly agree. This scale
achieved a reliability index of 0.81. A subsequent 3-item scale was adapted from this
version by the same authors (Chaudhuri and Holbrook, 2002); however, it achieves a lower
reliability index; probably because of the deletion of a determinant item. The 4-item scale
is considered as best suited to this study firstly because of its successful adaptations in
other branding studies (e.g. Mazler et al., 2008; Marzocchi et al., 2013). Its wide usage is
complemented by its focus on trust with a brand, rather than another organisational focus,
its conciseness, and high psychometric values.

5.3.5. Brand commitment

The measurement of commitment in organisational sciences is a problematic issue, guided

by the lack of consensus on its dimensionality. A large stream of the literature concerned
with the measurement of commitment threats it as a unidimensional behavioural variable,
or bi-dimensional, consisting of attitudes and behaviours (e.g. Warrington and Shim, 2000;
Eisingerich and Rubera, 2010; De Wulf et al., 2013). A further problem with the uni-
versus multidimensionality of commitment it that, some scales’ dimensionality seems
inconsistent with their conceptualisation, rendering their use subject to question. For
instance, Beatty and Khale (1988) define commitment as an attitudinal concept, yet their 3-
 191

item scale is reflective of behavioural intentions rather than attitudes (e.g. ‘When another
brand is on sale, I will generally purchase it rather than my usual brand’).

Semantic differential scales have also been used to measure commitment. Taking this
particular approach to the measurement of commitment, Ahluwalia et al. (2013) propose a
semantic differential scale with 4 items: (1) Good/bad; (2) Beneficial/Harmful; (3)
Desirable/Undesirable and (4) Nice/Awful. This scale allows expressing a value judgment
of a brand, however not to measure a preference for said brand.

Another line of measures of commitment embraces the idea that commitment is attitudinal
only (often relating it to attitudinal loyalty). One study taking such a perspective is
Pritchard et al.’s (1999). They conceptualise and measure commitment as resistance to
change, based on a past preference. It is indicative of attitudinal preference, but also
indicates cognitive aspects of this preference in its potential re-evaluation. In line with the
attitudinal uni-dimensionality, Jacoby and Chesnut (1978), measure commitment using a
4-item Likert scale, which achieved a high reliability score. It has however been criticised
as encompassing measures of intentional behaviour rather than attitudinal preference.
Other validated commitment scales such as Allen and Meyer’s (1990) were also discarded
despite their popularity, based on the grounds that they display a strong overlap with other
relational constructs used in this study, including dimensions of consumer engagement.

After a thorough review of the commitment scaling literature, this study takes the stance
advocated by El-Manstrly and Harrison (2013), which views brand commitment as an
attitudinal dimension of loyalty. After reviewing the literature on loyalty, they developed a
loyalty scale with data from Scottish consumers of a company in the service sector.
Commitment is equalled to attutidinal loyalty and, using a 5-item, 7-point Likert scale it
achieves high reliability (Cronbach’s Alpha = 0.88) and convergent validity (Variance
Extracted = 0.65). The 5-item scale is reviewed for the purpose of this research, and 3
items out of the initial five are kept. The two items ‘Compared with other service
providers, I am happy with the services x provides’ and ‘I am usually pleased with my
purchase decisions from x’ were not used, as they were deemed to tap into brand
satisfaction rather than commitment. The remaining three items are consistent with the
affective and attitudinal view of brand commitment as defined in chapter 3.
 192

5.3.6. Brand loyalty

Brand loyalty has been the subject of many measurement exercises, all of which are rooted
in the way the construct is defined. This study takes a behavioural approach to loyalty, and
views it as ‘repeat purchase behaviours’, considering it to be conceptually distinct from
any attitudinal dimension. Keeping in line with this conceptualisation, this study focuses
on measurement that encapsulates this behavioural dimension. Since attitudinal loyalty
(brand commitment) and behavioural loyalty (repeat purchase) are two distinct conceptual
constructs within the scope of this study, this study also considers them to be operationally
distinct.

Behavioural loyalty can be measured as retention, lifetime duration, usage, share of wallet
and cross buying. More specifically, usage can be measured as the number of repurchase
visits, amount of spending or number of transactions (Kumar et al., 2013). Scales of brand
loyalty often encompass attitudinal and behavioural aspects of loyalty under the same
scale. This is the case with Chaudhuri and Holbrook (2001) who, based on Jacoby and
Chesnut (1978) measure purchase loyalty with two items along with attitudinal loyalty
with two items. This scale was also used by Marzocchi et al. (2013) in the context of brand
communities. Similarly, Pritchard et al. (1999) use a composite measure with four
attitudinal items and two behavioural items, which are intended to specifically quantify the
amount of purchase occurrences within a specific timeframe. These scales however suffer
from being framed in a multidimensional approach to loyalty, and as such, only grant
limited importance to the behavioural items.

For this reason, it is deemed essential to use an existing scale that is entirely dedicated to
the measurement of repeat purchasing behaviour, such as the one developed by Odin et al.
(2001). This scale is composed of 4 items capturing past and future purchasing behaviour
of one specific brand within a product category, and is therefore the one most adapted to
our context of investigation and conceptual approach of brand loyalty. This scale also
evidenced high reliability with a Cronbach’s Alpha of 0.96.
 193

5.4. Summary

The first section of the chapter focused on the development of a reliable and valid scale of
consumer engagement in the context of OBC in order to answer the first research question.

Through a series of steps following strict psychometric processes and tests, two mirrored
scales of consumer engagement were created: one with the OBC as a partner, the other one
with the brand. The duality of engagement partners as well as the dimensions and sub-
dimensions of consumer engagement were proposed in the conceptual framework, and
further evidenced in the 25 interviews. A series of iterative item generation steps ensued,
and the content validity of these items was ensured with the help of academic experts.
Using a final pool of 2x35 items, data were collected on a sample of Facebook page
members. The items were first purified and then submitted to a two-level CFA analysis.
This analysis was first successfully computed on the calibration sample, and on the
validation sample next, exhibiting adequate goodness of fit indices at both stages and for
both first-order and second-order levels, accounting for the sub-dimensions of the scales.
The two consumer engagement scales also exhibited good measures of convergent,
discriminant and construct validity. Lastly, measurement and structural invariance for the
consumer engagement scales were tested across the English and the French samples,
evidencing complete invariance between the two linguistic contexts.

The chapter also presented the scales that were chosen to tap into all other variables
included in the causal model of this study. The next chapter focuses on the testing of the
hypotheses related to this model, using the scales chosen and developed in this chapter.
 194

Chapter 6: Hypothesis testing

6.1. Introduction

This chapter focuses on addressing research questions 2 and 3. Using the developed
measures of consumer engagement, the relationships with their antecedents and outcomes
is tested in this chapter using SEM procedures. A CFA is computed first to assess the
measurement model’s reliability and validity, followed by statistics regarding the structural
model. The results show that the hypotheses are partially supported.

In order to validate the research questions across different linguistic contexts, all tests
include analyses of the French and English samples in parallel, and further tests of group
invariance are computed at each stage of the process to assess the equivalence of the
model.

6.2. Approach to hypothesis testing

The aim of the SEM detailed in this section is to test the set of hypotheses developed in the
conceptual stages of this research. Figure 18 below offers a reminder of the visual
representation of the hypotheses.
 195

Figure 18: The causal model (2)

This section adopts a two-phase SEM analysis, focusing first on the measurement model to
assess the factor structure, and then the structural model to test the hypothesised links
between the variables and assess the fit of the full structural model with the data (Anderson
and Gerbing, 1988).

For the measurement model, the same logical series of steps as detailed in the scale
development CFAs are applied: model identification, model estimation, goodness-of-fit
evaluation and check of the parameter estimates. Reliability and validity of the models are
subsequently reported for the CFAs and a measurement invariance test proving that the
French and English samples exhibit measurement invariance. The samples used to test the
hypotheses are the English validation sample (n=224) previously used in the consumer
engagement scale development, and the French sample (n=273).

Then, the SEM results are presented and verification or rejection of the hypotheses
examined. Results are presented for the French and English samples concurrently, and a
test performed to account for structural invariance between the samples.
 196

6.2.1. The measurement model

Model fit
A first CFA is run as suggested by Andersson-Cederholm and Gyimothy (2010), including
in the measurement model all hypothesised antecedents and outcomes of online brand and
brand community engagement, as well as the two engagement constructs, using the
developed engagement scales.

The measurement model exhibited for the English model a chi-square of 1741.34 (p =
0.00) with 535 degrees of freedom, CFI = 0.84, TLI = 0.83, which are slightly below the
advocated guidelines, and an RMSEA = 0.06, which is deemed a good representation of
fit. For the French sample, these values are: chi-square of 1973.24 (p = 0.00) with 535
degrees of freedom, CFI = 0.84, TLI = 0.82, which are slightly below the advocated
guidelines, and an RMSEA = 0.06. All standardised loadings are above or close to 0.50
and t-values are all significant (p<0.01) both for the English and the French sample. The
reasons for CFI and TLI measures below the advocated guidelines are largely due to the
nature of the engagement scales included in the model, as lengthy and complex scales are
more difficult to use in models with many variables and may result in redundancy between
closely related items (Ruvio et al., 2008). An overview of the modification indices indeed
indicates high modification values for the covariances of the two engagement scales, which
directly affect the model fit. These values are however natural since the items measure the
same concept for different objects. However, the constructs perform perfectly at their
higher-order levels, showing no evidence of multicollinearity (see Table 21 below).

Configural and metric invariance of the measurement model

Assessing the applicability of framework developed in one context into another context is
important to establish its generalisibality. Following recognised multigroup invariance
tests which use increasingly constraining levels of invariance (Steenkamp and
Baumgartner, 1998), this section aims to first show the configural and metric (or
measurement) invariance of the hypothesis model across the English and French sample.
The AMOS multigroup analysis procedure is used, following (Byrne, 2010) 4 . The
configural invariance determines the similar structure of the measurement instrument re
4
The approach proposed by Byrne (2010) differs slightly from the omnibus test traditionaly run with Lisrel
(which assesses configural, metric, scalar and factor invariance simultaneously, see Steenkamp and
Baumgartner, 1998), as it decomposes the levels of invariance into different models, allowing the researcher
to first ensure less constrained levels of invariance (configural, measurement) before moving to more
restrictive models (with structural invariance). It is therefore a method based on a priori progression rather
than a posteriori relaxing of constraints.
 197

checks for the similarity of the patterns of factor loadings (Steenkamp and Baumgartner,
1998). Since the model fit of the two samples has been assessed separately in the above
paragraph, this section aims to assess the goodness-of-fit of the multigroup sample (that is,
where the French and English sample have been combined). The measurement model
exhibited a chi-square of 3112.304 (p = 0.00) with 1033 degrees of freedom and a chi-
square/df ratio of 3.00 (within the 2.00-3.00 bracket deemed acceptable), CFI = 0.87, TLI
= 0.85, which are slightly below the advocated guidelines, and an RMSEA = 0.06, which is
deemed a good representation of fit. In this multigroup sample, similarly to the French and
English samples separately, the goodness of fit is adequate, no salient factor loadings are
significantly different from zero and the correlations between factors are below the unity
(Steenkamp and Baumgartner, 1998). Since the model fit is achieved for the multigroup
sample, although this model has no constraints, this sets baseline values against which
more restrictive invariance models will be assessed.

Having established the configural invariance, the metric, or measurement invariance aims
to check the similarity of the loadings across samples. In order to test for metrics
invariance, the factor loadings are thus constrained to be the same across countries, using
the automated models function of AMOS (Byrne, 2010) and using the CFI difference value
to test the significance of the invariance (a more recent and practical approach than the chi-
square difference test, according to Cheung and Rensvold, 20025), with a cut-off criterion
of ΔCFI =< 0.01 applied to verify measurement invariance. The online brand engagement
scale exhibits a ΔCFI=0.001 between the configural and constrained model, whereas the
OBC engagement model has a ΔCFI=0.01, which both indicate measurement invariance
for the French and English sample. A further check of the modification indices does not
suggest the need to modify the factorial structure, further validating the measurement
invariance. Having established the configural and measurement invariance of the French
and English samples, this study can proceed to the test of the hypotheses, which is
complemented by further tests of invariance.

5
Note that the chi-square difference test is more strict than the CFI difference test, however hard toapply to
this study’s model given the complexity of the model yielding high degrees of freedom which do not feature
on chi-square distribution tables.
 198

Table 22: SEM measurement model – Standard loadings and t-values

EN Sample FR Sample
Latent factors and items St loading t-value St loading t-value
Product Involvement
This type of product is very important to me 0.97 15.26 0.92 19.54
This type of product matters to me 0.92 16.85 0.99 20.37
Attitude toward OBC participation
Participating in the online community is …Foolish/Wise 0.76 13.97 0.74 13.54
…Harmful/Beneficial 0.90 18.95 0.95 18.88
…Bad/Good 0.91 19.22 0.93 18.37
…Punishing/Rewarding 0.88 18.52 0.80 17.53
Online Interaction Propensity
In general, I like to get involved in online discussions 0.81 10.85 0.81 10.65
I am someone who enjoys interacting with like-minded people online 0.87 19.22 0.86 15.34
I am someone who likes actively participating in online discussions 0.85 11.31 0.86 11.55
In general, I thoroughly enjoy exchanging ideas with others online 0.98 15.43 0.90 16.02
Online Brand Engagement
Affective dimension 0.93 13.54 0.97 15.63
Enjoyment* 0.88 14.53 0.85 16.32
Enthusiasm* 0.86 16.87 0.69 12.72
Cognitive dimension 0.90 13.63 0.89 15.04
Attention* 0.89 17.05 0.88 18.19
Absorption* 0.88 18.22 0.87 16.53
Behavioural dimension 0.94 13.92 0.95 12.32
Sharing* 0.85 10.52 0.78 11.91
Learning* 0.56 8.24 0.70 10.72
Endorsing* 0.67 9.52 0.71 14.52
 199

EN Sample FR Sample
Latent factors and items St loading t-value St loading t-value
OBC Engagement
Affective dimension 1.00 16.52 1.00 16.33
Enjoyment* 0.89 14.27 0.86 163.21
Enthusiasm* 0.82 16.43 0.67 12.89
Cognitive dimension 0.92 17.78 0.94 18.78
Attention* 0.89 20.10 0.90 20.70
Absorption* 0.93 22.46 0.89 15.02
Behavioural dimension 0.84 12.00 0.83 11.67
Sharing* 0.90 14.15 0.91 13.48
Learning* 0.76 11.97 0.84 12.83
Endorsing* 0.79 13.27 0.70 12.64
Brand Trust
I trust this brand 0.94 20.58 0.90 15.95
I rely on this brand 0.76 13.11 0.80 14.15
This is an honest brand 0.87 22.60 0.80 18.12
This brand is safe 0.88 15.21 0.80 15.27
Brand Commitment
I have grown to like this brand more than others offering the same product/service 0.82 16.33 0.87 16.32
I like the product/services offered by this brand 0.90 16.95 0.94 21.67
To me, this brand is the one whose products/services I enjoy using most 0.87 15.74 0.80 16.91
Brand Loyalty
I am loyal to only one brand (the one I follow), when I buy this type of product 0.78 10.80 0.87 16.32
For my next purchase, I will buy this brand again 0.91 11.95 0.92 17.33
I always buy this brand 0.77 14.01 0.73 15.67
I usually buy this brand 0.71 15.49 0.80 14.26
* Asterisks indicate that the summated scale of the sub-dimensions is used
 200

Model reliability and validity

The constructs are internally consistent with Cronbach’s Alpha values consistently above
0.86, which is well above the value advocated cut-off point of 0.60 (Bagozzi and Yi,
1988). The convergent validity indicators are also satisfactory, with AVE values all above
0.61, supporting the measurement model’s convergent validity. The CR indicators are
equal or above 0.83 for all constructs, which further indicate reliability, as suggested by
Hair et al. (2010). Correlations among latent variables are all significant (CR ≥ 1.96). As
all AVEs are superior to the square of their related pairwise correlations, this also indicates
that the measurement models achieve discriminant validity.

Table 23: Measurement model – Reliability and validity

English Sample
Constructs 1 2 3 4 5 6 7 8
1. Online brand engagement 0.84 0.79 0.32 0.47 0.43 0.25 0.33 0.15
2. OBC engagement 0.89 0.86 0.21 0.24 0.20 0.13 0.29 0.22
3. Brand loyalty 0.56 0.46 0.61 0.30 0.43 0.17 0.17 0.05
4. Brand trust 0.69 0.49 0.55 0.74 0.67 0.16 0.28 0.11
5. Brand commitment 0.66 0.44 0.66 0.82 0.75 0.12 0.16 0.07
6. Product involvement 0.50 0.36 0.42 0.40 0.34 0.89 0.22 0.02
7. Attitude 0.58 0.54 0.41 0.53 0.40 0.46 0.74 0.05
8. OIP 0.39 0.47 0.22 0.33 0.26 0.14 0.23 0.78
CR 0.94 0.95 0.86 0.86 0.90 0.94 0.92 0.93
Alpha 0.90 0.93 0.86 0.91 0.90 0.94 0.91 0.93
French sample
Constructs 1 2 3 4 5 6 7 8
1. Online brand engagement 0.87 0.81 0.35 0.34 0.22 0.15 0.23 0.18
2. OBC engagement 0.90 0.86 0.17 0.08 0.04 0.07 0.21 0.15
3. Brand loyalty 0.59 0.41 0.69 0.34 0.43 0.20 0.09 0.02
4. Brand trust 0.58 0.29 0.58 0.69 0.54 0.15 0.15 0.04
5. Brand commitment 0.46 0.19 0.66 0.74 0.76 0.17 0.07 0.01
6. Product involvement 0.38 0.26 0.45 0.39 0.41 0.92 0.09 0.08
7. Attitude 0.48 0.46 0.30 0.39 0.27 0.30 0.75 0.07
8. OIP 0.43 0.38 0.14 0.19 0.07 0.28 0.27 0.73
CR 0.95 0.95 0.90 0.83 0.90 0.96 0.92 0.92
Alpha 0.90 0.91 0.90 0.90 0.90 0.96 0.91 0.92
Note: The diagonal represents the AVEs of each construct; below the diagonal are the
pairwise correlations between constructs and above the squared pairwise correlations.
 201

6.2.2. The structural model

SEM with Maximum Likelihood estimation is used to test the hypotheses. In the causal
path model, the statistics only partly support that the data fit the model at adequate levels
with a chi-square= 3441.881 (p=0.00) with 1066 degrees of freedom. The CFI is 0.86, TLI
is 0.84, and RMSEA equals 0.06. This partial support for model fit is discussed prior in the
SEM measurement model results.

Table 24: Structural model – Results

Standardised
path estimate Path estimate CR Significance
(β)
Constructs EN FR EN FR EN FR EN FR
Online Brand Engagement (R2= 0.73/0.73)

OIP -0.20 0.04 -0.19 0.04 -0.39 1.01 0.69 0.31

Attitude toward
0.09 0.06 0.12 0.08 1.78 1.33 0.07 0.18
online participation
Product
0.20 0.19 0.20 0.16 4.19 4.30 0.00 0.00
involvement
OBC Engagement 0.75 0.77 0.74 0.83 9.75 12.05 0.00 0.00
OBC Engagement (R2= 0.40/0.26)

OIP 0.39 0.26 0.37 0.23 6.41 4.42 0.00 0.00

Attitude toward
0.47 0.41 0.58 0.54 7.49 6.87 0.00 0.00
online participation
Product
0.18 0.12 0.18 0.09 3.03 2.14 0.03 0.03
involvement
Brand Trust (R2= 0.52/0.34)
Online Brand
0.72 0.58 0.70 0.46 10.50 8.38 0.00 0.00
Engagement
Brand Commitment (R2= 0.52/0.34)
Online Brand
0.69 0.48 0.69 0.50 9.39 7.44 0.00 0.00
Engagement
Brand Loyalty (R2= 0.42/0.41)
Brand Trust 0.09 0.26 0.09 0.35 1.28 4.29 0.19 0.00
Brand Commitment 0.60 0.52 0.58 0.54 6.80 8.14 0.00 0.00
Note: R2 are given for the English, then the French sample
 202

Most of the hypothesised relationships are verified through the path analysis, with
different, yet largely consistent estimates for each sample. More specifically, H1b and H2b
are rejected both in the English and French sample, failing to account for the impact of OIP
and attitude toward OBC engagement on online brand engagement. Additionally, the
English sample also fails to support H7, evidencing rejection of the impact of brand trust
on brand loyalty in the English context. This hypothesis is however supported for the
French sample.

Table 25: Hypotheses results

Hypotheses English French

Antecedents of OBC engagement

H1a OIP is positively related to OBC engagement. Supported Supported

Attitude toward OBC participation is positively related to

H2a Supported Supported
OBC engagement.
Product involvement is positively related to OBC
H3a Supported Supported
engagement.

Antecedents of online brand engagement

H1b OIP is positively related to online brand engagement. Rejected Rejected

Attitude toward OBC participation is positively related to

H2b Rejected Rejected
online brand engagement.
Product involvement is positively related to online brand
H3b Supported Supported
engagement.
OBC engagement is positively related to online brand
H4 Supported Supported
engagement.
Outcomes of online brand engagement
Online brand engagement is positively related to brand
Supported Supported
H5 trust.
Online brand engagement is positively related to brand
H6 Supported Supported
commitment.
H7 Brand trust is positively related to brand loyalty. Rejected Supported
H8 Brand commitment is positively related to brand loyalty. Supported Supported
 203

H1 to H4

All the hypotheses related to the drivers of OBC engagement exhibit significant values in
both samples. Support is therefore granted to H1a (β = 0.39 English/0.26 French, sig =
0.00), H2a (β = 0.47 English/0.41 French, sig = 0.00) and H3a (β = 0.18 English/0.12
French, sig = 0.00). In other words, this means that OBC engagement is significantly and
positively influenced by consumer’s OIP, attitude toward OBC participation and product
involvement. Data show that the strongest influence of OBC engagement derives from
consumer’s general attitude toward online participation (β = 0.47 English/0.41 French),
directly followed by OIP (β = 0.39 English/0.26 French) in both samples.

The significance of the model’s path coefficient shows that online brand engagement is
positively influenced by product involvement (β = 0.20 English/0.19 French, sig = 0.00)
and OBC engagement (β = 0.75 English/0.77 French, sig = 0.00), showing support for H3b
and H4. No support is however granted to H1b (β = -0.20 English/-0.19 French, sig = 0.69
English/0.31 French) and H2b (β = 0.09 English/0.06 French, sig = 0.07 English/0.18
French), who exhibit in both samples levels of significance above the cut-off value of 0.05.
OIP and attitude toward OBC participation have therefore no effect on online brand
engagement. The results also show that the most powerful predictor of online brand
engagement is by far OBC engagement (β = 0.75 English/0.77 French).

H5 to H8

Brand commitment is positively influenced by online brand engagement, in support for H 5

(Model 1: β = 0.72 English/0.58 French, sig=0.00). The influence of online brand
engagement on brand commitment is expressed in hypothesis 6. This hypothesis is
validated, as evidenced by the significant beta values (β = 0.69 English/0.48 French,
sig=0.00). Online brand engagement therefore impacts both brand trust and brand
commitment, with a stronger influence on brand trust. Lastly, the impact of brand trust and
brand commitment on brand loyalty is hypothesised respectively with H7 and H8. Where
the French sample succeeds in accepting H7 (β = 0.26 sig=0.00), the English sample
evidences rejection of it (β = 0.09, sig=0.19). The impact of brand commitment on brand
loyalty is supported in both samples, denoting a full acceptance of H8 (β = 0.60
English/0.52 French, sig=0.00).
 204

Structural invariance

To verify structural invariance, the same procedure of constrain has been applied to the
structural weights, covariances and residuals following in AMOS (Byrne, 2010). The
multirgoup sample (French+English) exhibits clear invariance between the less and more
constrained models, with a ΔCFI=0.01 between the model with only the measurement
weights constrained and the fully constrained model (structural weights, structural
covariances and structural residuals constrained). It can thus be concluded from these
findings that there is structural invariance in the SEM model as developed and validated in
this section. A further examination of the French and English values separately, as
previously summarised in Table 25, shows that the structural paths between brand trust and
brand loyalty behave differently across samples: the relationship is significant for the
French but not for the English sample. Indeed, although the ΔCFI indicates invariance
between the models, the analysis of the significance levels of the Brand Trust  Brand
Loyalty relationship on the English sample shows that, as the model becomes contrained at
the structural level, the relationship becomes significant, whereas it is not for the
unconstrained model. This relationship is the only one for which the significance level
changes across models, suggesting that there is no other variation in the structural paths
across models. Based on this analysis, the interpretation of the structural invariance of this
study’s model across cultures must thus be careful and it can be concluded that partial
structural invariance is achieved across linguistic samples.

6.2.3. Testing of alternative model

Testing a structural model against another theoretical model is important in order to

validate the consistency of a given model specifications (Baumgartner and Homburg,
1996). This validation method has been used repeatedly in the online community literature
to verify the strength of models (e.g: Kim et al., 2008). One of the theoretical questioning
concerning the relationship between consumer engagement and other relational variables is
whether engagement comes first or after brand commitment and trust (Brodie et al., 2013).
Although loyalty is often understood as an outcome of engagement (Hollebeek, 2011a), the
precedence of engagement over trust and commitment is often left open (Hollebeek,
2011a). An alternative model whereby brand commitment and brand trust are hypothesised
to be antecedents of consumer engagement rather than outcomes is thus run, following the
same statistical techniques as previously detailed. Additionally, in this model, brand trust
and brand commitment being antecedents of engagement, they do no longer mediate the
 205

relationship between consumer brand engagement and brand loyalty. Figure 19 depicts this
model.

Figure 19: Alternative model

Table 26 shows that OBC engagement is not significantly impacted by brand trust and
brand commitment (β are all below 0.14 and significance levels above 0.05). This holds
for both French and English samples. Brand relational variables can therefore not be
modelled as antecedents of OBC engagement. When it comes to the precedence of brand
commitment and brand trust over online brand engagement though, significance levels
tend to be strong (between 0.00 and 0.07), although β are relatively low (between 0.21 and
0.14). This suggests that although weaker than in the proposed model of this study, the
precedence of brand trust and commitment over online brand engagement might be
considered. This potentially is however highly debatable, given the goodness of fit values
of the alternative model. Indeed, the fit of the alternative model is clearly not as strong as
the fit of the hypothesised model. The goodness-of-fit values are, for the English sample:
chi-square= 2705.92 (p=0.00) with 685 degrees of freedom. The CFI is 0.74, TLI is 0.76,
and RMSEA equals 0.11. For the French sample, chi-square= 2641.28 (p=0.00) with 685
 206

degrees of freedom. The CFI is 0.79, TLI is 0.77, and RMSEA equals 0.10. Overall, the
testing of the alternative model broadly supports the same set of relationships as this
study’s model. It also shows that brand trust and commitment should definitely not be
modelled as antecedents of OBC engagement, and that trust and commitment are also
better modelled as outcomes of online brand engagement. In other words, the testing of the
alternative model further validates the adequacy of this study’s model.

