Transparency is

the new black

A study of how transparent apparel supply chains
influence Generation Y.

MASTER
THESIS WITHIN: Business Administration
NUMBER OF CREDITS: 15 ECTS
PROGRAMME OF STUDY: International Marketing
AUTHOR: Louice Jonsson, Sara Ygge
TUTOR: Tomas Müllern
JÖNKÖPING May 2018
Master Thesis in Business Administration

Title: Transparency is the new black – A study of how transparent apparel

supply chains influence Generation Y

Authors: Louice Jonsson & Sara Ygge

Tutor: Tomas Müllern

Date: 21 May 2018

Acknowledgement

We want to express our warmest gratitude towards everyone who have offered us support during
the process of writing our thesis. First and foremost, we want to thank our supervisor Tomas
Müllern whose constructive, yet enthusiastic guidance and optimism have been invaluable.
Secondly, a special thank you to Johan Larsson who has shared his expertise in the quantitative
field of study, and who encouraged us to challenge ourselves. Lastly, we want to express our
appreciation towards our seminar group who have provided us with insights, support and kind
words.

Thank you,

Louice Jonsson & Sara Ygge

i
Abstract
The apparel industry is changing. Due to increased public pressure and in an effort to abide to
stakeholders’ demand, the CSR-efforts from apparel companies are shifting towards focusing on
the social practices in the supply chain. This has given an uprise to the notion of supply chain
transparency where companies disclose valuable information about who did what, in what factory
and under what circumstances. Previous studies on supply chain transparency either studies it from
a” best practice” business perspective, while others conclude a positive effect of the general
understanding of transparency on consumer perceptions. This study applies business theory of
supply chain transparency on theories of consumer behavior, namely the theory of reasoned action
(Ajzen & Fishbein, 1980). To this widely utilized behavioral model, the two dimensions of supply
chain transparency: degree of disclosure, i.e. the amount of information the company shares, and
degree of assessment, i.e. the perceived knowledge and insight the company appears to have, were
applied. This in order to provide a consumer perspective of supply chain transparency strategies,
more specifically the one of generation y. As sustainable and transparent business strategies are
gaining momentum, it is of essence to understand the outcome of such specific strategies. In order
to do so, a research model and hypotheses for further testing were established. A quantitative
survey was conducted, generating 108 responses on four different kinds of supply chain strategies.
By using AMOS 25.0 we performed structural equation modeling which tests all of the relationships
proposed in the model simultaneously. The results concluded that the perceived degree of
assessment is influenced by the degree of disclosure, meaning that consumers partly base the
perceived company assessment of the supply chain on the amount of information that is available.
Furthermore, the findings also suggest that consumers are more likely to base their trust, attitude
and purchase intention on the perceived degree of assessment. This study provides valuable
information regarding the generation y consumer perceptions of supply chain transparency in an
apparel context, and further contributes to theory by proposing a model for how the two
dimensions of supply chain transparency affects trust, attitudes and purchase intention.

ii
Table of Contents

1. Introduction.................................................................................. 1
1.1 Background....................................................................................... 1

1.2 Problem definition and purpose ........................................................ 3

1.3 Research Questions ......................................................................... 4

1.4 Key Words ........................................................................................ 5

2. Theoretical framework ................................................................ 6

2.1 Corporate Social Responsibility practices ........................................ 6

2.1.1 Apparel Corporate Social Responsibility ................................... 6

2.1.2 CSR-washing ............................................................................ 8

2.2 Transparency .................................................................................... 8

2.2.1 Supply chain transparency ........................................................ 9

2.2.2 The dimensions of supply chain transparency ........................ 10

2.2.2.1 Degree of disclosure ........................................................... 10

2.2.2.2 Degree of assessment ........................................................ 11

2.2.2.3 The supply chain transparency matrix ................................. 11

2.2.3 Transparency in apparel supply chains ................................... 13

2.3 Consumer behaviour ...................................................................... 15

2.3.1 Theory of Reasoned Action ..................................................... 15

2.3.2 Attitudes and social responsibility ........................................... 16

2.3.3 Trust and social responsibility ................................................. 17

2.3.4 Purchase intention and social responsibility ............................ 18

2.4 Generation Y - the conscious generation ....................................... 19

2.5 Research model and hypotheses ................................................... 20

3. Methodology .............................................................................. 23

iii
 3.1 Research philosophy ...................................................................... 23

3.1.1 Epistemological considerations .............................................. 23

3.1.2 Ontological considerations ..................................................... 24

3.2 Research approach ........................................................................ 25

3.2.1 Deduction ................................................................................ 25

3.2.2 Quantitative research .............................................................. 26

3.3 Research design ............................................................................. 27

3.4 Data collection method ................................................................... 29

3.4.1 Secondary data......................................................................... 29

3.4.2 Primary data ............................................................................. 29

3.5 Sampling ......................................................................................... 30

3.5.1 Non-probability sample ............................................................. 30

3.5.2 Sample selection ...................................................................... 30

3.6 Survey method................................................................................ 31

3.6.1 Self-completion questionnaire .................................................. 31

3.6.2 Scenario design ........................................................................ 31

3.6.3 Questionnaire design ................................................................ 33

3.7 Analysis of data .............................................................................. 36

3.8 Reliability and validity ..................................................................... 37

3.8.1 Reliability .................................................................................. 37

3.8.2 Validity ...................................................................................... 38

3.8.3 Pilot study ................................................................................. 38

3.9 Ethical considerations ..................................................................... 39

4. Empirical findings and analysis ............................................... 40

4.1 Sample demographics .................................................................... 41

4.2 Descriptive ...................................................................................... 42

4.2.1 Initial positions on social practices in the apparel industry ...... 42

iv
 4.2.2 Reliability ................................................................................. 44

4.2.3 Scenario comparison ................................................................ 45

4.3 Testing the assumptions for SEM ................................................... 47

4.4 Structural Equation Modelling ......................................................... 50

4.4.1 Model fit indexes ....................................................................... 50

4.4.2 Measurement model ................................................................. 52

4.4.2.1 Measurement model fit ....................................................... 52

4.4.2.2 Reliability and validity of the measurement model ............. 55

4.4.3 Structural model........................................................................... 55

4.4.3.1 Structural model fit ................................................................. 56

4.4.3.2 Hypotheses testing ................................................................ 57

4.4.4 Revised model ............................................................................. 58

4.4.4.1 Revised theory ....................................................................... 58

4.4.4.2 Revised model ....................................................................... 60

4.4.4.3 Revised model fit ................................................................... 61

4.4.4.4 Testing the revised model...................................................... 62

4.5 Hypotheses conclusion ................................................................... 64

5. Discussion ................................................................................. 65
5.1 Initial positions on social practices.................................................. 65

5.2 Scenario comparison ...................................................................... 66

5.3 Proposed model.............................................................................. 68

5.3.1 The relationship between degree of assessment and degree of

disclosure ......................................................................................... 68

5.3.2 Transparency as predictor of trust ........................................... 69

5.3.3 Transparency as predictor of general attitude ......................... 71

5.3.4 Proposed model....................................................................... 74

v
6. Conclusion ................................................................................. 75
6.1 Purpose and research questions .................................................... 75

6.2 Implications ..................................................................................... 76

6.2.1 Theoretical implications ............................................................ 76

6.2.2 Managerial implications ............................................................ 77

6.2.3 Societal implications ................................................................. 78

6.3 Future research .............................................................................. 79

6.4 Limitations....................................................................................... 81

References ..................................................................................... 83

vi
Figures
Figure 1: The supply chain transparency matrix by Marshall et al. (2016) ............... 12
Figure 2: Theory of Reasoned Action by Fishbein and Ajzen .................................. 16
Figure 3: Research model ........................................................................................ 20
Figure 4: Research overview .................................................................................... 23
Figure 5: Scenarios .................................................................................................. 32
Figure 6: Radar chart of scenario means ................................................................. 45
Figure 7: Structural model results ............................................................................ 58
Figure 8: Revised model .......................................................................................... 60
Figure 9: Structural revised model results ................................................................ 62
Figure 10: Proposed model of how supply chain transparency of social practices influence
consumers. ................................................................................................. 74

Tables
Table 1: Instruments ................................................................................................. 35
Table 2: Sample demographics ................................................................................ 41
Table 3: Initial positions on social practices in the supply chain .............................. 43
Table 4: Correlations on initial positions ................................................................... 44
Table 5: Measurement reliability - Cronbach’s Alpha ............................................... 44
Table 6: Scenario comparison .................................................................................. 46
Table 7: Model fit indexes with thresholds ............................................................... 51
Table 8: Measurement model fit ............................................................................... 53
Table 9: Measurement model results ....................................................................... 54
Table 10: Structural model fit ................................................................................... 56
Table 11: Revised structural model fit ...................................................................... 61
Table 12: Hypotheses conclusion ............................................................................ 64