Table 26: Alternative model SEM results

Standardised
path estimate Path estimate CR Significance
(β)
Constructs EN FR EN FR EN FR EN FR
2
Online Brand Engagement (R = 0.87/0.92)

OIP -0.04 0.06 -0.04 0.05 -0.98 1.90 0.32 0.05

Attitude toward
-0.01 -0.06 -0.06 -.005 -1.18 -1.57 0.25 0.11
online participation
Product
0.15 0.03 0.17 0.03 3.46 1.07 0.00 0.28
involvement
OBC Engagement 0.75 0.81 0.69 0.64 10.40 10.48 0.00 0.00
Brand commitment 0.17 0.14 0.20 0.12 2.71 2.98 0.02 0.01
Brand trust 0.15 0.21 0.13 0.16 2.27 4.14 0.07 0.00
OBC Engagement (R2= 0.42/0.28)

OIP 0.34 0.22 0.43 0.25 5.35 3.63 0.00 0.00

Attitude toward
0.37 0.34 0.46 0.39 4.97 5.25 0.00 0.00
online participation
Product
0.58 0.05 0.07 0.05 0.86 0.81 0.38 0.41
involvement
Brand commitment 0.14 0.06 0.17 0.01 1.36 0.07 0.17 0.94

Brand trust 0.03 0.13 0.03 0.13 0.33 1.50 0.74 0.13
Brand Loyalty (R2= 0.28/0.32)
Online brand
0.53 0.57 0.64 0.95 7.02 8.16 0.00 0.00
engagement
Note: R2 are given for the English, then the French sample
 207

6.3. Summary

The aim of this chapter was to provide answers to research questions 2 and 3. Following
the development of the consumer engagement scales, testing of the study hypotheses was
completed, addressing the second and third research questions. The factor structure was
first assessed using CFA procedures on the measurement model, followed by the
evaluation of the structural model properties to test the causal hypotheses. The
measurement model exhibited adequate fit with excellent standard loadings for each item
as well as excellent values of reliability and validity. The structural model also exhibited
adequate fit, and the significance of the path estimate indicated the support of 17
hypotheses of the 22 hypotheses tested across the English and French samples. This partial
support of the hypothesised relationships is further discussed in the next chapter, along
with the meaning of the scale development results. Despite one of the hypotheses
performing differently on the French and English sample, a group invariance test on the
causal model also indicated that the two samples have are overall equivalent.

Chapter 7: Discussion

7.1. Introduction

This chapter discusses the implications of the results presented in the previous chapter and
it is structured as follows. Firstly, the consumer engagement measurement is addressed and
the discussion shows how the new scales of consumer engagement developed in this study
significantly deepen the meaning of existing conceptualisations of consumer engagement
and advance its measurement. The discussion also demonstrates how they differ from two
scales of consumer engagement – published after this study’s data were collected –
particularly by being better suited to highly interactive and social OBC environments.

Secondly, the findings related to the research hypotheses are discussed, focusing on the
relationships between the different constructs in the causal model. This section opens with
 208

a reminder of the different research hypotheses and their result, which is followed by a
detailed discussion of the implication of each of these results in light of the existing
literature. After the discussion of these core findings, other findings are discussed,
including the discussion of the study validation in another linguistic context.

7.2. Consumer engagement concept and measurement

7.2.1. Dimensionalisation of engagement

This study advances the notion of consumer engagement as a new approach to consumer
participation in OBC. This is achieved by refining the conceptual meaning of consumer
engagement and proposing a way to measure it. In this way, the study offers conceptual
and methodological advancement to current literature.

Conceptually, two views of consumer engagement dimensionality dominate the marketing

literature: the unidimensional and the multidimensional view. While the unidimensional
view tends to focus on consumer engagement being a behavioural concept (Van Doorn et
al., 2010; Verhoef et al., 2010) the multidimensional view is usually composed of a
cognitive, an affective and a behavioural dimension and it currently prevails in the
literature (Brodie et al., 2011; Hollebeek, 2011; Brodie et al., 2013). This three-
dimensional perspective was deemed the most appropriate in light of the OBC context in
which this study operates, which calls for a rich, multi-faceted measure of members’
participation (Bagozzi and Dholakia, 2006; Mollen and Wilson, 2010). In this sense, the
scales differ from other recent measure development of consumer engagement in social
media that take a unidimensional, action-based view of engagement, whether they are
based on platform-extracted metrics (Gummerus et al., 2012) or consumers’ self-reported
metrics (Schivinski et al., forthcoming).

Conceptually, the scale development and validation procedure undertaken in this study
resonates with the multidimensional view of consumer engagement. The study validates
the existence of cognitive, affective and behavioural dimensions of consumer engagement,
which were validated through thorough scale development procedures. In doing so, the
study supports the widespread understanding that consumer engagement manifests itself in
these three ways (Hollebeek, 2011a).
 209

Since a clear meaning of each of these dimensions was until recently lacking in the
literature (Brodie et al., 2013), this study worked to enhance this understanding and
provide precise definitions of each of these dimensions. This was achieved through a
review of the engagement literature in marketing and other fields of the social sciences, as
well as qualitative interviews, and validated in first-order CFA development. Although the
three-dimensional view and definitions adopted in this study broadly mirror the dimensions
recently developed in other consumer/customer engagement conceptualisations, they
explicitly depart from them. In contrast to Hollebeek et al. (2014) and Vivek et al., (2014),
this study offers a dimensionality of consumer engagement, which bears a clear long-term
orientation, makes a sharp distinction between the three dimensions and holds an
underlying social element to each of these three dimensions. All these three elements are
core to the conceptualisation of consumer engagement in an OBC, as expressed in Chapter
3.

A careful analysis of the definitions of the three dimensions in this study, in Hollebeek et
al. (2014) and Vivek et al. (2014) reveals their conceptual similarities but also differences,
as evidenced in Table 25. Hollebeek et al. (2014) identify cognitive processing, affection
and activation as dimensions of customer engagement. Since the conceptualisation of the
present research is largely based on, and in agreement with, the work by Hollebeek, Brodie
and colleagues (Brodie et al., 2011; Hollebeek, 2011a; Brodie et al., 2013), it is natural for
the conceptualisation of this research to be reflecting the same broad dimensions.
However, in Hollebeek et al. (2014), the affective and cognitive dimensions of engagement
do not seem to bear any long-term, enduring characteristic since they are based on specific
time-bound interactions. This is a major departure from the core premise of consumer
engagement being an enduring rather than transient phenomenon (Schaufelli et al., 2002),
which this study aims to reflect. Moreover, Hollebeek et al.’s (2014) ‘activation’
dimension refers to a level of energy, effort and time spent, in contrast to this study’s
behavioural dimension, which aims to reflect only clear behavioural manifestations.

Table 27: Comparison of this study's dimensions with two other studies

This study Hollebeek et al. (2014) Vivek et al. (2014)

 210

This study Hollebeek et al. (2014) Vivek et al. (2014)

Cognitive engagement Cognitive processing Conscious attention

A set of enduring and Consumer’s level of brand- The degree of interest the
active mental states that a related thought processing person has or wishes to have
consumer experiences with and elaboration in a in interacting with the focus
respect to his/her particular consumer/brand of their engagement.
engagement partner. interaction.

Affective engagement Affection Enthused participation

The summative and A consumer’s degree of The zealous reactions and

enduring level of emotions positive brand-related affect feelings of a person related
experienced by a consumer in a particular to using or interacting with
with respect to his/her consumer/brand interaction. the focus of their
engagement partner. engagement.

Behavioural engagement Activation Social connections

The behavioural A consumer’s level of Enhancement of the

manifestations toward an energy, effort and time interactions based on the
engagement partner, spent on a brand in a inclusion of others with the
beyond purchase, which particular consumer/brand focus of engagement,
results from motivational interaction. indicating mutual or
drivers. reciprocal action in the
presence of others.

Another customer engagement scale is developed by Vivek et al. (2014), also presented in
Table 25, who suggest the following dimensions: conscious attention, which is similar to
this study’s cognitive dimension, enthused participation, which underlies elements of
affect and behaviour, and social connections. The present study somewhat reinforces and
validates these findings, although casting them in a different light. Vivek et al. (2014)
contend that social connections with friends, or other people, are a core dimension of
customer engagement with a brand, using items such as ‘I enjoy [brand name] more when I
am with others’.

The conceptualisation of this study differs from Vivek et al. (2014) in a significant way, as
social connections with other users are not considered to be a dimension of engagement as
such. In contrast, this study understands consumer engagement as inherently social and
interactive (Hollebeek et al., 2014), since it inevitably involves a relationship with another
 211

partner (the brand, the brand community, or another engagement partner). This aspect of
engagement is therefore taken into account by attributing consumer engagement to specific
partners. However, this study does not consider engagement with a brand to be necessarily
involving others users. The sociality of the consumer engagement concept was particularly
important to be considered as underlying all its dimensions rather than being a dimension
itself. Being inherently social constitutes one of the key benefits of consumer engagement
as a measure of OBC participation. In this study, the social aspect of consumer
engagement is captured by focusing on consumer’s interactions with an engagement
partner, which is viewed as a core constituent of any engagement measure. The conceptual
development of the present research shows that consumer can be engaged with a brand and
with a group of other consumers (here the OBC). This evidences the inherently social
aspect of engagement and is further discussed in the next sections of this chapter.

Furthermore, in contrast to other scale development studies which combine two aspect of
engagement into one (e.g. Vivek et al. (2014)’s ‘enthused participation’ is reflective of
both emotions and behaviours), this study proposes that individuals exhibit emotions,
cognition and behaviours as part of engagement and these are three separate yet
complementary aspects of engagement. Discriminant validity between the three
dimensions evidences their distinctiveness, whereas the chi-square difference test supports
that they cannot be reduced to less dimensions. It is thus essential to consider engagement
as made of cognitive, affective and behavioural manifestations, which are separate and all
essential.

To summarise, the conceptual dimensions of this study reaffirm the general understanding
of the multidimensionality of consumer engagement, and support its cognitive, affective
and behavioural characteristics. This study contributes to existing literature by providing
clear definitions for each of the dimensions, which complement yet contrast other
definitions developed in Hollebeek et al. (2014) and Vivek et al. (2014). More specifically,
this study advances the extant literature by clearly delineating the three dimensions,
avoiding dimensions overlaps; by including an inherently social aspect to all dimensions
(see section on locus of engagement for further insight); and by also attributing a long-term
aspect to the three dimensions.

Conceptualisation of the sub-dimensions

A key contribution of this study’s findings concerns the refinement and clarification of the
 212

meaning of the dimensions of consumer engagement by proposing seven sub-dimensions:

enthusiasm and enjoyment, attention and absorption, and sharing, learning and endorsing.
Indeed, the conceptual blur around the meaning of consumer engagement dimensions
cannot be addressed solely by defining the dimensions, as presented above: a more refined
meaning had to be sought, particularly in order to create measurement items. Based on a
thorough literature review and qualitative interview, this study developed the following
seven sub-dimensions: enthusiasm and enjoyment for the affective dimension, attention
and absorption for the cognitive dimension and sharing, learning and endorsing for the
behavioural dimension.

Considering the cognitive dimension, it is conceptualised as composed of attention and

absorption, which are precise cognitive processes, in contrast to the more general idea of
‘thought-processing’ (Hollebeek et al., 2014). This also means that cognitive engagement
is necessarily involving both focused attention in the form of active cognitive availability
and absorption, reflecting the difficulty to mentally detach oneself from the focus of
engagement. This brings depth to the understanding of cognitive engagement, and shows
that it goes beyond a mere interest in interacting with something (Vivek et al., 2014) and
involves active attention and engrossment with respect to the engagement partner.

The affective dimension of engagement is made of enjoyment and enthusiasm. Here again,
this study deepens the meaning of affective engagement beyond a simple understanding of
it being related to positive affect or feelings (Hollebeek et al., 2014). It also shows the
enduring, intrinsic and motivational aspect of engagement (Vivek et al., 2012). In order to
be affectively engaged, a consumer needs to exhibit intrinsic levels of enthusiasm about an
engagement object, as well as subsequently gain pleasure in interacting with it. Both
enjoyment and enthusiasm support the interactive and long-term nature of engagement, as
they can only be sustained through reciprocal and valued engagement from the focal
partner.

The behavioural dimension of engagement is composed of sharing, learning and endorsing

behaviours. This is somewhat conceptually reflective of Brodie et al.’s (2011) engagement
sub-processes, which are sharing, learning, advocating, socialising and co-developing.
Whereas sharing, learning and advocating (a synonym of endorsing) are validated sub-
dimensions of behavioural engagement in this study, socialising and co-developing are not
included in the model. As explained earlier, this study considers the social aspect of
engagement to be inherent to each of its dimensions rather than a separate dimension, and
 213

co-developing to be an outcome rather than component of engagement (e.g. Sawhney et

al., 2005). If co-development is to be understood from a UGC approach though, it could be
viewed as reflective of this study’s ‘sharing’ behavioural sub-dimensions which implies
the diffusion of content by consumers, however not necessarily original.

Another parallel can be made with the study by Kumar et al. (2010) which focuses on
customer engagement value as a higher-order concept composed of purchasing behaviour
(the loyalty outcome in this study), referral behaviour (endorsing sub-dimension),
influencer behaviour and knowledge behaviour (sharing). This study expands on their
approach by proposing the learning aspect of engagement, in addition to its sharing and
endorsing aspects. It also excludes purchasing behaviour as a dimension of engagement
and rather proposes it as an outcome of engagement. Indeed, this study adopts the opinion
that consumer engagement refers to any interaction going beyond purchase behaviours:
anything that is related to purchasing is therefore excluded as a dimension of engagement.

Measurement

Based on a thorough conceptual domain definition and with the help of 25 qualitative
interviews, this study was able generate items to empirically develop two valid and reliable
scales of consumer engagement in an OBC. The items closely reflect the sub-dimensions
and dimensions of consumer engagement, as supported by good first-order and second-
order CFA goodness-of-fit indices and indicators, as well as reliability and validity
indicators.

This study sought to create definitions of the dimensions and sub-dimensions that are
going back to the meaning of engagement in the social sciences and marketing in order to
provide a strong ground for the generation of items. Supported by rich consumer data, this
study avoided culling items from other previously developed measures of concepts akin to
engagement or similar to its dimensions.

The general premise that consumer engagement is best viewed as a multidimensional

rather than unidimensional concept was reaffirmed through multiple tests of convergent
and discriminant validity, as the scale could not be reduced to less dimensions. The same
confirmation has been obtained for the sub-dimensions using second-order CFA
procedures. Overall, this study’s behavioural, affective and cognitive dimensions and sub-
dimensions work toward empirically validating current conceptual understandings of
engagement behaviours such as Brodie et al.’s (2011). However, they significantly deepen
 214

the conceptual meaning of consumer engagement dimensions in existence, and depart in

several important ways from recently developed scales (e.g. Hollebeek et al., 2014; Vivek
et al., 2014).

To summarise, this study offers a richer and more specific understanding of the meaning,
and thus operationalisation of the construct of consumer engagement. Empirical validation
is achieved, evidencing that each of the (sub-) dimensions is unique and necessary.
Moreover, the underlying social and enduring aspects of engagement are evidenced.

7.2.2. Locus of engagement

The study explicates an important oversight concerning consumer engagement: its locus.
Although engagement has been investigated for a number of engagement objects, or
partners, such as the brand (e.g. Mollen and Wilson, 2010), the brand community (e.g.
Algesheimer et al., 2005) or a firm (Kaltcheva et al., 2014), limited attention has been paid
to how the locus of engagement matters and how consumer can be engaged with several
partners at the same time. This thesis proposes a consumer engagement scale that can be
applied to different engagement partners in a given context, using the same set of items,
sub-dimensions and dimensions. A key advancement here is thus the explicit recognition
that in the context of OBC, consumer can engage concurrently with two engagement
partners, i.e. the brand and the brand community (Wirtz et al., 2013).

The validation of this core boundary assumption of the research (consumer engagement in
an OBC happens with respect to two different engagement partners, the other members of
the OBC and the brand itself) is a powerful contribution to the consumer engagement
literature. It shows that consumers can be engaged in the exact same ways with different
engagement partners in relation to their consumption experience, in the same virtual
context, and at the same time (Dessart et al., 2015). So far, consumer engagement research
had mainly focused on consumer-brand engagement (e.g. Hollebeek et al., 2014) and
community or OBC engagement had been treated as a distinct phenomenon (e.g.
Algesheimer et al., 2005). Reconciling the conceptual and operational approach of
consumer engagement for different partners in a given context is a key take from this
study, which advances the OBC and engagement literature.

From the OBC perspective, parallels can be made with the framework proposed by
McAlexander et al. (2002), reflecting on the multi-object approach to consumer
relationships within OBCs. Although this study focused on two of the relationship partners
 215

identified by McAlexander et al. (2002), there may be room for further application of the
consumer engagement construct to a product or firm partner, which this study has not
investigated in the context of OBC.

Considering the consumer engagement literature, this study’s multi-partner perspective

somewhat parallels the approach taken by Vivek et al. (2014) in their scale development
procedure, where they emphasised the need to validate their consumer engagement scale
across contexts. A change of context implied a change of partner, and they thus validated
their scale using a brand partner (Apple) and a retail partner (chosen by the participants).
In contrast to Vivek et al.’s (2014) study however, this research shows that several partner
can co-exist in the same context and that engagement with one can be related to
engagement with another, rather than being a matter of contextual validation.

In terms of measurement, the only difference between the two engagement scales is the
adaptation of the engagement partner within the item wording. For instance, when
measuring enthusiasm for the OBC partner, one of the items reads: ‘I feel enthusiastic
about the group’, whereas brand engagement was measured with ‘I feel enthusiastic about
the brand’. The same approach was adapted to all items.

Lastly, the fact that consumer engagement with the OBC positively impacts consumer
engagement with the brand (H4) leads to two key considerations. Firstly, it is another
empirical indication of the need to consider and measure engagement with different objects
separately. Secondly, it indicates that engagement with one partner might influence
engagement with another partner (Wirtz et al., 2013). A notion of precedence can therefore
exist when engaging with different partners, depending on the context. For instance, in the
context of retail grocery shopping, a consumer might be engaged with the retail brand and
with different product brands that the retailer stocks. This study suggests that previous
engagement with the product might influence engagement with the retailer, or the other
way around.

7.3. Discussion of the hypotheses

The results of the SEM confirmed the existence of several significant antecedents and
outcomes of consumer engagement. Table 26 summarises the research hypotheses and
 216

their result for the English and French sample respectively. It presents an overview of the
confirmed or rejected relationships between the two consumer engagement scales, as well
as its antecedents and outcomes.

The antecedents include OIP, product involvement and attitude toward community
participation, which are all hypothesised to impact online brand engagement and OBC
engagement positively. OBC engagement is then hypothesised to positively influence
online brand engagement. The hypothesised direct outcomes of online brand engagement
are brand trust and brand commitment, and its indirect outcome, mediated by trust and
commitment, brand loyalty. All hypotheses are supported except H1b and H2b, both for the
English and French sample, as well as H7, for the English sample only. The section below
offers a discussion of these findings.
 217

Table 28: Summary of the research hypotheses

Hypotheses English French

Antecedents of OBC engagement

H1a OIP is positively related to OBC engagement. Supported Supported

Attitude toward OBC participation is positively related to
H2a Supported Supported
OBC engagement.
Product involvement is positively related to OBC
H3a Supported Supported
engagement.
Antecedents of online brand engagement

H1b OIP is positively related to online brand engagement. Rejected Rejected

Attitude toward OBC participation is positively related to

H2b Rejected Rejected
online brand engagement.
Product involvement is positively related to online brand
H3b Supported Supported
engagement.
OBC engagement is positively related to online brand
H4 Supported Supported
engagement.
Outcomes of online brand engagement
Online brand engagement is positively related to brand
H5 Supported Supported
trust.
Online brand engagement is positively related to brand
H6 Supported Supported
commitment.
H7 Brand trust is positively related to brand loyalty. Rejected Supported
H8 Brand commitment is positively related to brand loyalty. Supported Supported

7.3.1. Antecedents of consumer engagement

The first set of hypotheses concerns the antecedents of consumer engagement. The
conceptual model posited that three individual antecedents positively impact OBC and
online brand engagement: OIP, attitude toward OBC participation and product
involvement. The results provide partial support for the hypotheses and the section below
discusses their implications for OBC engagement and online brand engagement
respectively.

All relationships concerning the antecedents of OBC engagement are supported. In other
words, OBC engagement is positively impacted by OIP, attitude toward community
participation and product involvement.
 218

Online interaction propensity  OBC engagement

The data support the positive impact of OIP on OBC engagement. This relationship
resonates well with the literature on virtual communities, as OIP is known to increase
member’s contribution to OBC (Wiertz and de Ruyter, 2007). For example, the current
findings provide support for Wiertz and de Ruyter (2007) who identified OIP as a potent
moderator of the individual’s quantity and quality of knowledge contribution within
community settings. Similarly, the study reinforced Chang and Chuang’s (2011) study,
which proved the positive impact of social interaction on the quality of information shared.
These two studies show a clear connection between OIP and behavioural engagement in
the form of information sharing.

Findings from this study extend this research by showing that OIP positively influences
engagement and by offering a refined notion of engagement. That is, engagement here is
composed of emotional and cognitive dimensions, in addition to participation behaviours.
This means that individuals who are more prone to online interaction are not only likely to
subsequently exhibit behaviour in the community as previous studies show (Wiertz and de
Ruyter, 2007), but will also, de facto, be more cognitively engaged and think about the
community, as well as feel more positively toward the community.

These findings also support research in the consumer engagement literature. For example,
they resonate well with Blazevic et al. (2014) who recently confirmed the positive impact
of OIP on engagement on social media based brand communities. Similarly, Calder et al.
(2009) had advanced the need for social interactions as a driver of consumer engagement.
There again, however, the conceptualisation of engagement is limited to behaviours,
limiting the explanatory power of OIP to one dimension of consumer engagement (Calder
et al., 2009; Blazevic et al., 2014).

Additionally, this finding reinforces the importance of consumer engagement as an

interactive concept. The fact that OBC engagement is reliant on consumers’ intrinsic level
of interaction propensity corroborates the fact that consumer engagement is an interactive
phenomenon (Brodie et al., 2013; Breidbach et al., 2014), which is one of its core strengths
in comparison to other of OBC participation approaches. Furthermore, this finding
evidences that the lack of research into OIP in OBC settings needs to be addressed (Wiertz
and de Ruyter, 2007), particularly when considering manifestations of consumer
engagement.
 219

Attitude toward OBC participation  OBC engagement

A second verified driver of OBC engagement is consumer’s attitude toward OBC

participation. Previous studies argued that attitudes are evaluative (cognitive) reactions to
an action and are thought to reflect predispositions to respond in a favourable or
unfavourable manner (Eagly and Chaiken, 1993). The fact that consumer’s attitudes
toward participation might be a driver of actual behaviour has been the subject of much
research, stemming largely from the TAM, TPB and TRA. In this context, online
community and social networking sites research has shown that member’s participation
was driven in part by their attitude toward participating, albeit always with the moderating
effect of intentions to participate (Casaló et al., 2010; Gironda and Korgaonkar, 2014) as
well as desires (Bagozzi and Dholakia, 2006). In the OBC applications of the TAM and
TPB, the impact of attitude toward participation over actual participation is always
mediated by intentions to participate (e.g. Bagozzi and Dholakia, 2006).

The findings from this study correspond to and extend this existing understanding of the
role of attitudes on OBC participation in a major way. The findings extend current
cognitive behaviour models such as the TPB and TRA by showing that attitude are directly
related to OBC engagement, without the mediating effect of intentions to act. This parallels
the existing critiques of these cognitive behaviour models that report an intention-
behaviour gap (e.g. Carrington et al., 2014), giving impetus to the test of direct attitudes-
actions relationships.

In addition to supporting and simplifying past understandings through the removal of the
intentions construct, the findings show that attitudes are not only antecedents of social
interactions in the context of online brand communities, but also of social emotions and
cognition. As an individual gets more positively inclined toward the community, his
feelings for the community members evolve and the mental focus he places on them as
well. If community participation is perceived to be rewarding, beneficial, or wise
(components of a positive attitude), one will be likely to pay more attention to the
community and its content (Nonnecke et al., 2006) and get more pleasure from
participating (Wiertz and de Ruyter, 2007).

Overall, these findings bring important insights to the understanding of OBC engagement.
The attitude that one has toward an action such as OBC participation not only directly
 220

influences the behavioural participation itself, but also the cognition and emotions
associated with it, making up for increased levels of OBC engagement.

Product involvement  OBC engagement

The support of the relationship between individual product involvement and OBC
engagement highlights several current shortcomings of both OBC and engagement
literature. It challenges the usual focus on brand involvement when exploring drivers of
consumer engagement (Bowden, 2009; Brodie et al., 2011). Indeed, the engagement
literature had so far only, and mainly conceptually, explored the impact of brand
involvement on consumer engagement (e.g. Brodie et al., 2011). This study therefore
expands previous research by showing that involvement with a product category will likely
drive consumers to be actively engaged with a community of other consumers.

These interesting findings lend support to existing OBC studies. Community studies have
shown that consumer’s relationship with a product is an integral part of OBC participation
(McAlexander et al., 2002; Stokbürger-Sauer, 2010). More specifically, it is understood
that initial product involvement can affect different levels of behavioural community
participation, from lurking (Shang et al., 2006) to active future community participation
(Nambisan and Baron, 2007). This phenomenon can be explained by the very nature of
engagement, which implies a higher commitment to searching and finding information
about a product, in order to reduce uncertainty and risk (Chaudhuri, 2000; Dholakia, 2001).
Learning is an important part of behavioural engagement, which is potentially directly
impacted by the cognitive aspect of product involvement. Because OBCs are rich and
trusted resources for information search (Bagozzi and Dholakia, 2006), members rely on
them extensively when they have high levels of product involvement requiring thorough
information search (Sanchez-Franco and Rondan-Cataluña, 2010).

This study’s results emulate and extend previous findings, by suggesting that involvement
with a product category is a driver of OBC engagement. In other words, this research
implies that an individual interested in a particular type of product such as coffee, but not
initially in a particular brand like Lavazza, could become actively engaged in a community
focusing on Lavazza, out of involvement in coffee consumption. Because this study
conceptualises engagement as being affective, cognitive and behavioural, product
involvement here is not only conducive of higher levels of behavioural participation and
information search, but also of higher levels of affective and cognitive engagement with
 221

the other consumers, which is an aspect of the involvement-engagement relationship that

has received little attention so far (Hollebeek et al., 2014).

Overall, the fact that all three hypothesised antecedents impact OBC engagement has two
key implications. Firstly, it shows that OBC engagement can be largely predicted by
individual traits and predispositions. These traits and predispositions have an enduring
aspect, similarly to consumer engagement itself (Vivek et al., 2012). They are not
situation-dependent but are rather engrained in consumers’ self. Secondly, out of these
three drivers, attitude toward OBC participation proved to be the most powerful, followed
by OIP and product involvement, and this for both French and English samples. This study
extends existing research by showing the impact of these three factors on OBC
engagement as a multidimensional construct, whereas most OBC research investigating
these traits had so far only focused on their behavioural participatory outcomes (e.g.
Wiertz and de Ruyter, 2007). These findings also imply that OBC engagement can be
understood as a function of individual traits and predispositions, a postulate that was so far
only conceptual (Wirtz et al., 2013). Lastly, they suggest the need to consider consumer
individual profiles and identity traits in OBC engagement strategies (Van Doorn et al.,
2010; Campbell et al., 2014).

7.3.2. Antecedents of online brand engagement

The hypotheses relating to the drivers of online brand engagement have proven to be more
problematic than those related to OBC engagement. Out of the three hypothesised
relationships, the data only supports the positive impact of consumers’ product
involvement.

Online interaction propensity  online brand engagement

Despite theoretical support for the relationship found in previous studies (Wiertz and de
Ruyter, 2007), the results failed to confirm the link between OIP and online brand
engagement, and this seems true for both French and English samples. There are several
potential explanations for this finding. It seems that although an individual might be prone
to interact online, there might be some barriers to engaging directly with a brand in these
settings. The barriers might be linked to the brand not fostering communication, either by
absence of cue to interact, or by absence of ways to interact. Having an official page on
social media is not enough to signal that reaching out is welcomed by the brand and that
consumers’ communications will be reciprocated, the brand needs to be active in signalling
 222

this (Kaplan and Haenlein, 2010). Some companies might altogether lack an engagement
of social media strategy, which consumers can easily pick upon when frequenting the page.
Consumers might feel hindered in their communications with brands because they do not
feel welcome to interact, or because they are afraid to display this engagement publicly
(Nonnecke et al., 2006). It seems that these effects have been underplayed in some
literature (e.g. Wiertz and de Ruyter, 2007; Blazevic et al., 2014).