Appendix
Appendix 1: Scenarios ............................................................................................. 93
Appendix 2: Questionnaire ....................................................................................... 97
Appendix 3: Item correlations matrix ...................................................................... 100
Appendix 4: Pearson correlation between constructs ............................................ 101
Appendix 5: Regression ......................................................................................... 102
Appendix 6: Curve Estimation ................................................................................ 107
Appendix 7: Cook’s distance .................................................................................. 110
Appendix 8: Kurtosis and Skewness ...................................................................... 112
Appendix 9: Homoscedasticity ............................................................................... 113
Appendix 10: Measurement model results ............................................................. 115
Appendix 11: Measurement model validity and reliability ....................................... 119
Appendix 12: Structural model results .................................................................... 120
Appendix 13: Revised model results ...................................................................... 122

vii
1. Introduction

1.1 Background
One of the most devastating event that has happened in the history of apparel production
took place in Bangladesh in 2013: The Rana Plaza factory collapse that resulted in the death
of 1133 and injured over 2500 workers (Centre for Policy Dialogue, 2013; Hobson, 2013).
The top floors of the building were built without planning permission, and the building itself
was not even intended to be used as a factory for thousands of workers and their machines
(Reinecke & Donaghey, 2015). While the multinational companies that the factories
produced for, such as Primark and Walmart, did not have any legal obligations in the
situation, the pressure on them quickly increased to take responsibility (Reinecke &
Donaghey, 2015). Consequently, in the aftermath of the tragic event an increasingly public
demand was placed on supply chains transparency, and the consumer awareness regarding
the issue quickly grew (Jacobs & Singhal, 2017; Methven O'Brien & Dhanarajan, 2016). This
also resulted in a multi-actor global mobilisation from unions, non-governmental
organizations and governmental institutions working to assure that companies act
responsible in terms of social rights, including working conditions such as labour rights, fair
payments and safety (Methven O'Brien & Dhanarajan, 2016).

A crucial aspect of corporate social responsibility (CSR) efforts in the apparel industry is the
public reporting of policies and practices related to the human rights (Dickson, 2013). Many
researchers even claim that the terrible incident at Rana Plaza could have been avoided if
there was an adequate supply chain transparency regarding the social and labour conditions
in the apparel industry (Burke, 2015). The apparel industry is highly globalized and often
have complex labour intensive supply chains, and therefore it is associated with many ethical
issues related to social practices (James & Montgomery, 2017). Apparel companies
sometimes work with thousands of factories at the same time, and many do not even know
where all of the different components of their garments are made (Fashion Revolution,
2017). The supply chain is further complex when accounting for the additionally thousands
of sub suppliers and sources of raw materials. This presents an explanation for why many

1
apparel companies do not have control of their supply chain, and sometimes not being aware
of the poor social conditions in some factories. Historically apparel companies have been
very selective in the information they provide about their practices regarding these matters,
but lately there is a clear shift in the industry where companies increasingly seem to engage
in social responsibilities (James & Montgomery, 2017). This shift is mostly due to the
increased public pressure, and companies aspires to proactively prevent external criticism
(Awaysheh & Klassen, 2010). The consumer demand for transparency in supply chains has
further been increased by the development of new technologies, mainly the internet which
allows consumers to access and spread information, making it difficult for companies to
control what information is available to the public (Awaysheh & Klassen, 2010).
Consequently, the power shift to the consumer, and pressure is put on the company to act
in accordance. On a market that demands transparency, an ethical supply chain therefore not
only is beneficial, but even necessary in order to comply with consumers and stakeholders
(Van Der Zee & Van der Vorst, 2005).