Another explanation for this lack of support is that consumers, although highly prone to
interact online with other actors, do not want to interact with brands in particular (Fournier
and Avery, 2011). Social media were initially platforms for individuals to connect with
peers (Kaplan and Haenlein, 2010) and have thus started growing in this direction. Brands
hopped on the bandwagon later on and positioned themselves as rightful actors of the
social media ecosystem. Research however shows that they are not always welcome to do
so (Fournier and Avery, 2011; Hollebeek and Chen, 2014). For this reason, individuals
might be less used to, and more reluctant to communicate with brands on social media,
even if they normally have a high OIP. Additionally, OBC users might feel that engaging
with the brand in public setting is inadequate, preferring to keep direct brand interactions
in the private sphere. This might be due to self-presentation motives (Hollenbeck and
Kaikati, 2012); for fear of what others might think of you if you interact too closely with
brands (Patterson, 2012); or for privacy concerns (Murphy et al., 2014). The finding thus
may add credence to the literature that views brands as agents of consumers’ social and
private identity formation efforts (Hollenbeck and Kaikati, 2012).

It is also possible that other factors, unaccounted for in this study, are indeed reflected in
those findings. For example, the size of the community might act as a moderator of the
relationship between OIP and online brand engagement (Wirtz et al., 2013). Indeed,
research shows that the bigger the OBC becomes, the less it feels like a community in
which people feel free to share their feelings and opinions (Dholakia et al., 2004). In
contrast to this proposition, a smaller community might also make its individual members
and their actions more visible, which might deter self-conscious members from
participating. The fact that the community size could moderate member’s inherent OIP
impact on online brand engagement is a thesis worth further exploration. The current state
of research however suggests that both large and small community settings could keep
inherently social members from participating for different reasons (Dholakia et al., 2004).
 223

Attitude toward OBC participation  online brand engagement

The findings failed to confirm that attitude toward OBC participation is a driver of online
brand engagement. This finding is surprising, especially. To illustrate, Bagozzi and
Dholakia (2006) linked attitude toward online participation to brand behaviours in an
online community setting. In their study, attitude is far removed from behaviours, as their
interplay is mediated by desires and social intentions first, followed by group behaviour.
The findings indicates that the results of this study may suggest a much more complex
relationship between the attitude toward OBC participation and online brand engagement
than what has been expected. For instance, there might be several mediators between the
attitude toward OBC participation and effective brand-related actions, such as consumers’
desires and intentions, as suggested by Bagozzi and Dholakia (2006).

Another explanation for this relationship not receiving support is a potential

misunderstanding of the attitude referent in the questionnaire. Although the study used an
existing and validated operationalisation of attitude toward OBC participation (Bagozzi
and Dholakia, 2006), the respondents might have not clearly understood that the
community represented both the other consumers and the brand. Respondents might
therefore have understood the question as referring to interactions with the other
consumers only, therefore reducing the explanatory power of their attitude toward OBC
participation on their online brand engagement. Following this line of thinking it maybe
questionable whether the typical OBC member is really conscious of the brand being a
member of the community on social media (Baird and Parasnis, 2011). Further in-depth
qualitative studies in that direction may help to answer this doubt.

Product involvement  online brand engagement

Another interesting finding concerns the impact of product involvement on online brand
engagement. Product involvement has repeatedly been show to increase consumer’s
affective, cognitive and behavioural responses to brands (e.g. Traylor and Joseph, 1984;
Mittal and Lee, 1989; Pritchard et al., 1999), which lead to the hypothesis that it would
equally lead to higher online brand engagement. The impact of product involvement is the
only hypothesis that was supported in this research when it comes to the individual
antecedents of online brand engagement, making its contribution all the more important to
understand. These findings strongly support existing consumer engagement research,
which posits brand involvement to be a driver of consumer engagement (Brodie et al.,
 224

2011; Hollebeek et al., 2014). Although an enduring interest in the brand is likely to lead to
direct interactions with the brand, interest in a product category has an equal explanatory
power over online brand engagement. In other words, this means that when a consumer has
an increased interest or concern for a type of product or service, for instance airline
companies, he or she will tend to be more engaged with particular brands belonging to this
category (Beatty et al., 1988).

If a person wants to know more about a certain type of product, turning to the community
for a balanced point of view seems like a frequently used strategy (see discussion of
hypothesis H3b), but so is engaging directly with a brand in order to get first-hand
experience. In their exercise to segment brand community members, Ouwersloot and
Odekerken-Schröder (2008) show that strong consumer-product relationships are often
combined with strong consumer-brand relationships in most engaged consumer profiles.

To conclude the section on the drivers of consumer engagement, it appears that

individuals’ existing attitudes, interests and propensities have varying impacts on OBC and
online brand engagement. Although product involvement positively impacts engagement
with both partners, attitude toward community participation and OIP only drive OBC
engagement. The overall picture that emerges from the study’s results in terms of the
drivers of both types of engagement is therefore that, although OBC engagement is driven
by all three individual traits considered relevant, only one of them impacts online brand
engagement. This means on the one hand that the drivers of community and brand
engagement differ, and on the other hand calls for another angle to understand the sources
of online brand engagement, which may lie in social influences rather than individual
traits.

OBC engagement  online brand engagement

The fourth research hypothesis has posited that the more a consumer is engaged with the
OBC, the more he or she will be engaged online with the host brand (Wirtz et al., 2013).
The support of this hypothesis found in both French and English samples represents an
important contribution, also because the hypotheses regarding the antecedents of online
brand engagement were in part rejected. Importantly, impact of OBC engagement over
online brand engagement is much stronger than the impact of product involvement,
highlighting the vital explanatory power of community participation over online brand
engagement.
 225

These findings expand the notion that interaction with a community of consumers fosters
stronger and more frequent brand-related behaviours and attitudes (e.g. Algesheimer et al.,
2005; Veloutsou and Moutinho, 2009). Looking at each facet of the OBC engagement
construct, several conclusions can be drawn. Firstly, being emotionally engaged with a
group of peers seems to positively affect engagement with a brand. This means that
positive emotions, such as enjoyment or enthusiasm, derived from interactions with other
consumers on a Facebook brand page lead consumers to be more engaged with the host
brand itself. This finding complements the contention that positive and negative feelings
generated from interactions in a virtual environment can shape consumers’ attitudes and
perceptions regarding the host brand (Nambisan and Baron, 2007). Moreover, cognitive
engagement with the community leads to increased brand engagement, and so do more
sharing, learning and endorsing behaviour with the focal community.

The reasons for this may stem from a simple observation: while OBC engagement means
that a consumer engages with other consumers in various forms and manifestations
(affective, cognitive and behavioural), the content of this engagement is often directly, or
at least indirectly linked to the focal brand, which is a fundamental premise of the OBC
literature (Muñiz and O’Guinn, 2001). Being more exposed to brand-related information
and gaining increased brand-related experiences and practice (Calder et al., 2009; Schau et
al., 2009) through community engagement, one’s level of brand engagement is triggered
and enhanced. Engaging in brand-related community interactions is inevitably going to
make the consumer think about the brand more and at the very least increase its share of
mind.

Interestingly, the precedence of OBC engagement to online brand engagement also allows
for drawing parallels with the discussion on the evolving roles of community members and
community practices. Research has noted that people assume different roles in OBCs and
that these roles are not static over time (Schau et al., 2009; Fournier and Avery, 2011). If
the ultimate goal of brand-hosted communities is to increase brand relationships through
brand engagement, the community can act as a catalyst to make members assume more and
more actively engaged roles (Fournier and Lee, 2009; Schivinski et al., forthcoming). As
people engage with the online brand community, new comers can assume roles of learners
while more experienced members act as greeters, supporters or guides (ibid., 2009), all the
while mingling in a multidimensional engagement process. Consumers holding a
particularly central position in the community’s relational network are also likely to act as
 226

influencers and stimulate less knowledgeable members (Trusov et al., 2010). Relating back
to the individual drivers of engagement, consumers exhibiting high levels of OIP, positive
attitude toward the community and high product involvement are most prone to engage.
The interactive and social nature of engagement combined with the dynamic of the
different community roles indicates that ultimately, all OBC members should be engaged
on some level.

Similarly, parallels can be made with the practices developed by Schau et al. (2009), as
they are interlinked with repeat consumer-to-consumer engagement. When experienced
members ‘welcome’ and ‘empathise’ with new members, they can foster their enthusiasm
and enjoyment of community participation, strengthening their affective community
engagement. Through competent practices of ‘milestoning’, ‘badging’ and ‘documenting’,
sharing, endorsing and learning dimensions of engagement are enacted. Likewise,
‘justifying’ the time spent on the community reinforces the cognitive engagement one
might experience. A continuous engagement with the community will not only see
practices being reinforced. As community members become more competent and fluent in
engaging with the community, they can evolve –almost ‘graduate’, to a higher level of
community integration and membership, which will place them in a more comfortable
position to engage directly with the brand.

In addition to complementing the OBC literature, this finding advances the current state of
the consumer engagement literature, where the role of OBC engagement is still
misunderstood (Wirtz et al., 2013). This study’s results empirically supports the
conceptual premise that OBC engagement drives online brand engagement in the context
of OBC, as proposed by Wirtz et al. (2013). Interestingly, the opposite causal relationship
has been put forth in the literature as well, giving precedence to brand engagement over
community engagement (Brodie et al., 2013). The relationship might be cyclical and more
engagement with an actor might lead to more engagement with another, and vice-versa.
The community/brand engagement link might therefore be a circle that feeds itself, and
increased brand engagement could lead to further community engagement. However, in the
context of this study, the dependent variables of interest are brand related and the focus
placed on understanding how consumer engagement contributes to enhance brand
relationships. In this context, the impact of community over brand engagement is of
particular relevance and the present study supports this hypothesis.
 227

Considering the broad agenda of consumer-centred relationships in OBC, the link between
consumer engagements with two different partners is in line with McAlexander et al.’s
(2002) concept of brand community integration. This study’s findings show that in a
highly interactive and social environment like an OBC, consumers are the focal subjects of
engagement but that their engagement with one partner is inevitably connected with their
engagement with another partner. This evidences the existence of many-to-many
interactions (Hoffman and Novak, 1996) and represent the ultimate level of individual’s
sociality through community-based social interactions (Murphy et al., 2014).

Lastly, these findings shed light on recent advances regarding social media branding
strategies, and Facebook pages management in particular. Malhotra et al. (2012) suggest
that in order to increase levels of behavioural engagement on Facebook (through Likes,
Comments and Shares), brands should not hesitate heavily promoting the brand and its
products and directly engage with consumers with calls-to-action. Although these
strategies should prove very powerful to create brand engagement, this study brings a
nuance to their analysis, by suggesting that there should be a sequence, or progression
from community engaging content to more brand-focused content. In this spirit, the
suggestion to post topical content that is not related to the brand should therefore also be
considered and even higher the agenda during the early stages of community building.

To conclude, engagement with one partner (here, the OBC as personified by its members)
might impact engagement with another partner (the brand) and thus occur in a sequential
manner. Being invested in an OBC (e.g. exhibiting positive emotions toward it, investing
time in it, and choosing to actively share with it, learn from it and endorse it) is a driving
force of online brand engagement. The role of OBC engagement over online brand
engagement is particularly interesting, as two of the three individual traits do not directly
impact online brand engagement. This suggests that OBC participation with other
members derives from individual predispositions whereas direct brand interactions in OBC
are supported by prior engagement with the rest of the community of members.

This finding has broad implication in terms of brand management and meets the growing
recognition of the social nature of brands (Thought Economics, 2009). By showing the
vital role of OBC engagement in sustaining brand engagement, it supports Keller (2013)
who pointed out that exhibiting and acting upon a sense of community is a necessary and
ultimate stage into developing brand engagement.
 228

7.3.3. Outcomes of online brand engagement

Two direct and one indirect brand relationship outcomes of online brand engagement have
been advanced in the research hypotheses. Brand engagement has been conceptualised to
be a direct driver of brand trust and brand commitment, ultimately leading to brand loyalty
(Hollebeek, 2011a).

Online brand engagement  brand trust

As expected, the results show that trust and engagement are positively correlated. Brand
trust first is defined as consumer’s willingness to rely on the ability of the brand to perform
its stated function (Chaudhuri and Holbrook, 2001, 2002) and it is not surprising that
higher levels of online brand engagement contribute to reinforcing it. In the engagement
literature, brand trust is one of the aspects of brand relationship quality and brand
engagement is conceptualised as one of its antecedent both for existing and new customers
(Hollebeek, 2011a).

Trust is likely to be increased by engagement because, in the interactive process of

engagement, consumers give brands the opportunity to assure them of their quality as a
relationship partner (Hollebeek, 2011a). Like trust, engagement is built over the long term
and repeat occurrences, and if a brand behaves in a way to allow consumers to
satisfactorily engage with them, trust is likely to occur. Echoing the recent brand
community and social media literature, if the brands provide compelling content to share
and learn from, it they entertain and keep consumer’s attention through their actions on
social media (Malhotra et al., 2012), allowing consumers to be engaged with them, and if
they do so consistently over time, consumers are more likely to trust them (Mazorcchi et
al., 2013).

Online brand engagement  brand commitment

The research hypothesis concerning the link between the implications of online brand
engagement for commitment has also been confirmed. Brand commitment represents
consumer’s enduring desire to maintain a valued relationship with a brand in the long term
(Morgan and Hunt, 1994) and it is a recurring theme in the OBC literature (e.g. Kim et al.,
2008). The findings of this study prove that online brand engagement positively influences
brand commitment. From a contextual point of view, this supports the idea that being
engaged in OBC settings leads to increased brand commitment (Casaló et al., 2008; Kim et
al., 2008; Jahn and Kunz, 2012). As community members become more engaged with the
 229

brand in the OBC settings, brand commitment is sustained. This means that through
competent engagement strategies, brand can foster higher levels of consumer intentions to
remain committed to their relationship with them. Brand commitment being an enduring,
long term predisposition to maintain a relationship, the long term orientation of consumer
engagement is further substantiated through the support of this hypothesis.

The findings of this study suggest that brand engagement is conducive of brand trust and
brand commitment, which are two core aspects of brand relationship quality. So far, these
relationships had only been conceptualised (Van Doorn et al., 2010; Brodie et al. 2011;
Hollebeek 2011a) and still required empirical validation, which this study seems to
provide. The ability of online brand engagement to secure higher levels of brand
relationship quality through brand trust and brand commitment settles the vital contribution
of the consumer engagement construct as an integral part of relationship building in a
social context (Brodie et al., 2013).

These findings also expand Bowden’s (2009) conceptualisation of the brand engagement
process whereby trust and commitment are two important components of engagement.
Although this study does not abide by the principle that engagement is an overarching
process (Bowden, 2009), in line with seminal studies, it also considers brand loyalty to be
the ultimate goal of brand relationship building (Morgan and Hunt, 1994), an issue that is
discussed in detail in the next section.

Brand trust  brand loyalty

The last set of hypotheses refers to the impact of brand trust and brand commitment on
brand loyalty. Brand loyalty in this study is approached from it behavioural perspective
and refers to consumers’ repeat purchase behaviour (Jacoby and Chesnut, 1978). The
impact of trust and commitment on loyalty has been the subject to much empirical research
in the wider brand relationship literature (Chaudhuri and Holbrook, 2002), but has so far
lacked empirical validation in consumer engagement perspective (Bowden, 2009;
Hollebeek, 2011a).

Considering the brand trust-loyalty relationship first, the study grants support to the effect
of brand trust on brand loyalty in the French sample. An interesting contrast between the
English and the French sample is detected, as the relationship is not supported for the
former. Differences between linguistic contexts is discussed in detail in section 7.4 of the
chapter.
 230

Overall, the findings resonate well with the widespread agreement that trust is positively
related with loyalty, which derive from general brand relationship studies (Garbarino and
Johnson, 1999). Trust is a particularly salient antecedent of loyalty for high-relational
consumers (ibid., 1999), which often qualifies OBC members. By actively displaying high
levels of reliability, integrity, and quality in the (online) brand community, brands are able
to sustain stronger levels of behavioural loyalty (Marzocchi et al., 2013).

Brand commitment  brand loyalty

Commitment represents the second hypothesised driver of loyalty and outcome of brand
engagement. In this instance, the hypothesis has been verified in both samples and no
cultural differences have been found. The commitment-loyalty link is well established in
the general relationship marketing literature (e.g. Morgan and Hunt, 1994), and this
relationship had also been conceptualised in the consumer engagement literature (van
Doorn et al., 2010; Hollebeek, 2011a). In the OBC literature as well, commitment to the
community and/or to the brand is generally viewed as a mechanism through which positive
behavioural outcomes are achieved (Algesheimer et al., 2005; Kim et al., 2008).

Two words of caution however need to be mentioned when analysing the relationship
between consumer engagement and its relational outcomes, in particular loyalty. Firstly,
this study defined consumers as any member of a Facebook page, irrespective of prior
purchase of the brand. Sample characteristics evidence that on average 84 percent of the
respondents are purchasing customers of the brand they like on Facebook (see Appendix
7). All respondents were however able to answer the questions related to brand loyalty,
thanks to the Likert scale type allowing negative answers. Any form of loyalty
measurement however needs for this reason to be taken with a pinch of salt, since OBC
membership does not necessarily entail brand custom, and even less so loyalty. Despite its
widespread adoption, assuming behavioural loyalty to be the number one indicator of OBC
success has clear shortcomings.

Furthermore, a discussion on the precedence of OBC participation over brand relationship

exists, and is referred to as the ‘advocacy paradox’ (Baird and Parnasis, 2011). This
paradox relates to the fact that companies tend to believe that engaging consumers on
social media will necessarily have positive loyalty and advocacy effects, whereas many
consumers say they need to be passionate before they’ll engage, and that this online
engagement with the brand is not systematically conducive of loyalty (ibid., 2011). This
 231

discussion relates to the debate in the consumer engagement literature whereby scholars
hesitate to conceptualise engagement as preceding or resulting from existing brand-
relationships (see Hollebeek, 2011). Rather than settle or provide a definite answer on this
matter, this study contributes to this discussion by evidencing the importance of engaging
with different partners to achieve relationship outcomes, since online brand engagement is
a function of OBC engagement. This echoes Baird and Parnasis (2011) findings that OBC
members rely on the community’s endorsements behaviour on social media to influence
their purchase behaviour. If behavioural brand loyalty needs to be considered carefully in
OBC contexts, the strength of engaging through endorsement behaviour is key as well as
the importance to engagement with the OBC and not just the brand.

Overall, and being cautious about their meaning, the findings regarding the drivers of
loyalty bring depth into the understanding of loyalty development as deriving from
consumer engagement. Loyalty is driven by brand commitment, and partly by brand trust
depending on the culture, both of them directly stemming from online brand engagement.
These results are interesting because they evidence that online brand engagement alone
cannot cater for loyalty, but works toward increasing fundamental brand relationship
elements such as trust and commitment, which ultimately lead to augmented instances of
purchase behaviours (Bowden, 2009; Hollebeek, 2011). In the OBC literature, several
studies had already highlighted the reinforcing potential of community participation over
adoption of new products by the brand (Thompson and Sinha, 2008) and general impact of
consumers’ lurking or posting behaviours on brand loyalty (Shang et al., 2006), albeit
without the mediating impact of brand trust and commitment, and without the integrative
approach to participation that consumer engagement affords.

Considering the consumer engagement literature, the engagement-loyalty nexus is a

complex one, and it has so far mainly been approached conceptually (van Doorn et al.,
2010; Hollebeek, 2011a). The present findings highlight the recent advances of Hollebeek
et al. (2014) who verified the direct impact of brand engagement on brand usage intent,
which can be assimilated to loyalty intent. In the present study however, brand trust and
commitment are mediators of the relationship, and direct reporting measures of purchase
behaviour are used rather than intentions to use, an approach that deepens past findings.

Overall, this study shows that through the mediating impact of brand trust and brand
commitment, consumer engagement drives consumer brand loyalty. This has major
 232

implications and shows the need for brands to sustain high levels of intra-community
engagement to ensure appropriate levels of brand relationship quality and subsequent
purchase behaviour. Furthermore, it advances the relevance of a multidimensional view of
community participation in brand relationship building. Consumer engagement is an
affective, cognitive and behavioural concept (Brodie et al., 2011), and it is the combination
of these three constituents that plays a role in driving trust, commitment and loyalty. Only
by ensuring appropriate levels of affect, cognitive processing and interactive behaviours
with a brand in the OBC context can managers reap the full benefits of their OBC.

7.4. Other findings: validation procedure

This section discusses the results concerning the group invariance tests computed with
AMOS to verify equality between the English and the French samples. As per the
recommended procedure (Byrne, 2010), configural, measurement and structural invariance
between the two samples has been tested. This method has first been applied to the
consumer engagement scales, and then to the whole structural model.

The initial assumption of invariance across samples has been verified thanks to a multi-
group confirmatory factor analysis, showing support to the applicability of the scales and
structural relationships to the two different languages: French and English. A concurrent
treatment of the two samples in hypothesis testing however highlighted a difference in data
behaviour between the two samples, as the brand trust-loyalty relationship was supported
for the French sample but not for the English one. The contextual invariance results are
first discussed with respect to the consumer engagement scales and then regarding the
whole structural model.

In the context of OBC embedded on social media—Facebook in this case, online consumer
engagement with a brand and with a brand community is adequately measured using the
same translated instrument, irrespective of whether the consumer is using the French or
English version of the platform. Since Facebook supports 91 languages (Facebook, 2014),
the linguistic replicability of any instrument related to Facebook activity is crucial,
especially as a growing amount of official Facebook pages have decided to create local
versions of their page and that these tend to perform better than their global counterparts
 233

(Social Bakers, 2012).

From a methodological point of view, overall, scales that are developed using different
linguistic samples have better cross-cultural applicability afterwards (e.g. Cadogan et al.,
1999). However, research instruments are often developed in a single linguistic setting and
then directly applied to another setting without considering theoretical fit, which can lead
to serious issues of construct bias (Douglas and Craig, 2006). The fact that the consumer
engagement scales exhibit group invariance between the two different languages is an
indication that the scale is conceptually and empirically applicable to different linguistic
contexts and possibly cultures, in the future.

Moreover, cross-context invariance support the inclusion of the consumer engagement

scales in a nomological network. Equal variability of the consumer engagement scales
across two languages suggests that correlation coefficients between this scale and other
related constructs should be comparable (Steenkamp and Baumgartner, 1998). This was
validated with the multi-group confirmatory factor analysis performed on the structural
model.

Linguistic group invariance was also found for the structural model. More specifically,
configural, measurement and structural invariance have been demonstrated. This finding
first implies that all construct measures included in the model are applicable across
languages, as well as their interrelationships. From a measurement standpoint, this is not
surprising. Apart from the consumer engagement scales, which were shown to be invariant,
all measurement scales used in the study have been adopted or adapted from existing
studies, and most of them had already been validated in various cultural settings.

From a structural standpoint, this implies that all hypothesis tests lead to the same results.
Although this was not proven to be entirely the case when performing the SEM tests on the
two samples separately, 10 out of 11 hypotheses (or 91 percent of the results) are the same
across samples. The results of this study therefore denote strong cultural similarities in
terms of what drives and results from consumer engagement in OBC embedded on social
media.

The only difference between the two samples has been found in the hypothesis linking
brand trust to brand loyalty. This difference in results can be explained by a series of
cultural and contextual factors differentiating the two languages. Firstly, English-speaking
and French-speaking cultures differ in terms of uncertainty avoidance, a major cultural
 234

dimension put forth in Hofstede’s cultural framework. French-speaking individuals are

known to exhibit higher degrees of uncertainty avoidance than English-speaking
individuals, according to Hofstede’s cultural framework and assessment tools (Hofstede,
2014). Uncertainty avoidance is related to risk, and has also been shown to be closely
related to the construct of trust (Doney et al., 1998). Following this logic, French-speaking
respondents are likely to ascribe increased levels of importance to brand trust. On this
basis, if and once brand trust is gained for the French speakers, is it probably more likely to
be more meaningful and have potent implications, such as loyalty, leading to the support of
H7 for the French sample.

On the other hand, English-speaking countries such as the UK or the US tend to score
much lower on the uncertainty avoidance dimension (Hofstede, 2014), meaning that trust
is not as important for them as it is for the French-speaking respondents. If trust is a less
potent cultural dimension in English-speaking countries, this can give an indication of why
brand trust fails to positively impact loyalty in this sample. The variation in loyalty for
countries with different levels of uncertainty avoidance has already been shown in cross-
cultural marketing research (El-Manstrly, 2014). Even if brand trust is gained, its
explanatory power over behavioural brand loyalty is likely to be weaker for the English-
speaking sample than it is for the French, leading to the different results for the two
samples.

Another potential explanation for this difference is that English-speaking countries tend to
be more advanced and innovative in terms of marketing techniques and brand management
(Garnier and McDonald, 2009). English-speaking consumers are therefore more exposed
to marketing content and have developed more understanding, as well as resistance to it.
Indeed, levels of marketing literacy in English-speaking countries tend to be higher
(Garnier and McDonald, 2009). This might be a reason why, even if a company is reliable
and worth their trust, English-speaking consumers will not necessarily keep buying from
them.

In addition to this, English is often the language in which international Facebook pages are
set up, meaning that, by definition, a brand operating under its English Facebook page is
likely to have many more global and local competitors than a local French or Spanish
page. For this reason, competition tends to be fiercer for OBCs in English and this might
contribute to offsetting the impact of brand trust built on social media on behavioural
loyalty.
 235

Moreover, if we consider that trust in this instance moderates the relationship between
brand engagement and brand loyalty, Hollebeek’s (2011) framework capturing the brand
engagement/loyalty nexus can prove useful. According to this framework, too much
engagement can lead to less loyalty. Indeed, when a consumer is too highly involved, he
can at some point feel drained and experience fatigue, which will decrease his/her loyalty
to the brand. Adding the mediating effect of brand trust, it might still be the case,
particularly in Anglo-Saxon cultures, that increased engagement eventually leads to less
loyalty, evidencing the complexity of the engagement/loyalty relationship.

Despite the difference between languages on the trust-loyalty relationship, this study’s
results seem to support the linguistic contextual validation it set out to achieve. As
explained in Chapter 4, the importance to determine contextual invariance of measures and
model is becoming a stringent need inherent to globalised environments such as social
media, where geographic boundaries are virtually inexistent (Andersen, 2005; Kaplan and
Haenlein, 2010; Faraj and Johnson, 2011). This study proposes a first step in addressing
this requirement by showing that: (1) consumer engagement, both with the OBC and brand
partners exists in the same way for English and French-speaking respondents; (2)
consumer engagement is triggered in the same way by individual and social factors for
English and French-speaking respondents; (3) consumer engagement equally leads to
brand trust and brand commitment in for English and French-speaking respondent and (4)
consumer engagement ultimately affects brand loyalty in a positive way across linguistic
groups, although trust cannot be relied on as effectively in Anglo-Saxon settings.

These findings have potent implications for the management of OBC on social media and
suggest that similar consumer engagement strategies can be replicated across local
Facebook pages in different languages. This cross-linguistic validation brings an important
contribution to social media studies in particular, which have so far been rather timid in
testing their results in various national or linguistic contexts (Okazaki and Taylor, 2013).
In a context of disputed cultural convergence, finding out whether results related to online
consumer behaviour and brand management are geographically transposable is urgently
needed (ibid., 2013). These results are a first step in this direction and their implications
for theory and practice are discussed in the concluding chapter.
 236

7.5. Summary

The objective of this chapter was to interpret the results of the data analysis presented in
Chapters 5 and 6. In order to do so, the findings were discussed in light of the framework
and model presented in Chapter 3 and its related research hypotheses. The results were
interpreted and their implications discussed mindful of recent research in the OBC and
consumer engagement literature.

The concept of consumer engagement in an OBC context has been conceptualised in a way
that extends and deepens existing dimensionalisations of consumer engagement.
Engagement is composed of three dimensions, reflected in seven sub-dimensions. By
adequately measuring engagement in an OBC context, the scale development is
particularly innovative in using two engagement partners concurrently: the brand and the
OBC.

Regarding the hypotheses, this study shows support for most of them and evidences the
central role of consumer engagement in the OBC context. Consumer engagement is driven
by some individual traits and pre-dispositions, which signals that differences of
engagement levels can be attributed to enduring personal characteristics, an innovative
empirical finding of this study. The recognition of these characteristics and the clustering
of OBC members on this basis is therefore an important endeavour to sustain consumer
engagement and adequately manage consumers based on their individual differences.

The results also show that although individual traits largely explain OBC engagement,
online brand engagement tends to rely less on individual characteristics (only one out of
three) and more on OBC engagement. This implies that the vitality of direct brand
behaviours, affect and cognition depends on the consumer’s level of engagement with
other consumers. C2C interactions are thus critical to triggering and sustaining online
brand engagement and an essential step in the formation of sustained brand relationships,
as the consequence of online brand engagement show.

Lastly, brand relationship outcomes are evidenced in the form of brand trust, brand
commitment and brand loyalty as expressed by repeat purchase. Consumer engagement as
an interactive, social and multi-dimension construct therefore finds its place in relationship
marketing models and proves to be a strong predictor of relationship quality and ensuing
 237

brand sales. These findings are largely consistent across two cultural contexts, as supported
by the group invariance results between the French- and English-speaking samples.

These findings contribute to the extant OBC and consumer engagement literature and have
several implications for the practice of OBC and brand relationship management, which
are discussed in the following concluding chapter.
 238

Chapter 8: Conclusion

8.1. Introduction

This study set out to advance the concept of consumer engagement as a valid approach to
the conceptualisation and measurement of online brand community participation and
examine some of its antecedents and outcomes. The thesis starts by demonstrating the
importance of this endeavour, pointing out the relevance of studying OBC participation
and how consumer engagement constitutes a better approach to OBC participation than
other existing views. The key argument advanced here is that the current approaches to
measuring participation in the context of online brand communities fail to account fully for
a multi-faceted notion of participation, one that is suited for the interactive nature of the
online contexts. A critical analysis of the OBC literature highlights the shortcomings of
current research in understanding OBC participation and consumer engagement appears as
a potent new concept to address the identified gaps. Put simply, consumer engagement
provides an approach to OBC participation that is more holistic, social and interactive.

Building on literature from multiple strands, the thesis then advances the concept of
consumer engagement as valid alternative to measure online participation. Based on the
extant literature on consumer engagement in marketing, and building from studies in other
fields of the social sciences, the study has proposed a framework of consumer engagement
in OBC, as well as a conceptual model of its hypothetical antecedents and outcomes. The
model presents consumer engagement in OBC as having one subject (the consumer) and
two partners (the brand and the OBC). The concept is composed of three dimensions,
which are further broken down into seven sub-dimensions. Consumer engagement in OBC
is hypothesised to derive from three individual traits and predispositions and result in three
brand-relationship outcomes. Engagement with the OBC is also hypothesised to precede
and impact online engagement with the brand. The integration of multiple theoretical
strands results in a conceptual model encompassing 11 hypotheses.

To test these hypotheses, the study has adopted a post-positivist methodological approach.
The main findings are derived from the analysis of quantitative data collected over four
 239

months from a total of 721 members of French- and English-speaking Facebook pages.
The communities examined cover a broad range of categories, which seem to adequately
capture the diversity of OBCs. The study data have been analysed using validated
confirmatory factor analyses and structural equation modelling techniques. Prior to testing
the eleven hypotheses, the study has offered a measure of consumer engagement,
following strict scale development procedures involving qualitative data, a panel of
academic experts and iterative item generation. Psychometric properties of the scales are
further evidenced through tests on a calibration (n=224) and validation sample (n=224).

Hypotheses were then tested using the English validation sample (n=224) and the French
sample (n=273) in parallel. Results provide support for the majority of the hypotheses (8
out of 11 for the English sample, and 9 out of 11 for the French sample). Data reveal that
OBC engagement is strongly dependent on individual traits and predispositions, and that it
strongly contributes to increased levels of online brand engagement, along with product
involvement. Online brand engagement has a powerful impact on brand loyalty through the
mediating impact of brand trust and commitment. These results are almost identical across
the two linguistic settings under investigation, as confirmed by multi-group analyses.

8.1. Theoretical contributions

The findings from this study significantly contribute to both OBC literature and consumer
engagement research in several aspects. The first contribution concerns the
comprehensive advancement and reconceptualisation of OBC participation through
consumer engagement. Member participation, affect and cognition in OBCs so far
benefited from a scattered and incomplete treatment, resulting in a plethora of overlapping
and intertwined concepts, theories and methodological approaches (Preece and Malhoney-
Krishmar, 2005). This study enhances the current theoretical and practical understanding
of OBC participation by proposing and empirically validating a framework of consumer
engagement in OBC that is more cognisant of online contexts. Advancing the notion of
consumer engagement as a valid conceptualisation of OBC participation, the
multidimensional, interactive and social nature of OBC participation is recognised and
integrated under a single construct. By doing so, a significant advance is granted to the
state of OBC participation (Dholakia et al., 2004; Stockburger, 2010) research and by
 240

extension, to online consumer behaviour in relation to brands (Mathwick, 2002; Labrecque

et al., 2013). The study also proposes that two engagement partners exist in OBCs and that
both need to be taken into account if one is to measure consumer engagement in an OBC
context. The findings show that their interrelationship seems to be a crucial element.
Through this exercise, precise insight is gained in terms of what engagement means and
how it can be operationalised. Although the framework is adapted to an OBC context by
focusing on engagement with a brand and with a community, this study aims to provide a
generic multidimensional conceptualisation of consumer engagement that is replicable in
any consumption-related context beyond OBCs, and with any engagement partner. In this
sense, the conceptual framework contributes to the wider consumer engagement literature
as well.

Indeed, the second contribution regards the conceptualisation of consumer

engagement. The proposed conceptual framework deepens the dimensionality of
consumer engagement at a level of detail so far unprecedented in the marketing literature.
Building on engagement research in marketing and incorporating insight from other fields
of the social sciences, dimensions of engagement are clarified and broken down into sub-
dimensions. These include enthusiasm, enjoyment, attention, absorption, learning, sharing
and endorsing. Further contribution lies in this study’s explicit recognition of two different
loci of engagement in the context of OBC participation. By showing that OBC engagement
and online brand engagement are two interrelated facets of consumer engagement, this
study also shows the need to break away from traditional one-on-one approaches to brand
relationships. This relates to an urge for more social orientation toward brand relationship
management, where C2C interactions are rightfully acknowledged and tapped into
(Thought Economics, 2009; Baird and Parasnis, 2011; Ashley and Tuten, 2015),
particularly in online settings.

A third significant theoretical contribution concerns the identification and validation

of antecedents and outcomes of engagement. The study develops and tests a causal
model explaining the role of individual traits and predispositions driving engagement and
its impact on consumer-brand relationships. This answers the call for empirical research
into the drivers and outcomes of consumer engagement (van Doorn et al., 2010), while
strengthening the position of consumer engagement as the appropriate approach to OBC
participation. Indeed, one of the key conclusions from the review of the consumer
engagement literature was that it cruelly lacked empirical validation of its conceptual
 241

propositions in terms of engagement antecedents and consequences, leaving engagement to

be a very attractive, yet extremely abstract marketing concept (Hollebeek et al., 2014).

A first understanding of the driving factors of increased levels of consumer engagement in

the OBC context is provided. Although some hypotheses are not supported, the findings
show that online brand community and brand engagement are to some extent influenced by
individual traits and predispositions (OIP, attitude toward OBC participation and product
involvement). This contributes to the consumer engagement research, which had so far had
difficulties pinpointing clear triggers to consumer engagement and, in particular, drivers
related to the individual consumer (Wirtz et al., 2013). Although van Doorn et al. (2010)
and Gambetti and Graffignia (2010) do acknowledge that individual traits and pre-
dispositions may affect the likelihood of consumer engagement and impact cognitive and
behavioural processes, this proposition had remained conceptual to date.

This study brings clarity into the individual traits and predispositions that trigger consumer
engagement in OBC. This also brings OBC research further, as extant literature exploring
the drivers of OBC participation had so far mainly focused on perceived value and
benefits, as well as social motivations, as identified in the literature review. OBC members
therefore differ not only in terms of their individual or social motivations and behaviours,
but also in terms of their inherent and enduring individual traits and characteristics.

The role of consumer engagement in enhancing consumer-brand relationships is verified.

This is achieved through a series of hypotheses linking consumer engagement to brand
trust, commitment and loyalty. These findings also significantly enrich the existing OBC
and consumer engagement literature. Both streams of literature had indeed identified brand
relationship building, including brand loyalty, as one of the top benefit sought after as a
result of consumer engagement. Research in this area was emerging in the OBC literature,
and remained conceptual in the area of consumer engagement. By proving that engaged
consumers show significantly higher levels of brand trust, commitment and subsequently
loyalty, this study timely addresses this issue and enriches the existing literature.
 242

8.3. Methodological contributions

The first methodological contribution concerns the operationalisation of consumer

engagement with several achievements. Firstly, the study advances an innovative
measure of consumer engagement in OBC. Integrating theory to arrive at a novel
conceptualisation, the conceptual framework provided the basis to the development of two
mirrored scales of consumer engagement, namely OBC engagement and online brand
engagement. The robust scale development procedure culminated in the creation of two
identical scales of consumer engagement. The procedure involved both qualitative and
quantitative data, as well as a panel of academic experts, and the scales both show high
degrees of reliability and validity. This endeavour provides a major contribution to the
existing literature on consumer engagement measurement, which is to date extremely
limited, particularly in the OBC context. This way, the study answers the call by Brodie et
al. (2011) and Mollen and Wilson (2010) to create a consumer engagement scale in an
online context. The scale development contributes significantly to empirically validating
the so far hesitant understanding of consumer engagement dimensionality. It reveals that
consumer engagement is best conceptualised and measured as a multi-dimensional
construct, and that each of its dimensions can be further broken down into sub-dimensions.
This provides a potent explanation of the exact meaning and applications of consumer
engagement.

Secondly, the generalisability of the study’s result has been evidenced thanks to a
validation procedure in different linguistic contexts. The findings show measurement
and structural invariance on the data, which, bar for one hypothesis, evidence complete
similarity of results in different contexts. In addition to strengthening the conceptual and
empirical validity of the study, this paves the way for more cross-context studies in social
media and OBC studies, which are currently very limited in scope and application.
However, the inconsistency surrounding one of the hypotheses highlights potential effects
of culture and more studies are needed to further explore these effects.
 243

8.4. Managerial implications

The results of this study provide valuable insight for online marketers on how to use OBC
embedded on social media as a channel to successfully enhance consumer-brand
relationships. Understanding what drives consumer engagement, how the community and
brand engagement should interplay and the relational benefits for brands is crucial for
managers. By addressing these issues, this study bears direct implications on brands’
content posting and online CRM strategies. The managerial implications of this thesis are
thus numerous and address key issues of OBC management (Gensler et al., 2013). In other
words, the study give pointers on how brands can be successful engagement partners in the
OBC they have created.

The first set of valuable guidelines for managers concern the key individual traits and
predispositions that motivate an active OBC participation (Chang et al., 2013). On this
ground, several ways to effectively build consumer engagement can be envisaged, as it is
the task of the marketer to identify who are their interactive consumer segments in OBCs
and recognise their individual specificities. For instance, segmentation of OBC users based
on their individual traits and characteristics should be done to drive community
engagement. Since consumers differ in terms of their OIP, and their attitude toward OBC
participation and product involvement, consumers can be segmented according to these
criteria and engaged with in a dedicated way.

More specifically, this pertains directly to the communication and content strategy of OBC
managers. For instance, while targeting community members with higher levels of product
involvement, OBC managers should seek to create engagement around general product
category information that will initiate category-level thinking and interest. This way, they
can try to best address their needs by means of interactive brand engagement and fostering
OBC engagement. For instance, Made&More, an online fashion retailer selling exclusively
goods manufactured in Europe in sustainable conditions, successfully engage their
members by regularly posting content about slow fashion. This type of informational
content foster in-group interaction and the fact that the company takes a stance on this
issue also fosters direct brand engagement. Given the link between product involvement
and engagement and the fact that information-based content strategies work well with
high-involvement consumers (Warrington and Shim, 2000), this suggests that information-
 244

rich content needs to be included by brands on their OBC. Overall, OBC managers should
aim to post content that is brand-related (directly about the brand or related topics), but
always remain on-topic, as congruency between the OBC purpose and content is known to
drive higher engagement behaviours (Breitshol et al., 2015).

Recognising that OIP and attitude toward the community are driving OBC engagement,
this study also suggests that brands should seek to activate their most interaction-prone
members to engage with other members. In terms of the content posted, this can be
achieved, for instance through educational posts. A number of studies show that highly
informational content that allows improving brand use can bear high C2C interaction
potential (Schau et al., 2009; Malhotra et al., 2012). If companies manage to engage
members with positive attitudes and interaction propensity, their OBC engagement will
naturally have a ripple effect and involve other less active OBC members, thus activating
the community as a whole.

Secondly, the findings of this study highlight the preceding role of OBC engagement over
brand engagement, a relationship that has major repercussions in terms of OBC
management (Fournier and Lee, 2009). Practically, this means that OBC managers should
strive to create engagement among consumers prior to directly with the brand. As
Colliander et al. (2015, p.11) point out, ‘effectiveness of OBCs is contingent on companies
realizing that ‘social’ is the operative word in social media’. As a result, brands need to
foster C2C interactions and provide an OBC environment conducive of community
interactions and freedom of expression. Brands should be eager to create a community of
users first, of which they are the catalyst and common denominator, but sit back initially to
allow for the community to form and get strong (Fournier and Lee, 2009).

Understanding that they need to build the community first prior to pushing promotional
brand content is crucial as a lot of companies still use social media and OBC simply for
broadcasting their own message (Colliander et al., 2015). Brand should not be self-centred
and push their own content; rather, they should foster OBC engagement as a priority
(Fournier and Lee, 2009). Brands should aim to be part of the conversation rather than
monopolise it and make it all about themselves (Malhotra et al., 2012). In Facebook
settings in particular, leaving the brand in the background to focus on broader issues is
likely to increase engagement (ibid., 2012). The results of the present study show that once
strong OBC engagement is asserted, consumers are more likely to engage with the focal
brand, rather than the other way around. As such, online brand engagement can be seen as
 245

an organic derivative from OBC engagement, which comes at later stages of the
community life cycle.

Building online brand engagement is, however, crucial if one wants to achieve brand-
relationship building. Therefore, after fostering OBC engagement among members, brands
need to activate their members in interactive participation with them. Simply moderating
the OBC is not enough, and one should not forget that the OBC should also revolve around
brand-related topics (Malhotra et al., 2012). Therefore, brands need to generate meaningful
direct interactions and conversations with their members. To achieve this aim, brands need
to adopt an open and positive approach to discussing with their consumers. Interactive
engagement with consumers is an indication that the company cares enough to discuss,
rather than just push their content, and this is likely to result in higher degrees of purchase
intentions (Colliander et al., 2015) and brand-relationship quality (Hollebeek, 2011a), as
evidenced in this study. A practical way to generate direct brand engagement and show
that the brand cares about their consumers is to ask them questions or invite them to take
action. Inviting consumers to share their views and answer a question, or incentivising
content sharing, are two ways to initiate direct online brand engagement.

Another managerial implication from the study concerns the multidimensionality and long-
term perspective of consumer engagement (Brodie et al., 2011). Managers need to adopt a
long-term, enduring and multi-dimensional perspective to their OBC management. Put
differently, brands need to move away from short term, behaviour-only and metrics-based
measures of OBC effectiveness. As much as the benefits and value of OBC activities need
to be measured, neither consumer engagement nor relationship building are likely to
happen overnight. This is indicated on the one hand by the very nature of all relevant
constructs tested in this study’s model: from the drivers to the outcomes of engagement, all
aspects of OBC participation have an enduring, long term nature.

Secondly, although managers have tended to measure OBC participation based on metrics
(such as number of likes, comments, time spent online, or number of members), the
findings of this study show that engagement is a multidimensional construct which goes
much further than considering members’ actions and activities (e.g. Gummerus et al.,
2012). Consumer engagement is not about big data. Rather, it encompasses emotions and
cognitive processing that require time and effort to develop. In this sense, basing one’s
understanding of consumer engagement in OBC solely on equations involving site metric
is a delusional, short-sighted approach unlikely to result in strong brand relationships.
 246

By showing the necessary multidimensionality of consumer engagement, this study offers

managers empirical evidence that their OBC strategy need to activate all three aspects. For
instance, emotions call emotions: research has shown that high social media engagement is
driven by emotional appeals (Ashely and Tuten, 2015). Therefore, content that humanises
the brand or gamifies it (brand entertainment) will go a long way into fostering
consumers’, enjoyment and excitement about the brand. On the other hand, behavioural
engagement is likely to increase with topical, informative and visual content, as well as
direct calls-to-action (Malhotra et al., 2012). Overall, this study shows that brands should
use a wide variety of appeals, content forms and types to foster deep, long-lasting and
multi-dimensional engagement, as recently suggested by Ashley and Tuten (2015).

Lastly, the results of this study are broadly indicative of the ability for OBC managers to
replicate social media strategies across OBCs in different languages, especially when the
similar results in terms of brand relationship building are expected. In the twenty-first
century, the role of cross-cultural marketing research has become increasingly important in
driving managerial decisions (Slater and Yani-de-Soriano, 2010). The results suggest that
there is no significant difference for most of the hypothesised paths across cultures,
allowing brand page managers to use a standardised marketing strategy for their OBCs and
Facebook pages across cultures. However, no single study can resolve the debate of the
standardised versus localised marketing strategy and further investigation into the cross-
cultural applicability of the results is warranted.

8.5. Limitations and directions for future research

Despite the valuable contributions that this study brings to the OBC and consumer
engagement literatures, the current examination is not without limitations, and further
research is warranted to explore the fascinating realm of consumer engagement in OBC.
Limitations concern the type of data collected, sampling and generalisability of results, as
well as the limitations inherent to the conceptual scope of the study. Several suggestions
are made in order to advance research in this emerging domain.

Firstly, this study has limitations pertaining to the type of data collected. Using
predominantly quantitative data focused on consumer engagement in OBC, the research
 247

design inherently aims to capture complex phenomena through numbers. As such, and
despite a supportive qualitative phase, it necessarily adopts a reductionist approach, one
that might not capture the full depth of the phenomena under investigation. A way to gain
depth into consumer engagement in OBC would be to complement this study’s data with
qualitative approaches and follow-up with more in-depth interviews, like a number of OBC
and consumer engagement studies have done to date (e.g. Schau et al., 2009; Brodie et al.,
2013).

Additionally, all the hypotheses are tested based on a cross-sectional design. Such data is
being collected at one point in time, and therefore it is impossible to have an indication of
the sequence of events. The results therefore only indicate a relationship between the
variables but do not confirm causality but only covariance. In order to tackle causality
more explicitly, follow-up studies could be carried out and longitudinal data collected in
order to fully understand the causal relationships between consumer engagement, its
antecedents and outcomes.

Moreover, this study is based on consumer self-reported survey data. Despite the clear
advantages of such methods and type of data, further research should seek to gain access
into ‘organic data’ (Murphy et al., 2014). Organic data, in contrast to ‘designed data’ like
the one used in this survey, emerge out of communication technologies and are available
either for a fee or for free (ibid., 2014). An example of organic data that has flourished in
online community and social media studies is netnographic data (Kozinets, 2002; Kozinets,
2010), which would allow capturing actual manifestations of engagement rather than
personal reflections of these manifestations. Sentiment or content analysis of social media
content could also be used and integrated with the analysis of netnographic data (Murphy
et al., 2014).

The study sampling approach and the resulting sample represent a second
shortcoming of this study. The nature of the population and study context did not allow
obtaining a probabilistic random sample, which is an issue inherent to OBC research
(Preece and Maloney-Krichmar, 2005), where case studies of single communities still
prevail (e.g. Maztler et al., 2011). Although the large sample size offsets the risk of bias,
the non-probabilistic research sample has implications for the gereralisability of the results.
The inability to estimate a response rate might also have lead to bias. Effort was made to
alleviate this issue as much as possible by adopting a thorough two-level sample design.
However, due to this approach, once the survey was posted on the page, no further control
 248

could be exerted to ensure that the respondents of the survey were indeed members of the
community. Further studies may try to avoid these sampling issues through methodological
advances or by using even larger sample sizes.

Beyond sampling, generalisability of the study could be enhanced in a number of

ways. The study focused on OBCs situated in one social network, Facebook. One way to
extend this study’s findings and enhance its generalisability would be to consider other
OBC platforms than social media, or Facebook pages. This study is based on a conscious
choice to focus on one type of social media platform and consumer grouping to represent
OBCs, however OBCs can thrive in other formats and environments such as blogs or wikis
(Kaplan and Haenlein, 2010), or other social media like Twitter or Youtube. Different
platforms afford different interactive functionalities, which could impact the way consumer
engagement is enacted (Hollebeek et al., 2014). As online platforms keep growing in size,
evolving in form and expanding in terms of marketing applications, it is expected that the
number of online consumer engagement options will grow exponentially.

This study is also limited in its ability to more effectively control for the effects of
different brands, or brand categories. The selection of communities included in the sample
was based on a series of criteria and aimed to represent all categories of Facebook pages
fitting the study requirements. Collecting data on a number of different product categories
was desirable since OBC research to date has largely collected data from one product
category at a time. However, time and access constraints forced the researcher to target a
very broad range of pages to secure access to enough data. As a result, sample spread
across product categories is not even. Further research is encouraged to draw from larger
samples of specific brand types in order to be able to statistically verify if there are
differences of engagement levels or relationships for different types of brand. This research
aimed to include as many types of brands as possible but did not allow to directly compare
brands or brand categories, due to a lack of consistency of the representation of each brand
category.

This study also paved the way in exploring consumer engagement in OBCs across cultures,
focusing on French and English-speaking communities. However, the comparison of two
languages only does not allow drawing strong conclusions on the cross-cultural validity of
a test (Cadogan, 2010). In order to further the cross-cultural applicability of the model,
scholars need to collect data from more than two cultures. More specifically, the
comparison between very individualistic and very collectivistic cultures would be of
 249

interest, as previous studies suggested that engagement in online communities is higher in

collectivist cultures (Park and Jun, 2003) and driven by more social motives (Madupu and
Cooley, 2010). Focusing on two Western cultures is a first step in showing the cross-
cultural validity of the study, but further confirmation is needed concerning other, more
culturally diverse nations.

A final possibility to extend this work further is to reconsider the conceptual frame.
The conceptual model presented here builds on key studies in OBC literature to generate a
conceptual framework. Necessarily, the number of antecedents is small and finite and other
antecedents and outcomes of consumer engagement in OBC may need to be explored in
future research. Conceptual research in consumer engagement has highlighted a plethora of
possible connections between consumer engagement and other relational and social
constructs (e.g. Hollebeek, 2011a). As empirical research aiming to validate these
relationships is only slowly emerging, engagement and OBC scholars should seek to
further validate the relationships that link consumer engagement to other constructs. More
specifically, studies could look at the impact of perceived costs and benefits over
engagement (van Doorn et al., 2010), or the impact it bears on brand recall and attention
(Sprott et al., 2009) or brand experiences (Hollebeek, 2011).

Other consumer engagement partners should also be considered. The developed measure
of consumer engagement has only been applied to two engagement partners in this study.
Since the consumer engagement scale is applicable to consumer-to-consumer relationships
and consumer-to-brand relationships, building on the framework of brand community
relationships by McAlexander et al. (2002)’s, it would be worthwhile to test the
generalisability of the scale to other relevant engagement objects such as the company, or
the product, within and outside the OBC contexts. Similarly, consumer engagement
research has suggested that consumers can be engaged with objects ranging from an event
to a piece of media. These constitute further engagement partners to which the developed
scale could be applied. This could also work to provide further validation of the consumer
engagement scale in different consumption-related contexts other than OBCs.

Eventually, further research could elaborate on the dynamics, interplay and specificities of
consumer engagement dimensions. Recent studies suggest that behavioural engagement
can follows a hierarchical sequence (Schivinski et al., forthcoming). Following this logic,
affective, cognitive and behavioural forms of engagement could also occur in sequence. It
might also be the case that some individuals exhibit extremely high levels of emotions
 250

toward a brand or brand community, but fail to exhibit high engagement behaviours, as
suggested by recent studies on consumer engagement (de Villiers, 2015). For instance, an
individual sharing extensively with a brand or community might be engaged in this
behaviour rather mindlessly and performing rather low on the cognitive dimension.
Similarly, a community member might be extremely highly emotionally engaged with a
brand but be reluctant to share on online social platforms, hence not enacting the
behavioural aspect of engagement to its full potential. Idiosyncratic consumer behaviour
might appear online at an individual or group level and these differences might also be
linked to engagement antecedents, outcomes and partners variations.

8.6. Summary

Notwithstanding some limitations, the study makes several significant contributions to the
fields of OBC and consumer engagement in terms of theoretical advances, methodology
and practice. This thesis proposed the notion of consumer engagement as a valid approach
to OBC participation. The findings indicate that consumer engagement in OBC is
dependent on three individual consumer traits and that it significantly contributes to
enhancing consumer-brand relationships. Two scales to capture consumer engagement
with the OBC and the brand were proposed, following strict scale development procedures.

The findings of the thesis reveal the need for a multidimensional, interactive and social
approach of OBC participation. The importance of considering a variety of engagement
partners relevant to given engagement contexts is also of particular interest, despite being
so far under researched. Empirical support is granted to the testing of relationships
between consumer engagement and other individual and relational factors, renewing the
call for further empirical work in this direction. The findings of this study advance the
growing and important research areas of consumer engagement and OBC participation.
 251

Appendixes
 252

Appendix 1: Cover letter to Facebook page

administrators

Dear administrator of [Facebook page name],

My name is Laurence Dessart, and I am the lead researcher for the Global Social Media
Survey launched by The University of Glasgow, Scotland. The goal of this study is to
understand consumer engagement on Facebook, and in particular:

1° what brand page engagement and brand engagement consist of;

2° what generates them;
3° which benefits they bring to brands.
In order to do so, we need access to members of Facebook pages. Your page has been
selected as extremely representative for this study, based on its high engagement profile.

We understand that page participation relies on mutual trust and relevant content. To avoid
spamming your page, we wish to have your help in posting a link to the survey, in
exchange of its results.

Results will be available upon request, and pages providing over 100 responses will
receive detailed analysis for their own brand.

We suggest sharing the link to the survey with a post of the type: ‘Help us create a better
page for you. Answer this survey about your experience on our page
https://www.surveymonkey.com/s/SocialMediaGlasgow‘. The survey is incentivised with 3
x £100 Amazon vouchers, which you can also mention in your post.

If you want to know more about the survey before promoting it, feel free to contact me.

Thank you for your time and help.

Best regards,

Laurence Dessart
[email protected]. PhD Researcher, University of Glasgow.
 253

Appendix 2: The online consumer survey (English

version)

Page 1: Screening

1. Are you 18 years old or above?

□ Yes  Logic: go to page 2

□ No  Logic: go to page 3

Page 2: Disqualification

We are sorry, but if you are younger than 18 years old, you do not qualify to answer this
survey. We thank you for your time anyway.

Page 3: Welcome

Thank you for taking part in this Global Social Media Survey by the University of
Glasgow.
You will be asked to answer questions about your interactions with the Facebook page that
posted the link to this survey.