Today it is evident that many apparel companies try to convey an image of acting responsible
through transparency, and sustainable business practices are even becoming mainstream
(Henninger, Alevizou & Oates, 2016; Mittelstaedt et al., 2014). Some companies only include
social sustainability efforts briefly in their annual report, but others have chosen to include
the notion of transparency in their business model. In the context of the business
management of supply chain transparency there are two dimensions companies take into
account in order to decide their strategy: (1) degree of disclosure of the supply chain, which
is how much information the company wants to share, (2) degree of assessment of the supply
chain, which is the level of insight and knowledge the company has about their supply chain
(Marshall, Mccarty, Mcgrath & Harrigan, 2016). For example, companies that have high
degree of disclosure and high degree of assessment are Everlane and Honestby. These are
two examples of radical transparency, which includes transparent supply chains (Everlane,
2018: Honestby, 2018). Including supply chain transparency in CSR-efforts is evidently an
emerging trend with other examples such as Nudie Jeans co (2018) and Patagonia (2018)
which clearly take a stand for sustainable apparel manufacturing processes. However,
claiming to act responsible and being transparent can also be misused by companies in order
to deceive customers by portraying themselves as something they are not. These companies

2
usually have a high degree of disclosure, but a low degree of assessment. This is called CSR-
washing or, more commonly in terms of environmental sustainability, greenwashing. These
terms are defined as misleading advertising of sustainability credentials implying that the
company purposely communicates a positive sustainable performance, even though it does
not reflect the truth (Delmas & Burbano, 2011; Du, 2015).

As consumers are getting more aware of unethical business behavior, they are more inclined
to analyze and reflect over their own behavior and starting to change accordingly
(Harrison, Newholm, & Shaw, 2005). Many researchers have therefore in the last decade
acknowledged the general importance of transparency in building relationships between
consumers and companies, especially when communicating its social responsibility related
achievements (Reynolds & Yuthas 2008). As sustainability is becoming more important,
social responsibility efforts too become a focal point in the minds of consumers. The
consumers of generation y, born between 1980-2000 are especially concerned of global issues
(Goldman Sachs, n.d.; Hill & Lee, 2012). They conclude a consumer segment that has a
substantial disposable income, and are informed and aware of environmental and social
issues, and are further sceptic towards companies claiming to manage these (Hill & Lee,
2012). In the light of the supply chain transparency as an emerging trend within the CSR and
social sustainability, the generation y therefore poses as an interesting consumer segment to
study as they are aware of such issues, but likely to be sceptical, thus not responding well to
CSR-washing strategies.

1.2 Problem definition and purpose

Much of the recent academic research have concluded the positive influence of apparel
products related to CSR efforts on consumers purchase intentions (Kang & Hustvedt, 2014;
Hustvedt & Dickson, 2009; Lee, Choi, Youn & Lee, 2012), yet few studies have examined
how the young consumers of generation y understand and perceive CSR in the apparel
industry. Previous studies on CSR and its influence of consumer behavior have mainly been
considering specific attributes such as organic cotton apparel (Hustvedt & Dickson, 2009)
or fair-trade apparel (Shaw, Hogg, Wilson, Shiu & Hassan, 2006), and only a few have been

3
studying the effect of perceived supply chain transparency on consumers (Kang & Hustvedt,
2014). It is however evident that no studies, to the best of our knowledge, address how the
specific different business supply chain transparency strategies are perceived by consumers,
especially in the context of social practices within it, and let alone how these relate to
generation y. As sustainable and thereby transparent business strategies are gaining
momentum, it is of essence to understand the outcome of these specific strategies. This is
naturally important for companies in order to predict consumer behaviours, but also in order
to comprehend how to convince consumers to consume responsible. With a foundation in
the high versus low degree of disclosure and degree of assessment, we propose that
previously discussed strategies evoke different responses of consumers.

The purpose of this study is therefore to determine how disclosure of CSR-efforts related to
social practices in terms of the two dimensions of supply chain transparency, Degree of
Disclosure and Degree of Assessment, affect generation y’s perceptions and intentions in the
context of the apparel consumption. We further aim towards theory development in terms
of establishing a model explaining how these dimensions predicts outcomes in terms of
attitude, trust and purchase intention.

1.3 Research Questions

RQ1. What is the relationship between the dimensions of supply chain transparency, degree
of disclosure and degree of assessment, in the minds of generation y?

RQ2. How do the dimensions of supply chain transparency, degree of disclosure and degree
of assessment, affect the intention, attitudes and trust towards a brand among generation y
in an apparel context?