The study is anonymous and follows the University of Glasgow ethics. For more details,
copy/paste this link in your browser:
https://drive.google.com/file/d/0B5UWs_j5SRuzYWN5SXRZOXhIcDg/edit?usp=sharing

The questionnaire takes about 15 minutes to answer, and there are three £100 Amazon
vouchers to be won and this only for fully answered questionnaires.

Page 4: Online Activity

The following questions are about your general online activity.

2. On average, how many hours per day do you spend online? [dropdown]

□ 0-1 □ 2-3 □ 4-5 □6-8 □ 8+

3. Indicate on a scale from 1 (completely disagree) to 7 (completely agree) to which

extent you agree with the following statements.

1 2 3 4 5 6 7
In general, I like to get involved in online discussions
I am someone who enjoys interacting with like-minded
people online
I am someone who likes actively participating in online
 254

discussions
In general, I thoroughly enjoy exchanging ideas with
others online

Page 5: Facebook Activity

The following questions are about your Facebook activity

4. When did you join Facebook? [dropdown]

□ 2004 □ 2005 □ 2006 □ 2007 □ 2008 □ 2009 □ 2010 □ 2011 □ 2012 □ 2013 □ 2014

5. How many times per day do you log on to Facebook? [dropdown]

□ I don’t log on every day

□ 1 to 3
□ 4 to 6
□6+
□ All the time: I get notifications on my phone

6. In a typical day, roughly how much time do you spend on Facebook? [dropdown]

□ Less than 10 minutes

□ 10 to 30 min
□ 31 to 60 min
□ 60 min +
 255

Page 6: Identification of the brand page

7. The link to this survey was posted by a page that you like, which has been carefully
selected. It is important that you answer this survey keeping in mind this particular
page, and the brand that is represents. Please tell us which brand this page is
about:…

8. Are you a customer of this brand?

□ Yes
□ No, I like the page but I have never bought the brand

Page 7: Brand-related variables

The following questions are about the brand that you have just identified.

9. Indicate on a scale from 1 (completely disagree) to 7 (completely agree) to which

extent you agree with the following statements

1 2 3 4 5 6 7
I have grown to like this brand more than others offering
the same product/service
I like the product/services offered by this brand
To me, this brand is the one whose product/services I
enjoy using most
I trust this brand
I rely on this brand
This is an honest brand
This brand is safe

Page 8: Affective brand engagement

10. Indicate on a scale from 1 (completely disagree) to 7 (completely agree) to which

extent you agree with the following statements.

1 2 3 4 5 6 7
I feel enthusiastic about the brand
The brand makes me enthusiastic
I am heavily into this brand
I am interested in anything about this brand
I find this brand interesting
I enjoy interacting with the brand
When interacting with the brand, I feel happy
I get pleasure from interacting with the brand
Interacting with the brand is like a treat for me
 256

Page 9: Cognitive brand engagement

11. Indicate on a scale from 1 (completely disagree) to 7 (completely agree) to which

extent you agree with the following statements.

1 2 3 4 5 6 7
I pay a lot of attention to the brand
Things related to the brand grab my attention
I spend a lot of time thinking about the brand
I make time to think about the brand
When interacting with this brand, I forget everything
else around me
Time flies when I am interacting with this brand
When I am interacting with this brand, I get carried away
When interacting with this brand, it is difficult to detach
myself

Page 10: Behavioural brand engagement

The following questions are about your interactions with the administrators of the page,
which are usually the brand managers. Think about all the interactions below as done
through ‘comments’, ‘likes’, ‘shares’, etc.

12. Indicate on a scale from 1 (completely disagree) to 7 (completely agree) to which

extent you agree with the following statements.

1 2 3 4 5 6 7
I reply to the questions of the brand page managers
I share my opinion with them
I share my experiences with them
I share my ideas with them
I share interesting content with them
I help them
I ask them questions
I seek ideas, or information from them
I seek help from them
I learn from the content they provide
I show support to what they say or do
I share their content to my wider network
I promote the brand
I try to get other interested in the brand
I actively defend the brand from its critics
I say positive things about the brand to other people

Page 11: Product Category

The following questions are about the type of product the brand belongs to.
 257

13. Which type of products does the brand belong to?

□ Food and beverage

□ Technology (software, telecom, computer products…)
□ Services (bank, insurance, education…)
□ Travel (airline, railways, travel agents…)
□ Fashion and beauty
□ Durable goods (automobile, electronics, home appliances)
□ Retail (stores, supermarket, e-shops)
□ Entertainment (sports, films, series, books, games, …)
□ Other, please specify: ………………………………….

14. Considering this type of product, indicate on a scale from 1 (completely disagree)
to 7 (completely agree) to which extent you agree with the following statements.

1 2 3 4 5 6 7
This type of product is very important to me
This type of product matters to me
When you buy this type of product, it’s a big deal if you
make a mistake
When you buy this type of product, it’s hard to make a
bad choice
I particularly like this type of product
You can really tell a lot about a person by the type of
product he/she picks out

15. Consider your overall purchases of this type of product. Indicate on a scale from 1
(completely disagree) to 7 (completely agree) to which extent you agree with the
following statements.

1 2 3 4 5 6 7
I am loyal to only one brand (the one I follow), when I
buy this type of product
For my next purchase, I will buy this brand again
I always buy this brand
I usually buy this brand

Page 12: Community-related variables

Now consider the page to answer the following set of questions.

16. For approximately how long have you liked the page?
□ Less than a year □ 1 – 5 years □ 5 – 10 years
17. How often do you actively click on the page?
□ Never □ Less than once a month □ About once a month □ About once a week
□ More than once a week
 258

18. How much time do you spend on the page per week?
□ 0 – 2 min □ 3 – 5 min □ 6 -10 min □ 11 – 15 min □ 15 min +
19. How big would you say the page is, in term of number of members?
□ Very small □ Fairly small □ Medium □ Fairly big □Very big
20. On the following scales, please express your attitude toward participating on the page
sometime during the next month

1 2 3 4 5 6 7
1 = Foolish; 7 = Wise
1 = Harmful; 7 = Beneficial
1 = Bad; 7 = Good
1 = Punishing; 7 = Rewarding

Page 13: Affective OBC engagement

21. Indicate on a scale from 1 (completely disagree) to 7 (completely agree) to which

extent you agree with the following statements.

1 2 3 4 5 6 7
I feel enthusiastic about the page
The page makes me enthusiastic
I am heavily into this page
I am interested in anything about this page
I find this page interesting
I enjoy interacting with the page members
When interacting with the page members, I feel happy
I get pleasure from page participation
Participating on the page is like a treat for me

Page 14: Cognitive OBC engagement

22. Indicate on a scale from 1 (completely disagree) to 7 (completely agree) to which

extent you agree with the following statements.

1 2 3 4 5 6 7
I pay a lot of attention to the page
Things related to the page grab my attention
I spend a lot of time thinking about the page
I make time to think about the page
When interacting with page members, I forget
everything else around me
Time flies when I am interacting with the page members
When I am interacting with the page members, I get
carried away
When interacting with the page members, it is difficult
to detach myself
 259

Page 15: Behavioural OBC engagement

The following questions are about your interactions with the other page members.
All these interactions can be done by commenting, sharing, posting, liking, etc.
23. Indicate on a scale from 1 (completely disagree) to 7 (completely agree) to which
extent you agree with the following statements.

1 2 3 4 5 6 7
I reply to their questions
I share my opinion with them
I share my experiences with them
I share my ideas with them
I share interesting content with them
I help them
I ask them questions
I seek ideas, or information from them
I seek help from them
I learn from the content they provide
I show support to what they say or do
I share the content they posted to my wider network
I promote the page
I try to get other interested in the page
I actively defend the page from its critics
I say positive things about the page to other people

Page 16: Demographics

24. What is your gender?

□ Male □ Female
25. What year were you born in? [dropdown 1996 – 1930]
26. What is the highest level of education you have completed?
□ Primary School □Secondary School □ Undergraduate degree □ Postgraduate
degree
27. What is your nationality? [dropdown with all countries]
28. What is your country of residence? [dropdown with all countries]

Page 17: Thank you

Your answers have been recorded. Thank you very much for completing this survey!
If you wish to take part in the Amazon voucher draw, please input your email address here
 260

Appendix 3: The online consumer survey (French

version)

Page 1: Question filtre

1. Etes-vous âgé de 18 ans ou plus?

□ Oui  Logique: aller à la 2

□ Non  Logique: aller à la 3

Page 2: Disqualification

Malheureusement, l’étude n’est pas accessible aux moins de 18 ans. Nous vous remercions
néanmoins pour votre temps.

Page 3: Bienvenue

Merci pour votre participation à cette Etude Internationale sur les Réseaux Sociaux. Nous
allons vous demander de répondre à quelques questions concernant votre participation sur
la page Facebook qui a posté le lien vers cette étude.

Cette étude est anonyme et régie par les règles éthiques de l’Université de Glasgow,
RoyaumeUni.
Pour plus de détails, veuillez copier/coller ce lien:
https://drive.google.com/file/d/0B5UWs_j5SRuzUE1qZmc0TElmVkk/edit?usp=sharing.

Répondre au questionnaire prendra environ 15 minutes. Vous aurez l'option de participer

au tirage au sort permettant de gagner un des trois coupons Amazon de €100 mis en jeu, et
ce uniquement en cas de questionnaire entierement complété.

Page 4: Activité en ligne

Les questions suivantes concernent votre activité en ligne, de façon générale.

2. En moyenne, combine d’heures par jour passez-vous en ligne? [dropdown]

□ 0-1 □ 2-3 □ 4-5 □6-8 □ 8+

3. Sur une échelle de 1 (pas du tout d’accord) à 7 (tout à fait d’accord), indiquez dans
quelle mesure vous êtes d’accord avec les énoncés suivants.

1 2 3 4 5 6 7
En général, j’aime être impliqué dans des discussions en
ligne
Je suis quelqu’un qui aime communiquer en ligne avec
d’autres personnes
Je suis quelqu’un qui aime participer activement à des
 261

discussions en ligne
En général, j’aime beaucoup échanger des idées en ligne

Page 5: Activité sur Facebook

Les questions suivantes concernent votre activité sur Facebook.

4. En quelle année avez-vous créé votre compte Facebook? [dropdown]

□ 2004 □ 2005 □ 2006 □ 2007 □ 2008 □ 2009 □ 2010 □ 2011 □ 2012 □ 2013 □ 2014

5. Par jour, combien de fois vous connectez-vous sur Facebook? [dropdown]

□ Je ne me connecte pas tous les jours

□ de 1 à 3
□ de 4 à 6
□ plus de 6
□ En continu: je reçois les notifications sur mon portable

6. En moyenne, combien de minutes passez-vous sur Facebook chaque jour? [dropdown]

□ Moins de 10 minutes □ de 31 à 60 min
□ de 10 à 30 min □ plus de 60 min

Page 6: Identification de la page

Le lien vers cette étude a été posté par une page dont vous êtes fan, qui a été
minutieusement sélectionnée pour cette étude. Il est important que vous répondiez à cette
étude en rapport avec cette page en particulier, et la marque qu’elle représente.

Veuillez préciser à quelle marque est dédiée la page qui a posté le lien vers cette étude:….

7. Etes-vous client de cette marque (si applicable)

□ Oui
□ Non, je suis fna sur Facebook mais je n’ai jamais acheté cette marque

Page 7: Questions sur la marque

Les questions suivantes sont à propos de la marque que vous venez de mentionner.

8. Sur une échelle de 1 (pas du tout d’accord) à 7 (tout à fait d’accord), indiquez votre
avis sur les énoncés suivants:

1 2 3 4 5 6 7
Je préfère les produits/services offerts par cette marque
par rapport à d’autres
J’aime les produits/services offerts par cette marque
Ce sont les produits/services de cette marque que j’ai le
plus de plaisir à utiliser
J’ai confiance en cette marque
Je me fie à cette marque
 262

Cette marque est honnête

Cette marque est une valeur sûre

Page 8: Engagement affectif avec la marque

9. Sur une échelle de 1 (pas du tout d’accord) à 7 (tout à fait d’accord), indiquez votre
avis sur les énoncés suivants:

1 2 3 4 5 6 7
Je suis enthousiaste par rapport à cette marque
Cette marque me rend enthousiaste
Je suis vraiment fan de cette marque
Tout ce qui se rapporte à cette marque m’intéresse
Je trouve cette marque intéressante
J’aime interagir avec cette marque
Je suis heureux/-se quand j’interagis avec cette marque
Je prends du plaisir à interagir avec cette marque
Interagir avec cette marque est une recompense pour
moi

Page 9: Engagement cognitive avec la marque

10. Sur une échelle de 1 (pas du tout d’accord) à 7 (tout à fait d’accord), indiquez votre
avis sur les énoncés suivants:

1 2 3 4 5 6 7
Je fais très attention à cette marque
Ce qui se rapporte à cette marque attire mon attention
Je pense beaucoup à cette marque
Je prends le temps de penser à cette marque
Quand j’interagis avec cette marque, j’oublie tout le
reste
Le temps passe vite quand j’interagis avec cette marque
Quand j’interagis avec cette marque, je suis transporté
J’ai du mal à revenir sur terre quand j’interagis avec
cette marque

Page 10: Engagement comportemental avec la marque

Les questions suivantes sont a propos de vos interactions avec les administrateurs de la
page Facebook, qui sont généralement les gestionnaires de la marque. Pensez aux
interactions ci-dessous en termes d'utilisation des boutons "j'aime", "commenter",
"partager", etc.
11. Sur une échelle de 1 (pas du tout d’accord) à 7 (tout à fait d’accord), indiquez votre
avis sur les énoncés suivants:
 263

1 2 3 4 5 6 7
Je réponds aux questions des gestionnaires de la marque
Je partage mon opinion avec eux
Je partage mon expérience avec eux
Je partage mes idées avec eux
Je partage du contenu intéressant avec eux
Je les aide
Je leur pose des questions
Je cherche des idées et informations auprès d’eux
Je recherche leur aide
J’apprends grâce au contenu qu’ils partagent
Je montre mon accord par rapport à ce qu’ils peuvent
dire ou faire
Je partage leur contenu avec mon réseau
Je promeus la marque
J’essaye d’intéresser d’autres personnes à la marque
Je défends activement la marque de ses critiques
Je dis des choses positives aux autres à propos de la
marque

Page 11: Product Category

Les questions suivantes ont rapport à la catégorie de produits à laquelle la marque

appartient.

12. A quelle catégorie de produits appartient cette marque?

□ Alimentation
□ Technologies (applications, telecoms, sites internet, ordinateurs…)
□ Services (banques, assurances, éducation…)
□ Tourisme/Voyages (agencies, companies aériennes…)
□ Mode et beauté
□ Biens durables (voitures, électroménagers)
□ Vente au detail (supermarchés, magasins, commerce en ligne…)
□ Divertissement (sports, films, séries, livres, jeux, …)
□ Autre: ………………………………….

13. Considérant la catégorie de produits que vous venez de sélectionner, et si

applicable, indiquez sur une échelle de 1 (pas du tout d’accord) à 7 (tout à fait d’accord)
votre avis par rapport aux phrases suivantes :

1 2 3 4 5 6 7
Ce type de produits est très important pour moi
J’accord de l’importance à ce type de produits
Quand on achète ce type de produits, il est grave de se
tromper
Quand on achète ce type de produits, il est difficile de se
 264

tromper
J’aime particulièrement ce type de produits
Le type de produits qu’une personne achète en dit
beaucoup sur elle

14. Si applicable, considérez vos achats dans cette catégorie de produits et indiquez
dans quelle mesure vous êtes d’accord avec les énoncés suivants (1= pas du tout
d’accord ; 7 = tout à fait d’accord)

1 2 3 4 5 6 7
Je suis fidèle à la marque dont je suis fan quand j’achète
ce type de produits
Pour mon prochain achat, j’achèterai à nouveau cette
marque
J’achète toujours cette marque
J’achète en général cette marque

Page 12: Community-related variables

Pensez maintenant à la page dont vous faites part pour répondre aux questions suivantes.
15. Depuis combien de temps êtes-vous membre de la page?

□ Moins d’un an □ 1 – 5 ans □ 5 – 10 ans

16. Avec quelle fréquence cliquez-vous activement sur la page?

□ Jamais □ Moins d’une fois par mois □ Environ une fois par mois □ Environ une
fois par semane □ Plus d’une fois par semaine
17. Combien de temps passez-vous sur la page par semaine?
□ 0 – 2 min □ 3 – 5 min □ 6 -10 min □ 11 – 15 min □ 15 min +
18. Comment qualifieriez-vous la taille de la page, en termes de nombre de fans?
□ Très petite □ Petite □ Moyenne □ Grande □ Très grande
19. Sur les échelles suvantes, veuillez decrire votre attitude quant au fait d’interagir sur la
page au cours du mois prochain:

1 2 3 4 5 6 7
1 = Idiot; 7 = Sage
1 = Nocif; 7 = Bénéfique
1 = Mauvais; 7 = Bon
1 = Pénalisant; 7 = Gratifiant

Page 13: Engagement affective avec la communauté

20. Sur une échelle de 1 (pas du tout d’accord) à 7 (tout à fait d’accord), indiquez votre
 265

avis sur les énoncés suivants.

1 2 3 4 5 6 7
Je suis enthousiaste par rapport à cette page
Cette page me rend enthousiaste
Je suis vraiment fan de cette page
Tout ce qui se rapporte à cette page m’intéresse
Je trouve cette page intéressante
J’aime interagir avec cette page
Je suis heureux/-se quand j’interagis avec les membres
de cette page
Je prends du plaisir à interagir au sein de cette page
Interagir avec cette page est une recompense pour moi

Page 14: Engagement cognitive avec la communauté

21. Sur une échelle de 1 (pas du tout d’accord) à 7 (tout à fait d’accord), indiquez votre
avis sur les énoncés suivants.

1 2 3 4 5 6 7
Je fais très attention à la page
Ce qui se rapporte à la page attire mon attention
Je pense beaucoup à la page
Je prends le temps de penser à la page
Quand j’interagis avec les membres de cette page,
j’oublie tout le reste
Le temps pass evite quand j’interagis avec les membres
de cette page
Quand j’interagis avec les membres de cette page, je
m’évade
J’ai du mal à retomber sur terre quand j’interagis avec
les membres de cette page

Page 15: Engagement comportemental avec la communauté

Les questions suivantes concernent vos interactions avec les autres membres de la page.
Pensez aux interactions ci-dessous en termes d'utilisation des boutons "j'aime",
"commenter", "partager", etc.

22. Sur une échelle de 1 (pas du tout d’accord) à 7 (tout à fait d’accord), indiquez votre
avis sur les énoncés suivants.

1 2 3 4 5 6 7
Je réponds à leurs questions
Je leur fais part de mon opinion
Je partage mon expérience avec eux
Je partage mes idées avec eux
 266

1 2 3 4 5 6 7
Je partage du contenu intéressant avec eux
Je les aide
Je leur pose des questions
Je cherche des idées et informations auprès d’eux
Je recherche leur aide
J’apprends grâce au contenu qu’ils partagent
Je montre mon accord par rapport à ce qu’ils peuvent
dire ou faire
Je partage leur contenu avec mon réseau
Je fais la promotion de la page
J’essaye d’intéresser d’autres personnes à la page
Je défends activement la page de ses critiques
Je communiqué des choses positives aux autres à propos
de la page

Page 16: Demographiques

23. Quel est votre sexe?

□ Homme □ Femme
24. En quelle année êtes-vous né(e)? [dropdown 1996 – 1930]
25. Quel est le plus haut niveau d’études que vous ayez atteint?
□ Ecole primaire □ Ecole Secondaire □ Supérieur de type court □ Supérieur de
type long
26. Quelle est votre nationalité? [dropdown avec tous les pays]
27. Dans quell pays résidez-vous? [dropdown avec tous les pays]

Page 17: Thank you

Vos réponses sont maintenant enregistrées. Nous vous remercions vivement pour votre
participation à cette étude.

Si vous souhaitez participer au tirage au sort pour gagner un des coupons Amazon de €100,
veuillez indiquer votre adresse email ci-dessous.
 267

Appendix 4: Sample of Facebook pages

Product Member count Answer Answer/

Facebook Pages
category Feb 2014 count Members
The Huggy's Bar 6,000 142 2.36%
Kate's Kitchen 4,536 38 0.83%
Agora Greek Delicacies 1,128 20 1.77%
Jupiler 267,900 15 0.01%
Food and MaBelle 773 11 1.42%
Beverage Edward & Irwyn 458 5 1.09%
The Belgian Owl 1,484 4 0.26%
Red Bull 38,456,000 1 0.01%
Nutella 218,000 1 0.01%
Nespresso 300,000 1 0.01%
Star Alliance 152,000 137 0.09%
Delta Airlines 130,000 4 0.01%
Lufthansa 150,000 3 0.01%
Travel US Airways 128,000 3 0.01%
United Airlines 685,000 1 0.01%
Swiss International Air
534,000 1
Lines 0.01%
ASOS 3100000 40 0.01%
Made&More 817 28 3.42%
Zara 18,567,000 11 0.01%
Smalltwongirl 1,456 8 0.54%
Fashion Too Belgista - Le Blog 281 4 1.42%
and Beauty
Scotts Guard Watches 567 3 0.52%
J&Joy 36,000 2 0.01%
Suit Supply 14,567,000 1 0.01%
Skin Clinics 786 1 0.01%
Fit Body Bootcamp 3,156 26 0.82%
Runner's World 1,134,000 20 0.01%
TEDx University of
4,123 11
Glasgow 0.26%
Entertainm Snooze Pure FM 52,000 11 0.02%
ent Borrowed Space 354 7 1.97%
Playstation 28,345,000 4 0.01%
ESN 657 3 0.45%
Citizen Mule 173 2 1.15%
Scottish Rugby 118,978 1 0.01%
Durable Porsche 630,000 54 0.01%
Goods Audi 7,690,545 12 0.01%
Services Santander UK 187,000 16 0.01%
 268

Product Member count Answer Answer/

Facebook Pages
category Feb 2014 count Members
Creative Wallonia 4,456 14 0.31%
Betacowork 1,879 8 0.42%
University of Glasgow 87,000 2 0.01%
Hot Dog Fashion 1,039 8 0.77%
Others L'Echo 8,979 6 0.06%
The Guardian 3,987,000 3 0.01%
Go Pro 7,678,000 8 0.01%
Technology Plug&Go 188 7 3.72%
Samsung Mobile 31,000,000 1 0.01%
Amazon 22,000,000 11 0.01%
Retail
TESCO 1,300,000 1 0.01%
Total 181,543,713 721 0.01%
 269

Appendix 5: Examples of Facebook pages

The above screen-shots depict a varied array of Facebook pages included in the sample.
From left to right, top to bottom: Kate’s Kitchen is a local fresh food deli in Sligo,
Ireland. Star Alliance is an international airline alliance. Runner’s World is a global, US-
based magazine targeting runners, and Made&More is a slow-fashion e-retailer based in
Belgium.
 270

Appendix 6: Data screening

Valid Missing
Item Mean St. dev. Skewness Kurtosis
count count
Online Interaction Propensity
OIP1 1296 393 3.58 1.69 0.21 -0.8
OIP2 1301 388 4.32 1.68 -0.18 -0.87
OIP3 1293 396 3.46 1.71 0.21 -0.96
OIP4 1295 394 3.80 1.75 -0.02 -0.94
Brand Trust
BTR1 971 718 5.42 1.54 -1.05 0.6
BTR2 973 716 4.85 1.77 -0.65 -0.52
BTR3 970 719 5.23 1.57 -0.79 0.04
BTR4 970 719 5.43 1.50 -1 0.6
Brand Commitment
BCO1 971 718 4.94 1.78 -0.76 -0.35
BCO2 968 721 5.26 1.60 -0.94 0.31
BCO3 969 720 4.94 1.73 -0.74 -0.3
Online Brand Enthusiasm
BENT1 926 763 5.20 1.60 -0.83 0.11
BENT2 925 764 4.97 1.68 -0.65 -0.34
BENT3 925 764 4.70 1.75 -0.45 -0.68
BENT4 924 765 4.57 1.80 -0.38 -0.74
BENT5 924 765 5.22 1.52 -0.8 0.2
Online Brand Enjoyment
BENJ1 924 765 4.13 1.85 -0.14 -0.98
BENJ2 923 766 3.90 1.88 -0.02 -1.04
BENJ3 922 767 3.92 1.89 -0.04 -1.07
BENJ4 921 768 3.40 1.91 0.24 -1.09
Online Brand Attention
BAT1 879 810 4.44 1.73 -0.34 -0.67
BAT2 877 812 4.90 1.63 -0.63 -0.22
BAT3 876 813 3.36 1.76 0.32 -0.81
BAT4 875 814 3.14 1.77 0.45 -0.73
Online Brand Absorption
BABS1 876 813 2.61 1.75 0.87 -0.25
BABS2 874 815 3.01 1.86 0.54 -0.84
BABS3 872 817 2.88 1.85 0.64 -0.69
BABS4 878 811 2.57 1.74 0.88 -0.35
Online Brand Sharing
BSH1 825 864 3.29 1.87 0.32 -1.02
BSH2 827 862 3.29 1.89 0.31 -1.05
BSH3 822 867 3.23 1.88 0.36 -1
BSH4 827 862 3.05 1.84 0.47 -0.92
 271

Valid Missing
Item Mean St. dev. Skewness Kurtosis
count count
BSH5 824 865 2.84 1.79 0.65 -0.69
BSH6 827 862 2.87 1.78 0.57 -0.82
Online Brand Learning
BLE1 825 864 3.15 1.88 0.35 -1.13
BLE2 826 863 3.49 1.97 0.11 -1.28
BLE3 825 864 2.99 1.89 0.51 -0.94
BLE4 825 864 4.21 1.93 -0.3 -1.01
Online Brand Endorsing
BEND1 823 866 3.77 1.93 -0.04 -1.17
BEND2 822 867 3.72 1.97 -0.01 -1.2
BEND3 826 863 4.05 2.02 -0.21 -1.18
BEND4 827 862 4.02 2.04 -0.18 -1.23
BEND5 825 864 3.53 2.03 0.18 -1.25
BEND6 825 864 4.73 1.95 -0.64 -0.66
Product Involvement
PI1 805 884 5.08 1.60 -0.59 -0.45
PI2 804 885 5.16 1.55 -0.67 -0.25
PI3 797 892 4.29 1.95 -0.16 -1.1
PI4 797 892 4.01 1.85 -0.02 -1
PI5 801 888 5.45 1.42 -0.91 0.48
PI6 798 891 4.67 1.77 -0.56 -0.58
Brand Loyalty
BL1 794 895 4.16 1.86 -0.24 -1.02
BL2 793 896 4.66 1.76 -0.53 -0.58
BL3 792 897 3.52 1.99 0.19 -1.18
BL4 785 904 4.35 1.95 -0.32 -1.02
Attitude Toward Online Participation
ATI1 747 942 4.50 1.49 -0.32 0.06
ATI2 743 946 4.81 1.28 -0.33 0.41
ATI3 743 946 4.97 1.33 -0.43 0.31
ATI4 742 947 4.77 1.26 -0.2 0.44
Online Brand Community Enthusiasm
OENT1 734 955 4.23 1.75 -0.27 -0.83
OENT2 734 955 3.81 1.82 0.02 -1.01
OENT3 729 960 3.48 1.84 0.22 -1.01
OENT4 732 957 3.84 1.85 0 -1.07
OENT5 732 957 4.60 1.67 -0.52 -0.45
Online Brand Community Enjoyment
OENJ1 734 955 2.86 1.71 0.61 -0.6
OENJ2 731 958 2.76 1.71 0.6 -0.72
OENJ3 731 958 2.96 1.75 0.47 -0.86
OENJ4 730 959 2.57 1.71 0.84 -0.35
Online Brand Community Attention
OAT1 718 971 3.65 1.76 0.07 -0.94
 272

Valid Missing
Item Mean St. dev. Skewness Kurtosis
count count
OAT2 717 972 4.31 1.71 -0.39 -0.7
OAT3 715 974 2.60 1.64 0.8 -0.29
OAT4 713 976 2.49 1.62 0.89 -0.09
Online Brand Community Absorption
OABS1 716 973 2.12 1.53 1.33 0.92
OABS2 714 975 2.25 1.55 1.12 0.34
OABS3 715 974 2.14 1.52 1.28 0.74
OABS4 709 980 2.01 1.47 1.45 1.27
Online Brand Community Sharing
OSH1 710 979 3.00 1.81 0.47 -0.92
OSH2 711 978 3.06 1.80 0.38 -1
OSH3 709 980 2.96 1.80 0.48 -0.92
OSH4 707 982 2.82 1.78 0.54 -0.9
OSH5 710 979 2.71 1.77 0.64 -0.75
OSH6 708 981 2.74 1.71 0.57 -0.8
Online Brand Community Learning
OLE1 712 977 2.92 1.84 0.54 -0.9
OLE2 711 978 3.24 1.92 0.3 -1.16
OLE3 708 981 2.77 1.80 0.66 -0.71
OLE4 709 980 3.79 1.95 -0.09 -1.18
Online Brand Community Endorsement
OEND1 711 978 3.30 1.89 0.23 -1.14
OEND2 708 981 3.30 1.97 0.28 -1.24
OEND3 709 980 3.47 2.06 0.15 -1.38
OEND4 711 978 3.38 2.05 0.25 -1.31
OEND5 709 980 3.08 1.98 0.48 -1.1
OEND6 705 984 3.77 2.15 0.01 -1.41
 273

Appendix 7: Sample characteristics

English French Total

Variables Count Percent Count Percent Count Percent

Age
18-24 80 18 90 33 170 24
25-34 193 43 138 51 331 46
35-44 104 23 33 12 137 19
45-54 52 12 8 3 60 8
55+ 19 4 4 1 23 3
Gender
Male 252 56 103 38 355 49
Female 196 44 170 62 366 51
Education
Primary school 2 0 1 0 3 0
Secondary school 56 13 41 15 97 13
Undergraduate degree 174 39 93 34 267 37
Postgraduate degree 216 48 138 51 354 49
Nationality
UK 76 17 0 0 76 11
GR 42 9 0 0 42 6
BE 35 8 238 87 273 38
FR 0 0 14 5 14 2
US 34 8 0 0 34 5
IE 33 7 0 0 33 5
Others 228 51 11 4 239 33
Country of residence
UK 124 28 0 0 124 17
US 40 9 0 0 40 6
IE 36 8 0 0 36 5
BE 35 8 245 90 280 39
FR 0 0 14 5 14 2
GR 23 5 0 0 23 3
Others 190 42 21 8 211 29
Brand category
Travel 148 33 1 0 149 21
Food and Beverage 87 19 151 55 238 33
Durable Goods 66 15 0 0 66 9
Entertainment 52 12 33 12 85 12
Fashion and Beauty 50 11 48 18 98 14
Services 23 5 17 6 40 6
Others 11 2 6 2 17 2
 274

English French Total

Variables Count Percent Count Percent Count Percent

Retail 6 1 6 2 12 2
Technology 5 1 11 4 16 2
Existing consumer
Yes 380 85 225 82 605 84
No 68 15 48 18 116 16
Facebook joining year
2004 13 3 5 2 18 2
2005 18 4 1 0 19 3
2006 45 10 25 9 70 10
2007 89 20 53 19 142 20
2008 103 23 132 48 235 33
2009 74 17 34 12 108 15
2010 50 11 12 4 62 9
2011 25 6 7 3 32 4
2012 19 4 3 1 22 3
2013 9 2 0 0 9 1
2014 3 1 1 0 4 1
Daily time on Facebook
Less than 10 minutes 37 8 14 5 51 7
10 to 30 min 129 29 55 20 184 26
31 to 60 min 143 32 81 29 224 31
60 min + 139 31 123 45 262 36
Daily Facebook log-ons
All the time
151 34 99 36 250 35
(notifications)
1 to 3 116 26 57 21 173 24
4 to 6 88 20 48 18 136 19
6+ 70 16 60 22 130 18
I don't log on every day 23 5 9 3 32 4
Page membership
duration
Less than a year 150 33 145 53 295 41
1-5 years 282 63 125 46 407 56
5-10 years 16 4 3 1 19 3
Active page visits
Never 43 10 34 12 77 11
Less than once a month 123 27 82 30 205 28
About once a month 99 22 83 30 182 25
About once a week 114 25 59 22 173 24
More than once a week 69 15 15 5 84 12
Time spend on page
monthly
0-2 min 188 42 167 61 355 49
 275

English French Total

Variables Count Percent Count Percent Count Percent

3-5 min 110 25 69 25 179 25
6-10 min 83 19 20 7 103 14
11-15 min 34 8 15 5 49 7
15 + min 33 7 2 1 35 5
Perceived page size
Very small 12 3 0 0 12 2
Fairly small 59 13 41 15 100 14
Medium 141 31 135 49 276 38
Fairly big 141 31 69 25 210 29
Very big 95 21 28 10 123 17
 276

Appendix 8: Profile of consumer interviewees

Country of Internet Social Media Daily time Length
Name Gender Age Nationality Language Medium
residence usage (ys) usage (ys) online (hs) (min)
Fred M 40 Belgian Belgium 14 10 10+ French Skype 60
Liam M 25 Chinese UK 15 8 10+ English F2F 40
James M 27 Scottish UK 13 7 7 English F2F 61
Sabrina F 27 Belgian Belgium 13 6 2 English Skype 53
Nigel M 28 Chinese/Canadian UK 15 7 7 English F2F 45
Sam M 29 Pakistani UK 13 6 4 English F2F 63
Sandra F 27 Belgian Belgium 13 9 3 French Skype 56
Judith F 28 Belgian Belgium 13 7 10 English Skype 94
Claire F 28 Scottish UK 14 7 10 English F2F 49
Anthony M 48 Belgian Belgium 15 8 10+ English Skype 60
Helen F 24 Greek UK 10 7 10+ English F2F 52
Maria F 25 Greek UK 10 6 10+ English F2F 52
Ray M 28 Belgian Belgium 13 6 3 French Skype 37
Derek M 33 Chinese/Canadian China 18 10 8 English Skype 140
Flora F 23 Peruvian Netherlands 10 7 3 English Skype 56
Matt M 25 Belgian Belgium 11 4 6 English Skype 83
Laura F 26 German UK 11 7 5 English F2F 47
Steven M 27 Belgian Belgium 10 6 3 French Skype 70
Sophia F 23 Pakistani UK 12 5 6 English F2F 36
Akim M 27 Pakistani UK 12 6 9 English F2F 35
 277

Appendix 9: Profile of industry expert interviewees

Marketing
Country of Length
Name Company experience Language Medium
residence (min)
(years)
Aubin Agentia Belgium 14 French Skype 60
Keith GreenSocial India 6 English Skype 35
Dan Freelance Belgium 13 English Skype 51
Benjamin SmartForest Belgium 9 French Skype 39
George IronValley Canada 6 English Skype 78
 278

Appendix 10: The interview guide

Developing a measure of consumer engagement

in Online Brand Communities

Introduction
- Turn on recording
- Thank participant for taking part
- Remind of the study context, hand out (or send) information sheet and consent
forms. Make sure forms are understood and have verbal or written consent.

Part 1: Your online activity in general

Let’s first discuss your online activity. Could you tell me a bit about what you usually do
online on an everyday basis (for personal, non-work related purposes)?
How about your activity on social media, could you tell me a little more about how you
use them6.

- Could you tell me which ones you use?

- Could you tell me for which purpose you usually use them?
- How about the benefits you get from using social media

Part 2: Your online activities related to brands

Do you participate, or are you a member of one or several online group(s) focused on a
specific brand? (I might be brands you particularly like or particularly dislike – give
examples)
Can you tell me more about the kind of groups are you part of? (Understand if they are
Facebook page/group member, Twitter follower, blog subscriber or casual follower, forum
member, website member...)
Can you tell me about your experience in these groups?
Could you tell me about your interactions with the other members of the group, if you have
any?
How about your interactions with the brand that the group is about?
Could you tell me about your level of involvement in these groups?

Part 3: Community engagement

Can you think about one group that you are a part of and that you particularly like? Which
brand is it about? Let’s talk about it a little bit...

6
Explained that social media are all types of platform or channels where users can generate content and
interact with others. Examples are social networking sites (Facebook, LinkedIn, Twitter, Google +), boards
(Pinterest0, blogs, forums, wikis, video and photo sharing (Youtube, Instagram, ...) etc.
 279

Could you explain to me what this group mean to you?

Can you tell how you feel about the other members of the group?
What can you say about your role in the group?
Can you recall and explain an instance in which you interacted with an (or several) other
member(s) of the group?
Can you think of an occasion when the group was valuable to you?
Can you recall a time when being part of this group helped you interact with the brand it is
focused on?

Part 4: Brand engagement

Looking at the brand that the group focuses on...
Could you explain what this brand means to you?
How would you describe the brand?
Can you explain to me how the brand makes you feel?
How do you feel about the brand in question? How do these feelings manifest themselves?
Could you explain what are your thoughts about the brand/how would you evaluate it?
Can you think of an occasion when you have done something to express your thoughts
about the brand?

Part 5: The concept of consumer engagement

In your own words, how would you define the term ‘engagement’?
In your opinion, what does it mean to be engaged as a consumer?
Can you give me an example of a time when you feel you were engaged as a consumer?

Part 6: Demographics

 Name
 Age
 Gender
 Years of Internet usage
 Years of social media usage
 Frequency of Internet usage (hours per day)
Thank participant and turn off recording.
 280

Appendix 11: Qualitative data summary

Theoretical
Keywords Sample quote 1 Sample quote 2 Sample quote 3
Themes
‘When you are really engaged, you
Emotion, love, ‘They feel probably as close as a family,
General ‘Being engagement, it's really just might loose the distance necessary to
feel, heart, the page is like a home for us, we are too
affect liking the brand.' Derek take sound decisions, your emotions
passion, like involved!' Sam
guide you.' Steven
‘If I like a brand post, or comment on a
‘Being engaged with (…) is to tell ‘I like this company because of Facebook!
brand status, my friends are likely to
Excitement, oneself that you are so interested in it, They have a nice face, they are very kind,
comment on it as well. And in these
Enthusiasm passion, high that you spend time and effort on it, and helpful and enthusiastic about their
cases, I am so excited that I am quite
interest and that to some extent, you even put products and I relate to that, you know.'
happy to keep the conversation going
yourself at risk for it'. Anthony. Maria.
and talk more.' Liam.
‘They always have something fun to tell
Pleasure,
‘It’s not necessarily important to have on their page, something that is really ‘I like that and I have participated very
pleasant,
comments on what you posted, but it’s ‘Nutella’, something that is really about often, I have not won yet but I don’t
Enjoyment pleasing,
a pleasure, it’s a nice added value.' gourmandise, fun...so I really like this care, I like it still, it’s funny! I really
aggreable,
Anthony. page: it represents me and it represents like participating in that page.' Maria
nice, fun
what I enjoy in life.' Sabrina.
It just depends on how much time you It’s about the involvement of your
General Think, mental, are willing to sacrifice for the It means that you spend time thinking mind with something. You are engaging
cognitive mind group…how much time you spend about it.' Laura your mind. Your focus is on something
thinking about it.' Flora. that you are focusing on.' Sophia
Attention, I pay attention to follow things that don’t
attract, spend make you look very stupid. So I would So I try not to go too often because it is
‘It is an engagement of the mind!’
Attention time, make follow, ask questions and interact with too much time consuming: from one
Sophia.
time, know, things that would keep me at a certain thing you go to another.' Judith
conscious professional level.' James
 281

Theoretical
Keywords Sample quote 1 Sample quote 2 Sample quote 3
Themes
‘So when I go on Facebook I
Focus, can't essentially read the newsfeed, but I There is so much going on, and
Twitter, well, I think it is the biggest
stop, really read every single thing, even if it sometime I interact with so many brands
Absorption waste of time ever but I like it (laughs).'
immersed, takes me hours. Facebook I am more that afterwards I don’t even remember
Flora
stuck able to turn it off when I really want which brand it was.' Claire
to. Pinterest it’s impossible!’ Judith.
Motivation,
General effort, action, ‘Proactively and physically involved ‘Going out of a passive situation and Constantly interacting with them.'
behaviour active, in an activity’, Nigel entering an active situation.' Sandra Sophia
interaction
‘[I use the Facebook group]…to I just want to share, about anything,
Sharing, ‘Well yes, for instance, if somebody asks
exchange experiences about visits. If clothes, a service, a restaurant, that I
letting know, a question about a football game (i.e. ‘Did
Sharing there is a place where we have been, like. I like to share, but I don’t expect
telling, you see what just happened?’) I would
we can inform other people who are that others will consume based on my
exhanging very quickly answer.’ James
interested in visiting!’ Maria review.' Liam
Finding out, ‘I got a pen burst out in one of my ‘I follow them just to make sure I know If you see that some comment got a lot
information, favourite bags and I tweeted about it which products they are launching, what of likes, it is as if the group has
Learning knowing, and asked if anybody had any ‘at- is in their new summer collection. I want authenticated the words for you. It gives
news, solving, home’ remedies for what to do, and I to know what is new at the moment.' some sort of seal of approval, or quality
question, learn got loads back.’ Claire Sophia seal.' James
It’s also like being a lawyer for the ‘I’m liking things a lot, I’m the kind of ‘I took part in the vote (launched by a
Like,
company, like a spokesperson. In their person that sees something and then, hop, design brand) and then promoted it on
comment,
social circles online, the person is I like it.’ Judith Facebook. It’s not only because I want
Endorsing approve,
going to let their friends, people to buy their product, but because
support,
acquaintances and family know about they are really nice and really good.’
promote
your company.' Laura
 282

Appendix 12: Email to academic experts

Dear [title, name]

My name is Laurence Dessart and I am a doctoral researcher at the University of Glasgow,

Scotland, working under the supervision of Dr. Cleopatra Veloutsou and Dr. Anna
Morgan-Thomas.

I am writing to you in your quality of expert in the field of [specific field], as I hope you
could help me in developing an appropriate scale for the core construct of my thesis,
consumer engagement.

My PhD focuses on consumer engagement in the context of online brand communities and,
as no appropriate scale could be found in the existing literature; I am developing one.

The following link will provide you the definitions of consumer engagement and each of
its dimensions and sub-dimensions. For each sub-dimension, a list of items is proposed,
and the questionnaire will allow you to assess the representativeness of each of them. It
should take a maximum of 25 minutes to complete.

Link to the survey: https://www.surveymonkey.com/s/engagement_experts

I hope you can find the time to fill in the questionnaire, and I thank you in advance for
sharing your insight.

Best regards,

Laurence Dessart
 283

References

Aaker, J. L. 1997. Dimensions of brand personality. Journal of Marketing Research, 34(3),

pp. 347–356.

Achterberg, W., Pot A., Kerkstra, A., Ooms, M., Muller, M. and Ribbe, M. 2003. The
effect of depression on social engagement in newly admitted Dutch nursing home
residents. The Gerontologist, 43(2), pp. 213–218.

Adjei, M. T., Noble, S. M., and Noble, C. H. 2010. The influence of C2C communications
in online brand communities on customer purchase behavior. Journal of the Academy of
Marketing Science, 38(5), pp. 634–653.

Agarwal, J. and Malhotra, N. K. 2005. An integrated model of attitude and affect:

theoretical foundation and an empirical investigation. Journal of Business Research, 58(4),
pp. 483–493.

Ahluwalia, R., Burnkrant, R. E., and Unnava, H. R. A. O. 2013. Consumer response to

negative publicity: the moderating role of commitment. Journal of Marketing Research,
37(2), pp. 203–214.

Ajzen, I. 1991. The theory of planned behavior. Organizational Behavior and Human
Decision Processes, 50(2), pp. 179–211.

Algesheimer, R., Dholakia, U. M. and Herrmann, A. 2005. The social influence of brand
community: evidence from European car clubs. Journal of Marketing, 69(3), pp. 19-34.

Allen, N. J., and Meyer, J. P. 1990. The measurement and antecedents of affective,
continuance and normative commitment to the organization. Journal of Occupational
Psychology, 63(1), pp. 1–18.

American Marketing Association. 2012. Marketing Dictionary [Online]. [Accessed 12

September 2012]. Available from:
http://www.marketingpower.com/_layouts/Dictionary.aspx

Andersen, P. H. 2005. Relationship marketing and brand involvement of professionals

through web-enhanced brand communities: the case of Coloplast. Industrial Marketing
Management, 34(1), pp. 39–51.
 284

Anderson, J. C., and Gerbing, D. W. 1982. Some methods for respecifying measurement
models to obtain unidimensional construct measurement. Journal of Marketing Research,
19(4), pp. 453–460.

Anderson, J. C., and Gerbing, D. W. 1988. Structural equation modeling in practice: a

review and recommended two-step approach. Psychological Bulletin, 103(3), pp. 411–423.

Andersson-Cederholm, E., and Gyimóthy, S. 2010. The service triad: modelling dialectic
tensions in service encounters. The Service Industries Journal, 30(2), pp. 265–280.

Armstrong, J. S. and Overton, T. S. 1977. Estimating nonresponse bias in mail

surveys. Journal of Marketing Research, 14(3), pp. 396–402.

Arnould, E. and Thompson, C., eds. 2007. Consumer Culture Theory (And We Really
Mean Theoretics) (Vol. 11): Emerald Group Publishing Limited.

Ashley, C. and Tuten, T. 2015. Creative Strategies in Social Media Marketing: An

Exploratory Study of Branded Social Content and Consumer Engagement. Psychology and
Marketing, 32(1), pp. 15–27.

Bagozzi, R. P. 1994. Structural equation models in marketing research: basic principles. In

Bagozzi, R. P. ed., Principles of marketing research, Cambridge, Massachusetts:
Blackwell. pp. 386–422.

Bagozzi, R. P. and Dholakia, U. M. 2006. Antecedents and purchase consequences of

customer participation in small group brand communities. International Journal of
Research in Marketing, 23(1), pp. 45–61.

Bagozzi, R. P., Gopinath, M. and Nyer, P. U. 1999. The role of emotions in

marketing. Journal of the Academy of Marketing Science, 27(2), pp. 184–206.

Bagozzi, R. P. and Phillips, L. W. 1982. Representing and testing organizational theories:

A holistic construal. Administrative Science Quarterly, 27(3), pp. 459–489.

Bagozzi, R. P., Yi, Y. and Phillips, L. W. 1991. Construct validity in organizational

research, Administrative Science Quarterly, 36(3), pp. 421–458.

Bagozzi, R. P. and Yi, Y. 1988. On the evaluation of structural equation models. Journal
of the Academy of Marketing Science, 16(1), pp. 74–94.
 285

Baines, P., Fill, C. and Page, K. 2001. Marketing. 2nd ed. New York: Oxford University
Press.

Baird, C. H. and Parasnis, G. 2011. From social media to social customer relationship
management, Strategy and Leadership, 39(5), pp. 30–37.

Baker, M. J. 2001. Selecting a research methodology. The Marketing Review, 1(3), pp.
373–397.

Baker, M. J. 2003. Data Collection: Questionnaire Design. The Marketing Review, 3(3),
343–370.

Bakker, A. B., Albrecht, S. L., and Leiter, M. P. 2011. Key questions regarding work
engagement. European Journal of Work and Organizational Psychology, 20(1), pp. 4–28.

Beatty, S. E., Homer, P. and Kahle, L. R. 1988. The involvement—commitment model:

Theory and implications. Journal of Business research, 16(2), pp. 149–167.

Bejerholm, U. and Eklund, M. 2007. Occupational engagement in persons with

schizophrenia: relationships to self-related variables, psychopathology, and quality of life.
The American Journal of Occupational Therapy, 61(1), pp. 21–32.

Belk, R. W. 1988. Possessions and the extended self. Journal of Consumer Research,
15(2), pp. 139–168.

Bentler, P. M. 1992. On the fit of models to covariances and methodology to the Bulletin.
Psychological Bulletin, 112(3), pp. 400–404.

Bergami, M. and Bagozzi, R. P. 2000. Self‐categorization, affective commitment and

group self‐esteem as distinct aspects of social identity in the organization. British Journal
of Social Psychology, 39(4), pp. 555–577.

Bhattacharya, C. B. and Sen, S. 2003. Consumer-company identification: a framework for

understanding consumers’ relationships with companies. Journal of Marketing, 67(2), pp.
76–88.

Blazevic, V., Wiertz, C., Cotte, J. de Ruyter, K. and Keeling, D. I. 2014. GOSIP in
cyberspace: conceptualization and scale development for general online social interaction
propensity. Journal of Interactive Marketing, 28(2), pp. 87–100.
 286

Bloch, P. H. 1981. An exploration into the scaling of consumers’ involvement with a

product class. Advances in Consumer Research, 8, pp. 61–65.

Bloch, P. H. 1986. Product enthusiasm: many questions, a few answers. Advances in

Consumer Research, 13(1), pp. 534–544.

Bogdan, R., and Taylor, S. 1975. Introduction to qualitative research methods. A

phenomenological approach to the social sciences. New-York: Wiley.

Bowden, J. L. H. 2009. The process of customer engagement: a conceptual

framework. Journal of Marketing Theory and Practice, 17(1), pp. 63–74.

Bowden, J. L., Gabbott, M. and Naumann, K. 2014. Service relationships and the
customer disengagement – engagement conundrum. Journal of Marketing Management,
31(7/8), pp. 774–806.

Brakus, J. J., Schmitt, B. H., and Zarantonello, L. 2009. Brand experience: what is it? How
is it measured? Does it affect loyalty? Journal of Marketing, 73(3), pp. 52–68.

Breidbach, F., Brodie, R. and Hollebeek, L. 2014. Beyond virtuality: from engagement
platforms to engagement ecosystems. Managing Service Quality, 24(6), pp. 592–611.

Breitsohl, J., Kunz, W. H. and Dowell, D. 2015. Does the host match the content? A
taxonomical update on online consumption communities. Journal of Marketing
Management, DOI: 10.1080/0267257X.2015.1036102.

Breugelmans, E. 2008. Marketing Research: International Marketing Management.

McGraw-Hill/Irwin.

Brodie, R. J., Hollebeek, L. D., Juric, B., and Ilic, A. 2011. Customer engagement:
conceptual domain, fundamental propositions, and implications for research. Journal of
Service Research, 14(3), pp. 252–271.

Brodie, R. J., Ilic, A., Juric, B. and Hollebeek, L. 2013. Consumer engagement in a virtual
brand community: an exploratory analysis. Journal of Business Research, 66(1), pp. 105–
114.

Bryman, A. 2008. Social research methods. 3rd ed. Oxford: Oxford University Press.

Bruhn, J. G. 2005. The sociology of community connections. New-York: Springer.

 287

Bryson, C. and Hand, L. 2007. The role of engagement in inspiring teaching and learning.
Innovations in Education and Teaching International, 44(4), pp. 349–362.

Buckingham, A. and Saunders, P. 2004. The survey methods workbook. Cambridge: Polity
Press.

Business Insider. 2014. Revealed: the demographic trends for every social network.
[Online]. [Accessed 15 January 2015]. Available at: http://uk.businessinsider.com/2014-
social-media-demographics-update-2014-9

Byrne, B. 2010. Structural equation modeling with AMOS: basic concept, applications and
programming, 2nd ed. New-York: Routledge, Taylor and Francis Group.

Cadogan, J. W. 2010 Comparative, cross-cultural, and cross-national research: A comment

on good and bad practice. International Marketing Review, 27(6), pp. 601–605.

Cadogan, J. W., Diamantopoulos, A. and Pahud de Mortanges, C. 1999. A measure of

export market orientation: Scale development and cross-cultural validation. Journal of
International Business Studies, 30(4), pp. 689–707.

Calder, B. J., Isaac, M. S. and Malthouse, E. C. 2013. Taking the customer’s point-of-
view: engagement or satisfaction? Marketing Science Institute Working Paper Series 2013.
pp. 13–102.

Calder, B. J., Malthouse, E. C. and Schaedel, U. 2009. An experimental study of the

relationship between online engagement and advertising effectiveness. Journal of
Interactive Marketing, 23(4), pp. 321–331.

Calhoun, C. 1983. The radicalness of tradition: community strength or venerable disguise

and borrowed language. American Journal of Social Sociology, 88(5), pp. 886–914.

Campbell, C., Ferraro, C. and Sands, S. 2014. Segmenting consumer reactions to social
network marketing. European Journal of Marketing, 48(3/4), pp. 432–452.

Campbell, D. T., & Fiske, D. W. 1959. Convergent and discriminant validation by the
multitrait-multimethod matrix. Psychological Bulletin, 56(2), pp. 81-105.

Carlson, B. D., Suter, T. A.and Brown, T. J. 2008. Social versus psychological brand
community: the role of psychological sense of brand community. Journal of Business
Research, 61(4), pp. 284–291.
 288

Carrington, M. J., Nelville, B. A. and Whitwell, G. J. 2014. Lost in translation: Exploring

the ethical consumer intention–behaviour gap. Journal of Business Research, 67(1), pp.
2759–2767.

Casaló, L., Flavián, C., and Guinalíu, M. 2007. The impact of participation in virtual brand
communities on consumer trust and loyalty: The case of free software. Online Information
Review, 31(6), pp. 775–792.

Casaló, L. V., Flavián, C. and Guinaliu, M. 2008. Promoting consumer's participation in

virtual brand communities: a new paradigm in branding strategy. Journal of Marketing
Communications, 14(1), pp. 19–36.

Casaló, L. V., Flavián, C. and Guinalíu, M. 2010. Determinants of the intention to

participate in firm-hosted online travel communities and effects on consumer behavioral
intentions. Tourism Management, 31(6), pp. 898–911.

Castells, M. 1996. The rise of the network society. Oxford: Blackwell.

Castells, M. 1997. The information age, Vol. 2: the power of identity. Oxford: Blackwell.

Catteeuw, F., Flynn E. and Vonderhorst, J. 2007. Employee engagement: boosting

productivity in turbulent times. Organization Development Journal, 25(2), pp. 151–157.

Chang, H. H. and Chuang, S-S. 2011. Social capital and individual motivations on
knowledge sharing:Participant involvement as a moderator. Information and Management,
48(1), pp. 9–18.

Chang, A. Hsieh, S. H. and Lin, F. 2013. Personality traits that lead members of online
brand communities to participate in information sending and receiving. International
Journal of Electronic Commerce, 17(3), pp. 37–61.

Chaudhuri, A. 2000. A macro analysis of the relationship of product involvement and

information search: the role of risk. Journal of Marketing Theory and Practice, 8(1), pp.
1–15.

Chaudhuri, A. and Holbrook, M. B. 2001. The chain of effects from brand trust and brand
affect to brand performance: the role of brand loyalty. Journal of Marketing, 65(2), pp. 81–
93.
 289

Chaudhuri, A. and Holbrook, M. B. 2002. Product-class effects on brand commitment and

brand outcomes: The role of brand trust and brand affect. The Journal of Brand
Management, 10(1), pp. 33–58.

Chauhan, K. and Pillai, A. 2013. Role of content strategy in social media brand
communities: a case of higher education institutes in India. Journal of Product and Brand
Management, 22(1), pp. 40–51.

Christodoulides, G. 2009. Branding in the post-internet era. Marketing Theory, 9(1), pp.
141–144.

Christodoulides, G., De Chernatony, L., Furrer, O., Shiu, E., and Abimbola, T. 2006.
Conceptualising and measuring the equity of online brands. Journal of Marketing
Management, 22(7/8), pp. 799–825.