4
1.4 Key Words
Corporate Social Responsibility (CSR): A firm’s voluntary initiatives to act responsible in
respect to all of its diverse shareholders’ interests (Maignan & Ferrell, 2004; Du, Bhattacharya
& Sen 2010).
CSR-washing: When a company is claiming to act responsible and being transparent in
order to deceive customers by portraying themselves as something they are not (Siano,
Vollero, Conte, & Amabile, 2017; Boiral et al., 2017).
Degree of assessment: Degree of assessment of the supply chain, is the level of insight and
knowledge the company claim to have about their supply chain (Marshall et al., 2016).
Degree of disclosure: Degree of disclosure of the supply chain, is how much information
the company want to share regarding their supply chain (Marshall et al., 2016).
Generation Y: Consumers born between 1980-2000 (Goldman Sachs, n.d.).
Social practices in the supply chain: Refers to the management of social issues such as
upholding the human rights, safety, equity, ethics, health and welfare and philanthropy within
the supply chain (Mani et al., 2016).
The dimensions of supply chain transparency: Refers to degree of disclosure and degree
of assessment as two factors that decide what kind of image of responsibility is portrayed by
a company in the context of transparency (Marshall et al., 2016).
Theory of Reasoned Action: Refers to a theoretical framework or attitude-behaviour
model developed by Fishbein and Ajzen in 1967 to examine consumer behaviour (Ajzen &
Fishbein, 1980).
Transparency: Refers to the disclosure of information in a business context regarding
labour practices and working conditions in supply chains, and incorporates the extent to
which information regarding the supply chain is available to other actors, such as consumers
and other companies (Mol, 2015; Doorey, 2011; Carter & Rogers, 2008).

5
2. Theoretical framework

2.1 Corporate Social Responsibility practices

Corporate Social Responsibility, CSR, and its related practices and activities is a highly
contemporary topic both in academic literature as well as in the context of business
management. As definitions of the term varies, it is in this study understood as being: a firm’s
voluntary initiatives to act responsible in respect to all of its diverse shareholders’ interests
(Maignan & Ferrell, 2004; Du et al., 2010). This includes both responsibility towards the
potential effects on the environment as well as the social well being.

The umbrella term CSR includes the effect of social aspects, which is the meaning that is of
biggest interest for the case of this study, especially in the context of supply chains. Supply
chain social sustainability is therefore a relevant term, which refers to the management of
social issues such as upholding the human rights, safety, equity, ethics, health and welfare
and philanthropy within the supply chain (Mani et al., 2016). These terms are also what we
hereby will refer to as the social practices within the supply chain.

2.1.1 Apparel Corporate Social Responsibility

In the highly competitive apparel industry, leveraging fashionable garments and maintaining
a competitive cost advantage is of considerable matter (New, 1997). As part of business
globalization processes, apparel companies have acknowledged the advantage of outsourcing
their production to under-developed and developing countries due to enhanced distribution
facilities, increased accessibility (Laudal, 2010) and where cost competitive labour force can
be easily maintained (Diddi & Niehm, 2017). Numerous apparel companies are signing
contracts with independent suppliers and manufacturers in far distance countries which in
their turn uses local sub-suppliers, making it hard for them to keep track on who did what,
in what factory and under what circumstances (Masson, Mackerron & Fernie, 2007) and
negative exploitation of human resources sadly becomes a consequence (Diddi & Niehm,
2017; New, 1997). The apparel industry is often characterized with complex, inflexible and
uncooperative supply chains (Barnes & Lea-Greenwood, 2006). Therefore, it is due to
complexities such as lack of resources, insights and control which have made it challenging

6
for apparel companies to ensure good CSR practices throughout the supply chain activities
(Perry & Towers, 2009).

With extensive media attention addressing labour issues, labour standards and CSR within
the apparel industry, this has come under scrutiny over the last decade where consumers are
increasingly aware of harmful conditions affecting people’s welfare down the supply chain.
As a consequence, consumers have placed pressure and demand on companies to act
proactive of issues that they help create (Waddock, 2008; Pirnea & Ghenta, 2013; New,
1997).

Consumers are specifically pressuring apparel companies to be open in their reporting of

CSR initiatives where detailed information should be accessible and visible (Burchell & Cook,
2006). Nevertheless, since companies CSR reporting are often functioning as a foundation
for consumers to evaluate and judge their decision to purchase garments in order to decrease
their own impact of social issues (Marin, Ruiz & Rubio, 2009). In the potential of losing
consumers by not responding to their increase demands of CSR-reporting (Marin et al., 2009;
Groza, Pronschinske & Walker, 2011; Diddi & Niehm, 2017; Dillman, 2000), social practices
have become more integrated in apparel companies CSR reporting and in their tactical
business strategy (Moosmayer & Fuljahn, 2010; Diddi & Niehm, 2017). For example, some
apparel companies have actively been applying codes of conducts in order to highlight poor
labour practices (Shaw et al., 2006) and implemented philanthropy i.e. corporate donating
(Singh, Salmones Sanchez & Bosque, 2008; Lee, Park, Moon, Yang, & Kim, 2009). Other
have started educating consumers in fields of improvements, such as recycled and reused
materials (Pedersen, & Gwozdz, 2014) or applied certain standards (E.g. ISO 26000:2010) to
be guided on how to act in an ethical and transparent way in regard to issues such as human
rights, environmental footprint and socially performances (Diddi & Niehm, 2017).