Christodoulides, G., Jevons, C. and Blackshaw, P. 2011. The voice of the consumer speaks
forcefully in brand identity: user-generated content forces smart marketers to listen,
Journal of Advertising Research, 51(1), pp.101–111.

Churchill, G. A. 1979. A Paradigm for Developing Better Measures of Marketing

Constructs. Journal of Marketing Research, 16(1), pp. 64–73.

Churchill, G. A. 1999. Marketing research: methodological foundations, 7th ed. Orlando,

FL: Harcourt Brace and Company.

Churchill, G. A., and Peter, J. P. 1984. Research design effects on the reliability of rating
scales: a meta-analysis. Journal of Marketing Research, 21(4), pp. 360–375.

Clark, L. A., and Watson, D. 1995. Constructing validity: basic issues in objective scale
development. Psychological Assessment, 7(3), pp. 309.

Colliander, J., Dahlén, M. and Modig, E. 2015. Twitter for two: investigating the effects of
dialogue with customers in social media. International Journal of Advertising, DOI:
http://dx.doi.org/10.1080/02650487.2014.996197.

Cook, C., Heath, F. and Thompson, R. L. 2000. A meta-analysis of response rates in Web-
or Internet-based surveys. Educational and Psychological Measurement, 60(6), pp. 821–
836.
 290

Cooke, M. and Buckley, N. 2008. Web 2.0, social networks and the future of market
research. International Journal of Market Research, 50(2), pp. 267–292.

Couper, M. P. 2000. Web surveys: a review of issues and approaches. Public Opinion
Quarterly, 64(4), pp. 464–494.

Couper, M. P., Traugott, M. W., and Lamias, M. J. 2001. Web survey design and
administration. Public Opinion Quarterly, 65(2), pp. 230–253.

Cova, B. 1997. Community and consumption: towards definition of the linking value of
products or services. European Journal of Marketing, 31(3/4), pp. 297–316.

Cova, B. and Cova, V. 2002. Tribal marketing: The tribalisation of society and its impact
on the conduct of marketing. European Journal of Marketing, 36(5/6), pp. 595–620.

Cova, B. and Pace, S. 2006. Brand community of convenience products: new forms of
customer empowerment–the case ‘my Nutella the community’. European Journal of
Marketing, 40(9/10), pp. 1087–1105.

Cova, B., Pace, S. and Park, D. J. 2007. Global brand communities across borders: the
Warhammer case. International Marketing Review, 24(3), pp. 313–329.

Cova, B. and White, T. 2010. Counter-brand and alter-brand communities: the impact of
Web 2.0 on tribal marketing approaches. Journal of Marketing Management, 26(3/4), pp.
256–270.

Cox, E. P. 1980. The optimal number of response alternatives for a scale: a review.
Journal of Marketing Research, 17(4), pp. 407–422.

Craig, C. S., and Douglas, S. P. 2006. Beyond national culture: implications of cultural
dynamics for consumer research. International Marketing Review, 23(3), pp. 322–342.

Creswell, J. W. 2003. Research design: qualitative, quantitative, and mixed methods

approaches, 2nd ed. Thousand Oaks, CA: Sage.

Creswell, J. W. 2007. Qualitative inquiry and research design: choosing among five
approaches, 2nd ed. Thousand Oaks, CA: Sage.

Creswell, J. W. and Clark, V. L. 2008. Designing and conducting mixed methods research
2nd ed. Thousand Oaks, CA: Sage.
 291

Czaja, R. and Blair, J. 2005. Designing surveys: a guide to decisions and procedures, 2nd
ed. Thousand Oaks, CA: Pine Forge Press.

Das, M., Ester, P. and Kaczmirek, L. 2011. Social and behavioral research and the
Internet: advances in applied methods and research strategies. New-York: Routledge,
Taylor and Francis Group.

Davis, F. D. 1989. Perceived usefulness, perceived ease of use, and user acceptance of
information technology. MIS Quarterly. 13, pp. 319–339.

Dawes, J. 2008. Do data characteristics change according to the number of scale points
used? International Journal of Market Research, 50(1), pp. 61–77.

de Chernatony, L. 1999. Brand management through narrowing the gap between brand
identity and brand reputation. Journal of Marketing Management, 15(1-3), pp. 157–179.

de Chernatony, L. and Dall'Olmo Riley, F. 1998. Defining a ‘brand’: beyond the literature
with experts' interpretations. Journal of Marketing Management, 14(5), pp. 417–443.

De Mooij, M. 2011. Consumer behavior and culture. consequences for global marketing
and advertising, 2nd ed. Thousand Oaks, CA: Sage.

De Valck, K., Van Bruggen, G. H. and Wierenga, B. 2009. Virtual communities: A

marketing perspective. Decision Support Systems, 47(3), pp. 185–203.

de Villiers, R. 2015. Consumer brand enmeshment: Typography and complexity modeling

of consumer brand engagement and brand loyalty enactments. Journal of Business
Research. DOI:10.1016/j.jbusres.2015.01.005.

De Vries, L., Gensler, S. and Leeflang, P. S. 2012. Popularity of brand posts on brand fan
pages: an investigation of the effects of social media marketing. Journal of Interactive
Marketing, 26(2), pp. 83–91.

De Wulf, K., Oderkerken-Schröder, G., and Iacobucci, D. 2013. Investments and

relationships : exploration. Journal of Marketing, 65(4), pp. 33–50.

Delanty, G. 2003. Community (Key ideas). London: Routledge.

Del Rio, A. B., Vazquez, R., and Iglesias, V. 2001. The effects of brand associations on
consumer response. Journal of Consumer Marketing, 18(5), pp. 410–425.
 292

Denzin, N. K., and Lincoln, Y. S. 2011. The sage handbook of qualitative research, 4th ed..
Thousand Oaks, CA: Sage.

Dessart, L., Veloutsou, C., and Morgan-Thomas, A. 2015. Consumer engagement in online
brand communities: A social media perspective. Journal of Product and Brand
Management, 24(1) DOI: 10.1108/JPBM-06-2014-0635.

DeVellis, R. F. 2012. Scaled development: theory and applications, 3rd ed. Thousand Oaks,
CA: Sage.

Dholakia, U. M., Bagozzi, R. P., and Pearo, L. K. 2004. A social influence model of
consumer participation in network-and small-group-based virtual
communities. International Journal of Research in Marketing, 21(3), pp. 241–263.

Dholakia, N. and Reyes, I. 2013. Virtuality as place and process. Journal of Marketing
Management, 29 (13/14), pp.1580–1591.

Dholakia, U. M. and Vianello, S. 2009. Effective brand community management: lessons

from customer enthusiasts. Available at SSRN 1512090.

Diamantopoulos, A., Reynolds, N., and Schlegelmilch, B. 1994. Pretesting in questionnaire

design: the impact of questionnaire characteristics on error detection. Journal of Market
Research Society, 36(4), pp. 295–313.

Diamantopoulos, A., and Siguaw, J. A. 2000. Introducing Lisrel: a guide for the
uninitiated. London: Sage.

Dillman, D. A. 2000. Mail and Internet surveys: the total design method. New-York:
Wiley.

Dillman, D. A., and Bowker, D. K. 2001. The Web questionnaire challenge to survey
methodologists in Batinic, B., Reips, U.-D. Bosnjak, M. and Werner A. eds., Online
Social Sciences, pp. 53–71.

Doney, P.M., Cannon, J.P. and Mullen, M.R. 1998. Understanding the influence of
national culture on the development of trust. Academy of Management Review, 23(3), pp.
601–620.

Douglas, S. P., and Craig, C. S. 2006. On improving the conceptual foundations of

international marketing research. Journal of International Marketing, 14(1), pp. 1–22
 293

Eagly, A. H. and Chaiken, S. 1993. The psychology of attitudes. Harcourt Brace

Jovanovich College Publishers.

Egan, J. 2007. Marketing Communications. London: Thompson Press.

Eisingerich, A. B. and Rubera, G. 2010. Drivers of brand commitment : a cross-national

investigation. Journal of International Marketing, 18(2), pp. 64–79.

El-Manstrly, D. 2014. Cross-cultural validation of switching costs: a four-country

assessment. International Marketing Review, 31(4), pp. 413–437.

El-Manstrly, D., and Harrison, T. 2013. A critical examination of service loyalty measures.
Journal of Marketing Management, 29(15/16), pp. 1–28.

Ewing, M. T., Wagstaff, P. E. and Powell, I. H. 2013. Brand rivalry and community
conflict. Journal of Business Research, 66(1), pp. 4–12.

Facebook. 2012. Announcing a New Pages Structure for Global Brands. [Online].
[Accessed 12 April 2014]. Available at https://www.facebook-
studio.com/news/item/announcing-a-new-pages-structure-for-global-brands

Facebook. 2014. Localization and translation. [Online]. [Accessed 02 November 2014].

Available at https://developers.facebook.com/docs/internationalization

Facebook. 2015a. Newsroom: Company Information. [Online]. [Accessed 16 January

2015]. Available at http://newsroom.fb.com/company-info/

Facebook. 2015b. Desktop help: Pages basics. [Online]. [Accessed 16 January 2015].
Available at https://www.facebook.com/help/155275634539412

Faraj, S., and Johnson, S. L. 2011. Network exchange patterns in online

communities. Organization Science, 22(6), pp. 1464–1480.

Field, A. 2009. Discovering statistics using SPSS, 3rd ed. London: Sage.

Fetscherin, M., and Heinrich, D. 2015. Consumer brand relationships research: a

bibliometric citation meta-analysis. Journal of Business Research, 68(2), pp. 380–390.

Fiedler, M., and Sarstedt, M. 2014. Influence of community design on user behaviors in
online communities. Journal of Business Research, 67(11), pp. 2258–2268.
 294

Fielstead, W. J. 1979. Qualitative methods: a needed perspective in evaluation research, in

Cook, T. D. and Reichardt, C. S., eds. Qualitative and Quantitative Methods in Evaluation
Research, Beverly Hills, CA: Sage, pp. 33–48.

Fishbein, M. and Ajzen, I. 1975. Belief, attitude, intention, and behavior: an introduction
to theory and research. Reading, MA: Addison-Wesley

Flynn, L. R., and Goldsmith, R. E. 1993. Application of the personal involvement

inventory in marketing. Psychology and Marketing, 10(4), pp. 357–366.

Flynn, L. R. and Pearcy, D. 2001. Four subtle sins in scale development: some suggestions
for strengthening the current paradigm. International Journal of Market Research, 43(4),
pp. 409–433

Fornell, C. and Larker, D. 1981. Evaluating structural equation models with unobservable
variables and measurement error. Journal of Marketing, 18(1), pp. 39–50.

Forrester Research. 2014. Predictions 2015: Social Media Grows Up - A Stable Social
Landscape Will Foster Marketing, Customer Insights, And Sales Success. [Online].
[Accessed 8 February 2015]. Available at
https://www.forrester.com/Predictions+2015+Social+Media+Grows+Up/fulltext/-/E-
RES119621?docid=119621andal=0

Fournier, S. 1998. Consumers and their brands: developing relationship theory in

consumer research. Journal of Consumer Research, 24(4), pp. 343–353.

Fournier, S. and Avery, J. 2011. The uninvited brand. Business Horizons, 54(3), pp. 191–
207.

Fournier, S. and Lee, L. 2009. Getting brand communities right. Harvard Business Review,
87(4), pp. 105–111.

Fowler III, A. and Krush, M. 2008. We happy few: redefining community in

marketing. Advances in Consumer Research, 35, pp. 851–852.

Franzak, F., Makarem, S. and Jae, H. 2014. Design benefits, emotional responses, and
brand engagement. Journal of Product and Brand Management, 23(1), pp. 16–23.
 295

Freling, T. H., Crosno, J. L. and Henard, D. H. 2010. Brand personality appeal:

conceptualization and empirical validation. Journal of the Academy of Marketing Science,
39(3), pp. 392–406.

Füller, J., Matzler, K. and Hoppe, M. 2008. Brand community members as a source of
innovation. Journal of Product Innovation Management, 25(6), pp. 608–619.

Gambetti, R. C., and Graffigna, G. 2010. The concept of engagement: a systematic

analysis of the ongoing marketing debate. International Journal of Market Research, (52),
pp. 801–826.

Gambetti, R. C., Graffigna, G. and Biraghi, S. 2012. The Grounded Theory approach to
consumer-brand engagement. International Journal of Market Research, 54(5), pp. 659–
687.

Ganley, D. and Lampe, C. 2009. The ties that bind: Social network principles in online
communities. Decision Support Systems, 47(3), pp. 266–274.

Garbarino, E. and Johnson, M. S. 1999. The Different Roles of Satisfaction , Trust , and
Commitment in Customer Relationships. Journal of Marketing, 63(2), pp. 70–87.

Garnier, M. and Macdonald, E. K. 2009. The savvy French consumer: a cross-cultural

replication. Journal of Marketing Management, 25(9-10), pp. 965–986.

Garnefeld, I., Iseke, A. and Krebs, A. 2012. Explicit incentives in online communities:
boon or bane? International Journal of Electronic Commerce, 17(1), pp. 11–38.

Gensler, S., Völckner, F., Liu-Thompkins, Y. and Wiertz, C. 2013. Managing brands in the
social media environment. Journal of Interactive Marketing, 27(4), pp. 242–256.

Gerbing, D. W., and Anderson, J. C. 1988. An updated paradigm for scale development
incorporating unidimensionality and its assessment. Journal of Marketing Research, 25(2),
pp. 186–192.

Gerbing, D. W. and Hamilton, J.G. 1996. Viability of exploratory analysis as a precursor to

confirmatory factor analysis. Structural Equation Modeling, 3(1), pp. 62–72.

Gironda, J. T. and Korgaonkar, P. K. 2014. Understanding consumers’ social networking

site usage. Journal of Marketing Management, 30(5-6), pp. 571–605.
 296

Goulding, C., Shankar, A. and Canniford, R. 2013. Learning to be tribal: facilitating the
formation of consumer tribes. European Journal of Marketing, 47(5/6), pp. 813–832.

Grudens-Schuck, N. 2000. Conflict and engagement: an empirical study of a farmer-

extension partnership in a sustainable agriculture program. Journal of Agricultural and
Environmental Ethics, 13(1), pp. 79–100.

Gruen, T. W., Summers, J. O. and Acito, F. 2000. Relationship marketing activities,

commitment, and membership behaviors in professional associations. Journal of
Marketing, 64(3), pp. 34–49.

Gruner, R. L., Homburg, C., and Lukas, B. A. 2014. Firm-hosted online brand
communities and new product success. Journal of the Academy of Marketing
Science, 42(1), pp. 29–48.

Guba, E. G. and Lincoln, Y. S. 2005. Paradigmatic controversies, contradictions, and

emerging influences . In Denzin N. K. and Lincoln, Y. S. eds. The Sage Handbook of
Qualitative Reserarch, 3rd ed. Thousand Oaks, CA: Sage.

Gummerus, J., Liljander, V., Weman, E. and Pihlström, M. 2012. Customer engagement in
a Facebook brand community. Management Research Review, 35(9), pp. 857–877.

Habibi, M. R., Laroche, M., and Richard, M. O. 2014. Brand communities based in social
media: How unique are they? Evidence from two exemplary brand
communities. International Journal of Information Management, 34(2), pp. 123–132.

Hagel, J. and Armstrong, A. 1997. Net gain: Expanding markets through virtual
communities. Boston: Harvard Business Press.

Hair, J., Bush, R. and Ortinau, D. 2006. Marketing research within a changing
environment, 3rd ed. New York: McGraw-Hill/Irwin.

Hair, J., Anderson, R. E., Tatham, R. L. and Black, W. C. 1998. Multivariate data
analysis, 5th ed. Upper Saddle River, New Jersey: Prentice Hall.

Hall, E. 1976. Beyond culture. New-York: Doubleday.

Hammond, M. 2000. Communication within on-line forums: the opportunities, the

constraints and the value of a communicative approach. Computers and Education, 35(4),
pp. 251–262.
 297

Havitz, M. and Howard, D. 1995. How enduring is enduring involvement? A seasonal

examination of three recreational activities. Journal of Consumer Psychology, 4(3), pp.
255–276.

Haynes, S. N., Richard, D. C. S. and Kubany, E. S. 1995. Content validity in psychological

assessment: a functional approach to concepts and methods. Psychological Assessment,
7(3), pp. 238–247.

Healy, J. C. and McDonagh, P. 2013. Consumer roles in brand culture and value co-
creation in virtual communities. Journal of Business Research, 66(9), pp. 1528–1540.

Healy, M. and Perry, C. 2000. Comprehensive criteria to judge validity and reliability of
qualitative research within the realism paradigm. Qualitative Market Research: An
International Journal, 3(3), pp. 118–126.

Heere, B., Walker, M., Yoshida, M., Ko, Y. J., Jordan, J. S. and James, J. D. 2011. Brand
community development through associated communities: grounding community
measurement within social identity theory. The Journal of Marketing Theory and
Practice, 19(4), pp. 407–422.

Hennig-Thurau, T., Malthouse, E. C., Friege, C., Gensler, S., Lobschat, L., Rangaswamy,
A. and Skiera, B. 2010. The impact of new media on customer relationships. Journal of
Service Research, 13(3), pp. 311–330.

Hess, J. 1995. Construction and assessment of a scale to measure consumer trust. American
Marketing Association, Chicago, IL, Summer, (6), pp. 20–26.

Hewson, C., Yule, P., Laurent, D. and Vogel, C. 2003. Internet Research Methods: a
practical guide for the social and behavioural sciences. London: Sage.

Higgins, E. T., and Scholer, A. A. 2009. Engaging the consumer: the science and art of the
value creation process. Journal of Consumer Psychology, 19(2), pp. 100–114.

Hinkin, T. R. 1995. A review of scale development practices in the study of organizations.

Journal of Management, 21(5), pp. 967–988.

Hoffman, D. L. and Novak, T. P. 1996. Marketing in hypermedia computer-mediated

environments: conceptual foundations. Journal of Marketing, 60(3), pp. 50–68.
 298

Hofstede, G., Hofstede, G.J. and Minkov, M. 2010. Cultures and organisation: software of
the mind. New-York: McGraw Hill.

Hofstede, G. 2014. Cultural tools: country comparison [Online]. [Accessed 29 December

2014]. Available from: http://geert-hofstede.com/countries.html

Homburg, C., Ehm, L., and Artz, M. 2015. Measuring and managing consumer sentiment
in an online community environment. Journal of Marketing Research. DOI:
10.1509/jmr.11.0448.

Hollebeek, L. D. 2011a. Demystifying customer brand engagement: Exploring the loyalty

nexus. Journal of Marketing Management, 27(7/8), pp. 785–807

Hollebeek, L. 2011b. Exploring customer brand engagement: definition and

themes. Journal of Strategic Marketing, 19(7), pp. 555–573.

Hollebeek, L., and Chen, T. 2014. Exploring positively-versus negatively-valenced brand

engagement: a conceptual model. Journal of Product and Brand Management, 23(1), pp.
62–74.

Hollebeek, L. D., Glynn, M. S. and Brodie, R. J. 2014. Consumer brand engagement in

social media: conceptualization, scale development and validation. Journal of Interactive
Marketing, 28(2), pp. 149–165.

Hollenbeck, C. R. and Zinkhan, G. M. 2010. Anti‐brand communities, negotiation of brand

meaning, and the learning process: the case of Wal‐Mart. Consumption, Markets and
Culture, 13(3), pp. 325–345.

Hollenbeck, C. R. and Kaikati, A. M. 2012. Consumers' use of brands to reflect their actual
and ideal selves on Facebook. International Journal of Research in Marketing, 29(4),
pp.395–405.

House, R.J., Hanges, P.J., Javidan, M., Dorfman, P.W. and Gupta, V. eds.. 2004. Culture,
Leadership, and Organizations: The GLOBE Study of 62 Societies. Thousand Oaks: Sage.

Hsieh, H. F. and Shannon, S. E. 2005. Three approaches to qualitative content analysis.

Qualitative Health Research, 15(9), pp. 1277-1288.
 299

Hsu, M. H., Yen, C. H., Chiu, C. M. and Chang, C. M. 2006. A longitudinal investigation
of continued online shopping behavior: An extension of the theory of planned
behavior. International Journal of Human-Computer Studies, 64(9), pp. 889–904.

Hu, L. T. and Bentler, P. M. 1999. Cut-off criteria for fit indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Structural Equation Modeling,
6(1), pp. 1–55.

Hung, K., Li, S. Y. and Tse, D. K. 2011. Interpersonal trust and platform credibility in a
Chinese multibrand online community. Journal of Advertising, 40(3), pp. 99–112.

Hunt, S. D. 1991. Modern marketing theory: critical issues in the philosophy of marketing
science. Cincinnati, OH: South-Western Publishing Co.

Hurley, A. E., Scandura, T. A., Schriesheim, C. A., Brannick. M. T., Seers, A.,
Vandenberg, R. J. and Williams, L.J. 1997. Exploratory and confirmatory factor analysis:
guidelines, issues and alternatives. Journal of Organizational Behavior, 18(6), pp. 667–
683.

Jaakkola, E. and Alexander, M. 2014. The Role of Customer Engagement Behavior in

Value Co-Creation A Service System Perspective. Journal of Service Research, doi:
1094670514529187.

Jacoby, J. and Chestnut, R. 1978. Brand loyalty: measurement and management. New
York: John Wiley and Sons.

Jackson, L. A. and Wang, J. L. 2013. Cultural differences in social networking site use: A
comparative study of China and the United States. Computers in Human Behavior, 29(3),
pp. 910–921.

Jahn, B. and Kunz, W. 2012. How to transform consumers into fans of your brand. Journal
of Service Management, 23(3), pp. 344–361.

Jang, H., Olfman, L., Ko, I., Koh, J. and Kim, K. 2008. The influence of on-line brand
community characteristics on community commitment and brand loyalty. International
Journal of Electronic Commerce, 12(3), pp. 57–80

Jennings, M. K. and Stoker, L. 2004. social trust and civic engagement across time and
generations. Acta Politica, 39(4), pp. 342–379.
 300

Jevons, C., Gabbott, M. and de Chernatony, L. 2005. Customer and brand manager
perspectives on brand relationships: a conceptual framework. Journal of Product and
Brand Management, 14(5), pp. 300–309.

Johar, J. S. and Sirgy, M. J. 1991. Value-Expressive Versus Utilitarian Advertising

Appeals : When And Why To Use Which Appeal. Journal of Advertising, 20(3), pp. 23–
33.

Johnson, P. and Duberley, J. 2000. Understanding management research: An introduction

to epistemology. London: Sage.

Johnson, M. D. and Fornell, C. 1991. A framework for comparing customer satisfaction

across individuals and product categories. Journal of Economic Psychology, 12(2), pp.
267–286.

Kahn, W. A. 1990. Psychological conditions of personal engagement and disengagement at

work. Academy of Management Journal, 33(4), pp. 692–724.

Kaltcheva, D., Patino, V., Laric, A. V., Pitta, D. and Imparato, N. 2014. Customers'
relational models as determinants of customer engagement value. Journal of Product and
Brand Management, 23(1), pp. 55–61.

Kaplan, A. M. and Haenlein, M. 2010. Users of the world, unite! The challenges and
opportunities of Social Media. Business horizons, 53(1), pp. 59–68.

Kates, S. M. 2003. The collective consumer-brand relationship. Advances in consumer

research, 30, pp. 159–159

Katz, J. E., and Rice, R. E. 2002. Social consequences of Internet use: Access,
involvement, and interaction. Cambridge, MA: MIT press.

Keller, K. L. 2013. Strategic brand management: building, measuring, and managing

brand, 4th Global ed. Harlow: Pearson Education.

Kennedy, R., Riquier, C. and Sharp, B. 1996. Practical Applications of Correspondence

Analysis to Categorical Data in Market Research. Journal of Targeting, Measurement and
Analysis for Marketing, 5, pp. 56–70.
 301

Klassen, R. D. and Jacobs, J. 2001. Experimental Comparison of Web, Electronic and Mail
Survey Technologies in Operations Management. Journal of Operations Management,
19(6), pp. 713–728.

Kim, J. W., Choi, J., Qualls, W. and Han, K. 2008. It takes a marketplace community to
raise brand commitment: the role of online communities. Journal of Marketing
Management, 24(3/4), pp. 409–431.

Kim, A. J. and Ko, E. 2012. Do social media marketing activities enhance customer
equity? An empirical study of luxury fashion brand. Journal of Business Research, 65(10),
pp. 1480–1486.

Koh, J., Kim, Y. G., Butler, B., and Bock, G. W. 2007. Encouraging participation in virtual
communities. Communications of the ACM, 50(2), pp. 68–73.

Koh, J., and Kim, Y. G. 2004. Knowledge sharing in virtual communities: an e-business
perspective. Expert Systems with Applications, 26(2), pp. 155–166.

Kozinets, R. V. 1999. E-tribalized marketing? The strategic implications of virtual

communities of consumption. European Management Journal, 17(3), pp. 252–264.

Kozinets, R. V. 2002. The field behind the screen: using netnography for marketing
research in online communities. Journal of Marketing Research, 39(1), pp. 61–72.

Kozinets, R., Wojnicki, A. C., Wilner, S. J. and De Valck, K. 2010. Networked narratives:
Understanding word-of-mouth marketing in online communities. Journal of Marketing,
74(2). pp. 71–89.

Kuenzel, S. and Halliday, S. V. 2010. The chain of effects from reputation and brand
personality congruence to brand loyalty: The role of brand identification. Journal of
Targeting, Measurement and Analysis for Marketing, 18(3/4), pp. 167–176.

Kuhn, T. S. 1962. The structure of scientific revolutions. Chicago: University

Kumar, V., Aksoy, L., Donkers, B., Venkatesan, R., Wiesel, T. and Tillmanns, S. 2010.
Undervalued or overvalued customers: capturing total customer engagement
value. Journal of Service Research, 13(3), pp. 297–310.
 302

Kumar, V., Pozza, I. D. and Ganesh, J. 2013. Revisiting the satisfaction–loyalty

relationship: empirical generalizations and directions for future research. Journal of
Retailing, 89(3), pp. 246–262.

Kuo, Y. F. and Feng, L. H. 2013. Relationships among community interaction

characteristics, perceived benefits, community commitment, and oppositional brand loyalty
in online brand communities. International Journal of Information Management, 33(6), pp.
948–962.

Kvale, S. 1996. Interviews: an introduction to qualitative research interviewing. Thousand

Oaks, CA: Sage.

Labrecque, L. I., vor dem Esche, J., Mathwick, C., Novak, T. P., and Hofacker, C. F. 2013.
Consumer power: Evolution in the digital age. Journal of Interactive Marketing, 27(4), pp.
257–269.

Laroche, M., Habibi, M. R., Richard, M. O., and Sankaranarayanan, R. 2012. The effects
of social media based brand communities on brand community markers, value creation
practices, brand trust and brand loyalty. Computers in Human Behavior, 28(5), pp. 1755–
1767.

Lastovicka, J. L. and Joachimsthaler, E. A. 1988. Improving the detection of personality-

behavior relationships in consumer research. Journal of Consumer Research, 14(4), pp.
583–587.

Lastovicka, J. L. and Gardner, D. 1979. Components of Involvement. In Maloney, J. C.

and Silverman, B. eds., Attitude research plays for high stakes. Chicago: American
Marketing Association, pp. 53–73.

Lasswell, H. D. 1965. World politics and personal insecurity. New York: Free Press.

Laurent, G. and Kapferer, J.-N. 1985. Consumer involvement profiles. Journal of

Marketing Research, 22(1), pp. 41–53.

Lee, M. S., Motion, J. and Conroy, D. 2009. Anti-consumption and brand

avoidance. Journal of Business Research, 62(2), pp. 169–180.

Libai, B., Bolton, R., Bügel, M. S., De Ruyter, K., Götz, O., Risselada, H. and Stephen, A.
T. 2010. Customer-to-customer interactions: broadening the scope of word of mouth
research. Journal of Service Research, 13(3), pp. 267–282.
 303

Lin, H. F. 2007. The role of online and offline features in sustaining virtual communities:
an empirical study. Internet Research, 17(2), pp. 119–138.

Lin, C. P. 2010. Learning virtual community loyalty behavior from a perspective of social
cognitive theory. International Journal of Human-Computer Interaction, 26(4), pp. 345–
360.

Little, R. J., and Rubin, D. B. 1989. The analysis of social science data with missing
values. Sociological Methods and Research, 18(2/3), pp. 292–326.

Loewy, E. H. 1993. Freedom and Community: The ethics of interdependence. Albany:

State University of New York Press.

London, B., Downey, G. and Mace, S. 2007. Psychological theories of educational

engagement: a multi-method approach to studying individual engagement and institutional
change. Vanderbilt Law Review, 60(2), pp. 455–481.

Louro, M. J. and Cunha, P. V. 2001. Brand management paradigms. Journal of Marketing

Management, 17(7/8), pp. 849–875.

Lytle, R. S., Hom, P. W. and Mokwa, M. P. 1998. SERV * OR : A Managerial Measure of
Organizational Service-Orientation. Journal of Retailing, 74(4), pp. 455–489.

Macey, W. H. and Schneider, B. 2008. The meaning of employee engagement. Industrial

and Organizational Psychology, 1(1), pp. 3–30.

Madupu, V. and Cooley, D. O. 2010. Antecedents and consequences of online brand

community participation: a conceptual framework. Journal of Internet Commerce, 9(2),
pp. 127–147.

Madupu, V. and Krishnan, B. 2008. The relationship between online brand community
participation and consciousness of kind, moral responsibility, and shared rituals and
traditions. Advances in Consumer Research, 35, pp. 853–853.

Maffesoli, M. 1996. The time of the tribes. London: Sage.

Mathwick, C. 2002. Understanding the on-line consumer: a typology of on-line relational

norms and behavior. Journal of Interactive Marketing, 16(1), pp. 40–55.

Mathwick, C., Wiertz, C. and De Ruyter, K. 2008. Social capital production in virtual P3
community. Journal of Consumer Research, 34, pp. 832–849.
 304

McAlexander, J. H., Schouten, J. W. and Koenig, H., F. 2002. Building brand community
Journal of Marketing, 66(1), pp. 38–54.

Malhotra, N. and Birks, D. 2006. Marketing research: an applied approach. Financial

Times/ Prentice Hall.

Malhotra, A., Malhotra, C. K. and See, A. 2012. How to create brand engagement on
Facebook. MIT Sloan Management Review, 54(2), pp. 18–20.

Malthouse, E. 2001. Factor analysis: checking assumptions of normality before conducting

factor analyses. Journal of Consumer Psychology, 10(1/2), pp. 79–81.

Matzler, K., Grabner-Kräuter, S. and Bidmon, S. 2008. Risk aversion and brand loyalty:
the mediating role of brand trust and brand affect. Journal of Product and Brand
Management, 17(3), pp. 154–162.

Matzler, K., Pichler, E., Füller, J., and Mooradian, T. A. 2011. Personality, person–brand
fit, and brand community: An investigation of individuals, brands, and brand
communities. Journal of Marketing Management, 27(9/10), pp. 874–890.

Marzocchi, G., Morandin, G. and Bergami, M. 2013. Brand communities: loyal to the
community or the brand? European Journal of Marketing, 47(1/2), pp. 93–114.

McCracken, G. D. 1988. The long interview. Beverly Hills, CA: Sage.

McQuail, D. 1983. Mass Communication Theory, 1st ed., London: Sage.

McQuarrie, E. F. and Munson, J. M. 1992. A revised product involvement inventory:

Improved usability and validity. Advances in Consumer Research, 19(1), pp. 108–115.

McWilliam, G. 2000. Building stronger brands through online communities. Sloan

Management Review, 41(3), pp. 43–54.

Merriam, S. B. (2009). Qualitative Research: A Guide to Design and Implementation. San

Francisco: Wiley.

Miles, M. B. and Huberman, A. M. 1994. Qualitative data analysis: An expanded

sourcebook. New-York: Sage.

Mittal, B. and Lee, M.-S. 1989. A causal model of consumer involvement. Journal of
Economic Psychology, 10, pp. 363–389.
 305

Mollen, A. and Wilson, H. 2010. Engagement, telepresence and interactivity in online

consumer experience: reconciling scholastic and managerial perspectives. Journal of
Business Research, 63(9), pp. 919-925.

Moorman, C., Zaltman, G. and Deshpande, R. 1992. Relationships between providers and
users of market research: The dynamics of trust. Journal of Marketing Research, 29(3), pp.
314–328.

Morgan, R. M. and Hunt, S. D. 1994. The commitment-trust theory of relationship

marketing. Journal of Marketing, 58(5), pp. 20–38.

Morgan-Thomas, A. and Veloutsou, C. 2013. Beyond technology acceptance: Brand

relationships and online brand experience. Journal of Business Research, 66(1), pp. 21–27.

MSI. 2010. 2010-2012 Research Priorities. Boston, MA: Marketing Science Institute.

Muñiz, A. M. and O'Guinn, T. C. 2001. Brand community. Journal of Consumer Research,

27(4), pp. 412–432.

Muñiz, A. M. and Schau, H. J. 2005. Religiosity in the abandoned Apple Newton brand
community. Journal of Consumer Research, 31(4), pp. 737–747.

Murphy, J. Hill, C. A. and Dean, E. 2014. Social Media, Sociality, and Survey Research,
First Edition. in Craig A. Hill, Elizabeth Dean, and Joe Murphy (eds). Social Media,
Sociality, and Survey Research. John Wiley and Sons, Inc. pp. 1–33.

Nambisan, S., and Baron, R. A. 2007. Interactions in virtual customer environments:

Implications for product support and customer relationship management. Journal of
Interactive Marketing, 21(2), pp. 42–62.

Netemeyer, R. G., Bearden, William, O. and Sharma, S. 2003. Scaling procedures: issues
and applications. Thousand Oaks, CA: Sage.

Nonnecke, B., Andrews, D. and Preece, J. 2006. Non-public and public online community
participation: needs, attitudes and behavior. Electronic Commerce Research, 6(1), pp. 7-
20.

Nunnally, J. C. and Bernstein, I. H. 1967. Psychometric Theory 3rd ed. New-York:

McGraw-Hill.
 306

O'Reilly, T. 2007. What is Web 2.0: Design patterns and business models for the next
generation of software. Communications and Strategies, 1, p. 17.

O'Sullivan, S. R., Richardson, B. and Collins, A. 2011. How brand communities emerge:
the Beamish conversion experience. Journal of Marketing Management, 27(9/10), pp.
891–912.

Odin, Y., Odin, N. and Valette-Florence, P. 2001. Conceptual and operational aspects of
brand loyalty: an empirical investigation. Journal of Business Research, 53(2), pp. 75–84.

Okazaki, S. and Taylor, C. R. 2013.Social media and international advertising: theoretical

challenges and future directions. International Marketing Review, 30(1), pp. 56–71.

Oppenheim, A. N. 2000. Questionnaire design, interviewing and attitude measurement.

London: Continuum.

Ouwersloot, H. and Odekerken-Schröder, G. 2008. Who's who in brand communities-and

why? European Journal of Marketing, 42(5-6), pp. 571–585.

Parasuraman, A., Zeithaml, V. A. and Berry, L. L. 1988. Servqual. Journal of retailing,

64(1), pp. 12–37.

Park, C., and Jun, J. K. 2003. A cross-cultural comparison of Internet buying behavior:
Effects of Internet usage, perceived risks, and innovativeness. International Marketing
Review, 20(5), pp. 534–553.

Patterson, A. 2012. Social-networkers of the world, unite and take over: A meta-
introspective perspective on the Facebook brand. Journal of Business Research, 65(4), pp.
527–534.

Patterson, P., Yu, T. and De Ruyter, K. 2006. Understanding customer engagement in

services. In Advancing theory, maintaining relevance, proceedings of ANZMAC 2006
conference, Brisbane, pp. 4–6.

Pavlou, P. A. and Fygenson, M. 2006. Understanding and predicting electronic commerce

adoption: An extension of the theory of planned behavior. MIS Quarterly, 30(1), pp. 115–
143.

Peter, J. P. 1979. Reliability: A review of psychometric basics and recent marketing

practices. Journal of Marketing Research, 16(1), pp. 6–17.
 307

Pew Research Center. 2015. Demographics of Key Social Networking Platforms [Online].
[Accessed 6 January 2015]. Available at:
http://www.pewinternet.org/2015/01/09/demographics-of-key-social-networking-
platforms-2/.

Pham, M. T. and Avnet T. 2009. Rethinking regulatory engagement theory. Journal of

Consumer Psychology, 19(2), pp. 115–123.

Phillips, B. J. and McQuarrie, E. F. 2010. Narrative and persuasion in fashion

advertising. Journal of Consumer Research, 37(3), pp. 368–392.

Podsakoff, P. M., and Organ, D. W. 1986. Self-reports in organizational research:

Problems and prospects. Journal of Management, 12(4), pp. 531-544.

Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., and Podsakoff, N. P. 2003. Common
method biases in behavioral research: a critical review of the literature and recommended
remedies. Journal of Applied Psychology, 88(5), pp. 879-903.

Porter, C. E. 2004. A typology of virtual communities: a multi‐disciplinary foundation for

future research. Journal of Computer‐Mediated Communication, 10(1), pp. 00-00.

Porter, C. E., Donthu, N., MacElroy, W. H. and Wydra, D. 2011. How to foster and sustain
engagement in virtual communities. California Management Review, 53(4), pp. 80–110.

Preece, J., Nonnecke, B. and Andrews, D. 2004. The top five reasons for lurking:
improving community experiences for everyone. Computers in Human Behavior, 20(2),
pp. 201–223.

Preece, J. and Maloney‐Krichmar, D. 2005. Online communities: design, theory, and

practice. Journal of Computer‐Mediated Communication, 10(4), 00.

Pritchard, M. P., Havitz, M. E. and Howard, D. R. 1999. Analyzing the commitment-

loyalty link in service contexts. Journal of the Academy of Marketing Science, 27(3), pp.
333–348.

Quinton, S. 2013. The community brand paradigm: A response to brand management's

dilemma in the digital era. Journal of Marketing Management, 29(7-8), pp. 912–932.
 308

Raacke, J. and Bonds-Raacke, J. 2008. MySpace and Facebook: Applying the uses and
gratifications theory to exploring friend-networking sites. Cyberpsychology and
behavior, 11(2), pp. 169–174.

Rafiq, M., Fulford, H. and Lu, X. 2013. Building customer loyalty in online retailing: the
role of relationship quality. Journal of Marketing Management, 29(3/4), pp. 494–517.

Resnick, E. 2001. Defining engagement. Journal of International Affairs, 54(2), pp. 551–
566.

Reto, F. 2012. Brand communities for mainstream brands: the example of the Yamaha R1
brand community Journal of Consumer Marketing, 29(3), pp. 225–232.

Rheingold, H. 1993. The virtual community: homesteading on the electronic frontier.

Reading, MA: Addison-Wesley.

Richins, L., and Bloch, P. H. 1991. Post-Purchase Incorporating and Time Product
Satisfaction : the Effects of Involvement and Time. Journal of Business Research, 23(2),
pp. 145–158.

Richins, M. L., Bloch, P. H., and McQuarrie, E. F. 1992. How Enduring and Situational
Involvement Combine to Create Involvement Responses. Journal of Consumer
Psychology, 1(2), pp. 143–153.

Ritchie, J., Lewis, J., McNaughton Nicholls, C. and Ormston, R. 2013. Qualitative
Research Practice: A Guide for Social Science Students and Researchers, 2nd ed.
Thousand Oaks, CA: Sage.

Robson, C. 2011. Real World Research, 3rd ed. New York: Wiley.

Rothaermel, F. T., and Sugiyama, S. 2001. Virtual internet communities and commercial
success: individual and community-level theory grounded in the atypical case of
TimeZone.com. Journal of Management, 27(3), pp. 297–312.

Rothbard, N. P. 2001. Enriching or depleting? The dynamics of engagement in work and

family roles. Administrative Science Quarterly, 46(4), pp. 655-684.

Ruiz Mafé, C. and Blas, S. S. 2006. Explaining Internet dependency: an exploratory study
of future purchase intention of Spanish Internet users. Internet Research, 16(4), pp. 380–
397.
 309

Ruvio, A., Shoham, A. and Brencic, M. M. 2008. Consumers' need for uniqueness: short-
form scale development and cross-cultural validation. International Marketing
Review, 25(1), pp. 33–53.

Saks, A. M. 2006. Antecedents and consequences of employee engagement. Journal of

Managerial Psychology, 21(7), pp. 600–619.

Salanova, M., Agut, S., and Peiro, J. M. 2005. Linking organizational resources and work
engagement to employee performance and customer loyalty: the mediation of service
climate. Journal of Applied Psychology, 90(6), pp. 1217–1227.

Salganik, M. J. and Heckathorn, D. D. 2004. Sampling and estimation in hidden

populations using respondent-driven sampling. Sociological Methodology, 34(1), pp. 193–
239.

Sanchez-Franco, M. J., and Rondan-Cataluña, F. J. 2010. Virtual travel communities and

customer loyalty: Customer purchase involvement and web site design. Electronic
Commerce Research and Applications, 9(2), pp. 171–182.

Sarason, S. B. 1974. The psychological sense of community: perspectives for community

psychology. San Francisco: Jossey-Bass.

Sarkar, A. and Sreejesh, S. 2014. Examination of the roles played by brand love and
jealousy in shaping customer engagement. Journal of Product and Brand
Management, 23(1), pp. 24–32.

Saunders, M., Lewis, P., and Thornhill, A. 2009. Research Methods for Business Students,
5th ed. Harlow: Pearson Education.

Sawhney, M., Verona, G. and Prandelli, E. 2005. Collaborating to create: the Internet as a
platform for customer engagement in product innovation. Journal of Interactive
Marketing, 19(4), pp. 4–17.

Schau, H. J., Muñiz Jr, A. M. and Arnould, E. J. 2009. How brand community practices
create value. Journal of Marketing, 73(5), pp. 30–51.

Schaufeli, W. B. and Bakker, A. B. 2004. Job demands, job resources, and their
relationship with burnout and engagement: A multi‐sample study. Journal of
Organizational Behavior, 25(3), pp. 293–315.
 310

Schaufeli, W. B., Bakker, A. B. and Salanova, M. 2006. The measurement of work

engagement with a short questionnaire a cross-national study. Educational and
psychological Measurement, 66(4), pp. 701–716.

Schaufeli, W. B., Salanova, M., González-Romá, V. and Bakker, A. B. 2002. The

measurement of engagement and burnout: A two sample confirmatory factor analytic
approach. Journal of Happiness studies, 3(1), pp. 71–92.

Schivinski, B. Christodoulides, G. and Dabrowski, D. Forthcoming. Developing and

validating a scale to measure consumers’ engagement with social media brand-related
content. Journal of Advertising Research.

Schouten, J. W. and McAlexander, J. H. 1995. Subcultures of consumption: an

ethnography of the new bikers Journal of Consumer Research, 22(1), pp. 43–61.

Schwandt, T. A. 2007. The Sage dictionary of qualitative inquiry. Thousand Oaks, CA:
Sage.

Scott, J. 1991. Networks of corporate power: A comparative assessment. Annual Review of

Sociology, 17, pp. 181–203.

Scott, J. and Craig-Lees, M. 2010. Audience engagement and its effects on product
placement recognition. Journal of Promotion Management, 16(1/2), pp. 39–58.

Seiders, K., Voss, G. B., Godfrey, A. L. and Grewal, D. 2007. SERVCON: development
and validation of a multidimensional service convenience scale. Journal of the Academy of
Marketing Science, 35(1), pp. 144–156.

Sen, S. and Bhattacharya, C. B. 2001. Does doing good always lead to doing better?
Consumer reactions to corporate social responsibility. Journal of Marketing Research,
38(5), pp. 225–243.

Shang, R. A., Chen, Y. C. and Liao, H. J. 2006. The value of participation in virtual
consumer communities on brand loyalty. Internet Research, 16(4), pp. 398–418.

Sharma, S. 1996. Applied multivariate techniques. New York: Wiley.

Sharma, P. and Chan, R. Y. K. 2011. Counterfeit proneness: conceptualisation and scale

development. Journal of Marketing Management, 27(5/6), pp. 602–626.
 311

Shen, Y. C., Huang, C. Y., Chu, C. H. and Liao, H. C. 2010. Virtual community loyalty: an
interpersonal-interaction perspective. International Journal of Electronic
Commerce, 15(1), pp. 49–74.

Shin, D. H. 2010. Analysis of online social networks: A cross-national study. Online

Information Review, 34(3), pp. 473–495.

Sicilia, M. and Palazón, M. 2008. Brand communities on the internet: A case study of
Coca-Cola's Spanish virtual community. Corporate Communications: An International
Journal, 13(3), pp. 255–270.

Slater, S., and Yani-de-Soriano, M. 2010. Researching consumers in multicultural

societies: Emerging methodological issues. Journal of Marketing Management, 26(11/12),
pp. 1143–1160.

Smith, A. N., Fischer, E. and Yongjian, C. 2012. How does brand-related user generated
content differ across YouTube, Facebook, and Twitter? Journal of Interactive
Marketing, 26(2), pp. 102–113.

So, K. K. F., King, C., Sparks, B. A. and Wang, Y. 2013. The influence of customer brand
identification on hotel brand evaluation and loyalty development. International journal of
hospitality management, 34(September), pp. 31–41.

Social Bakers. 2012. Proof: Local Facebook pages still work better than the global ones
[Online] [Accessed 4 February 2015]. Available at:
http://www.socialbakers.com/blog/366-proof-local-facebook-pages-still-work-better-then-
the-global-ones

Social Bakers. 2014. Top 15 Facebook Brand Pages by Engagement Rate [Socialbakers
Pro] [Online] [Accessed 4 February 2015]. Available at:
http://www.socialbakers.com/blog/370-top-15-facebook-brand-pages-by-engagement-rate-
socialbakers-pro

Sojka, J. Z. and Giese, J. L. 1997. Thinking and/or feeling: an examination of interaction

between processing styles. Advances in Consumer Research, 24, pp. 438–442.

Sprott, D., Czellar, S. and Spangenberg, E. 2009. The importance of a general measure of
brand engagement on market behavior: development and validation of a scale. Journal of
Marketing Research, 46(1), pp. 92–104.
 312

Sproull, L. 2003. Online communities. The Internet Encyclopedia. Wiley Online Library.

Stanoevska-Slabeva, K. and Schmid, B. F. 2001. A typology of online communities and

community supporting platforms. In System Sciences, 2001. Proceedings of the 34th
Annual Hawaii International Conference on (pp. 10). IEEE.

Statista. 2014. Global social networks ranked by number of users 2015. [Online]
[Accessed 4 February 2015]. Available at:
http://www.statista.com/statistics/272014/global-social-networks-ranked-by-number-of-
users/

Statista. 2015a. Distribution of active Facebook users worldwide as of 4th Quarter 2014,
by age. [Accessed 27 April 2015]. Available at:
http://www.statista.com/statistics/376128/facebook-global-user-age-distribution/.

Statista. 2015b. Most popular Facebook fan pages as of January 5, 2015. [Online]
[Accessed 10 February 2015]. Available at:
http://www.statista.com/statistics/269304/international-brands-on-facebook-by-number-of-
fans/

Steenkamp, J. B. and Baumgartner, H. 1998. Assessing Measurement Invariance in Cross-

National Research, Journal of Consumer Research , 25, (June), pp. 78-90.

Stokbürger-Sauer, N. 2010. Brand community: drivers and outcomes. Psychology and

Marketing, 27(4), pp. 347–368.

Szmigin, I., Canning, L., and Reppel, A. E. 2005. Online community: enhancing the
relationship marketing concept through customer bonding. International Journal of Service
Industry Management, 16(5), pp. 480–496.

Tabachnik, B. G. and Fidell, L. S. 2000. Using Multivariate Statistics, 4th ed, Boston, MA :
Allyn and Bacon, Pearson.

Tajfel, H., Turner, J. C. 2004. The social identity theory of intergroup behavior. In Jost, J.
T. and Sidanius, J. eds, (2004). Political psychology: Key readings. Key readings in social
psychology. New York: Psychology Press, xiii, 497 pp. 276–293.

Thomas, T. C., Price, L. L. and Schau, H. J. 2013. When differences unite: Resource
dependence in heterogeneous consumption communities. Journal of Consumer
Research, 39(5), pp. 1010–1033
 313

Thompson, S. A. and Sinha, R. K. 2008. Brand communities and new product adoption:
The influence and limits of oppositional loyalty. Journal of Marketing, 72(6), pp. 65–80.

Thorbjørnsen, H., Supphellen, M., Nysveen, H. and Egil, P. 2002. Building brand
relationships online: A comparison of two interactive applications. Journal of Interactive
Marketing, 16(3), pp. 17–34.

Thought Economics. 2012. The role of brands in human culture: interview of Kotler and
Lindstrom. [Online] [Accessed 12 September 2012]. Availabel at:
http://thoughteconomics.blogspot.co.uk/2012/02/role-of-brands-in-human-culture.html

Tildesley, A. E. and Coote, L. V. 2009. This Brand is Me : A Social Identity Based
Measure of Brand Identification. Advances in Consumer Research, 36, pp. 627–628.

Tönnies, F. [1887] 1957. Community and Society: Gemeinschaft und Gesellschaft

(Translated and Edited by Loomis, C. P.): The Michigan State University Press.

Traylor, M. B. and Joseph, W. B. 1984. Measuring Consumer Involvement i n Products

Developing a General Scale. Psychology, 1(2), pp. 65–77.

Trusov, M., Bodapati, A. V., and Bucklin, R. E. 2010. Determining influential users in
internet social networks. Journal of Marketing Research, 47(4), pp. 643–658.

Trusov, M., Bucklin, R. E. and Pauwels, K. 2009. Effects of word-of-mouth versus

traditional marketing: findings from an internet social networking site. Journal of
Marketing, 73(5), pp. 90–102.

Tucker, L. R. and Lewis, C. 1973. A reliability coefficient for maximum likelihood factor
analysis. Psychometrika, 38, pp. 1–10.

Tull, D. S., and Hawkins, D. I. 1987. Marketing Research: Measurement and Method (4th
Edition). London: Collier Macmillan Publishers.

Turner, B. 2001. Outline of a general theory of cultural citizenship. In Stevenson, ed.,

Culture and citizenship. London: Sage.

Tuškej, U., Golob, U., and Podnar, K. 2013. The role of consumer–brand identification in
building brand relationships. Journal of Business Research, 66(1), pp. 53–59.

Urban, G. 1996. Metaphysical community: The interplay of the senses and the intellect.
University of Texas Press
 314

Van de Vijver, F. and Tanzer, N. K. 1998. Bias and equivalence in cross-cultural

assessment. European review of applied psychology, 47(4), pp. 263–279.

Van Doorn, J., Lemon, K. N., Mittal, V., Nass, S., Pick, D., Pirner, P.and Verhoef, P. C.
2010. Customer engagement behavior: theoretical foundations and research
directions. Journal of Service Research, 13(3), pp. 253–266.

Vargo, S. L. and Lusch, R. F. 2004. Evolving to a new dominant logic for

marketing. Journal of Marketing, 68(1), pp. 1–17.

Veloutsou, C. 2007. Identifying the dimensions of the product-brand and consumer

relationship. Journal of Marketing Management, 23(1/2), pp. 7–26.

Veloutsou, C. and Moutinho, L. 2009. Brand relationships through brand reputation and
brand tribalism. Journal of Business Research, 62(3), pp. 314–322.

Veloutsou, C., Christodoulides, G. and de Chernatony, L. 2013. A taxonomy of measures

for consumer-based brand equity: drawing on the views of managers in Europe. Journal of
Product and Brand Management, 22(3), pp. 238–248.

Verhoef, P. C., Reinartz, W. J. and Krafft, M. 2010. Customer engagement as a new

perspective in customer management. Journal of Service Research, 13(3), pp. 247–252.

Vivek, S. D., Beatty, S. E. and Morgan, R. M. 2012. Customer engagement: Exploring

customer relationships beyond purchase. The Journal of Marketing Theory and
Practice, 20(2), pp. 122–146.

Vivek, S. D., Beatty, S. E., Dalela, V. and Morgan, R. M. 2014. A generalized scale for
measuring customer engagement. Journal of Marketing Theory and Practice, 22(4), pp.
401–420.

Wagner, C. and Majchrzak, A. 2006. Enabling customer-centricity using wikis and the
wiki way. Journal of Management Information Systems, 23(3), pp. 17–43.

Wallace, E., Buil, I. and de Chernatony, L. 2014. Consumer engagement with self-
expressive brands: brand love and WOM outcomes. Journal of Product and Brand
Management, 23(1), pp. 33–42.
 315

Walsh, G., and Beatty, S. E. 2007. Customer-based corporate reputation of a service firm:
scale development and validation. Journal of the Academy of Marketing Science, 35(1), pp.
127–143.

Wang, E. S. T., and Chen, L. S. L. 2012. Forming relationship commitments to online

communities: the role of social motivations. Computers in Human Behavior, 28(2), pp.
570–575.

Warrington, P. and Shim, S. 2000. An empirical investigation of the relationship between

product involvement and brand commitment. Psychology and Marketing, 17(9), pp. 761–
782.

Wiertz, C., and de Ruyter, K. 2007. Beyond the call of duty: why customers contribute to
firm-hosted commercial online communities. Organization Studies, 28(3), pp. 347–376.

Williams, R. L. and Cothrel, J. 2000. Four smart ways to run online communities. Sloan
Management Review, 41(4), pp. 81–91.

Wirtz, J., den Ambtman, A., Bloemer, J., Horváth, C., Ramaseshan, B., van de Klundert,
J., and Kandampully, J. 2013. Managing brands and customer engagement in online brand
communities. Journal of Service Management, 24(3), pp. 223–244.

Wright, K. B. 2005. Researching Internet‐based populations: advantages and disadvantages

of online survey research, online questionnaire authoring software packages, and web
survey services. Journal of Computer‐Mediated Communication, 10(3), 00–00.

Wu, L. and Chen, J. L. 2005. An extension of trust and TAM model with TPB in the initial
adoption of on-line tax: an empirical study. International Journal of Human-Computer
Studies, 62(6), pp. 784–808.

Yadav, M. S. and Pavlou, P. A. 2014. Marketing in computer-mediated environments:

research synthesis and new directions. Journal of Marketing, 78(1), pp. 20–40.

Yoo, B. and Donthu, N. 2001. Developing and validating a multidimensional consumer-

based brand equity scale. Journal of Business Research, 52(1), pp. 1–14.

Yoo, W. S., Suh, K. S. and Lee, M. B. 2002. Exploring the factors enhancing member
participation in virtual communities. Journal of Global Information Management, 10(3),
pp. 55–71.
 316

Zaglia, M. E. 2013. Brand communities embedded in social networks. Journal of Business

Research, 66(2), pp. 216–223.

Zaichkowsky, J. L. 1994. Research notes : the personal involvement inventory : reduction,

revision, and application to advertising. Journal of Advertising, 23(4), pp. 59–70.

Zeithaml, V. A. 1988. Consumer perceptions of price, quality, and value: a means-end

model and synthesis of evidence. The Journal of Marketing, 52(3), pp. 2–22.

Zhou, T. 2011. Understanding online community user participation: a social influence

perspective. Internet Research, 21(1), pp. 67–81.

Zhou, Z., Wu, J. P., Zhang, Q., and Xu, S. 2013. Transforming visitors into members in
online brand communities: evidence from China. Journal of Business Research, 66(12),
pp. 2438–2443.