7
2.1.2 CSR-washing

Companies often respond to this new demand from stakeholders by adapting a symbolic
compliance as a way to improve legitimacy (Boiral, Heras- Saizarbitoria & Testa, 2017). The
results from such acts is that these standards are not internalized in the company and its
practices, but rather superficially adopted. In environmental sustainability, this is commonly
known as greenwashing, while the umbrella term that is more relevant to this study is CSR-
washing (Siano, Vollero, Conte, & Amabile, 2017; Boiral et al., 2017). The term refers to
deceptive communication, and can be described as “a disconnect between the positive image
projected to stakeholders with regard to CSR and a company’s actual internal practices in
this area” (Boiral et al., 2017, p. 57). Thus, it can be understood as the gap between symbolic
statements and substantive actions, and by displaying a CSR-righteous image, companies seek
to be perceived as ethical and conscious in the eyes of the stakeholders without putting the
effort into actually being it. This emerging trend have several possible reasons besides
responding to external pressure, but also in order to attain reputational capital (Aras &
Crowther, 2011) and possible financial gains (Jonsen, Galunic, Weeks & Braga, 2015). Some
examples of this are companies that voluntary develop their own CSR-standards that appear
to be genuine and substantial, yet avoiding complying with industry standards, or companies
that augment their CSR efforts in order to shift stakeholders’ attention away from potential
wrongdoings (Marquis & Toffel, 2012).

2.2 Transparency
Within the CSR construct the word transparency has become a highly relevant, especially in
the apparel industry. The notion of transparency might not be new, yet seems to have had
quite the renaissance in the recent years due to increased awareness of consumers. Many
scholars have given a definition to the general understanding of transparency, however these
differ in approach depending on its context (James & Montgomery, 2017). Ball (2009) for
example uses a wide understanding of it being synonymous with openness, while Mol (2015)
and Doorey (2011) with a similar wide understanding describes it as the disclosure of
information in a business context. Transparency International (2018) provides a

8
straightforward definition of transparency: “Transparency is about shedding light on rules,
plans, processes and actions. It is knowing why, how, what, and how much.”

2.2.1 Supply chain transparency

In a highly globalized context where external suppliers and sub suppliers are common
business practices, the demand for transparency is extended beyond the corporate
boundaries and into supply chains (Mol, 2015). In this context researchers sought a need for
a more specific definition of what we hereby will refer to as supply chain transparency which
incorporates the extent to which information regarding the supply chain is available to other
actors (Awaysheh & Klassen, 2010).

Cramer (2008) refers to supply chain transparency as the disclosure of supplier information,
Carter and Rogers (2008) defines it as the amount of information a company is willing to
disclose about their supply chain practices, and O’Rourke (2003) even describes it as being
central factor of which one is able to judge a company’s supply chain. Similarly, Doorey
(2011) and Laudal (2010) adopts an ability of traceability through production to the
approach. Egels-Zandén, Hulthén and Wulff (2014) specifically describes it as disclosure of
information about supplier names, sustainability conditions at suppliers, and buyers
purchasing practices. Additionally, researchers often pay specific attention to the elements
of ethics and sustainability in the supply chains (Mol, 2015). Thus, the debates are plenty and
a shortage of empirical evidence on how a company should act in order to be classified as
transparent is still not cohesive.

Researchers today deem transparency to be desirable for corporations in order to achieve

favourable characteristics such as legitimacy (Carter & Rogers, 2008), accountability
(Dubbink, Graafland, & Van Liedekerke, 2008) and trust (Augustine, 2012), which all
ultimately relates to the perception of the company in the eye of the consumer. Reynolds and
Yuthas (2008) even argue that transparency built on disclosing CSR efforts is one of the
essential conditions for the development of positive relationships between consumers and
companies. This is due to the fact that disclosure of information of supply chains becomes
the foundation of knowledge and awareness of a company’s business practices for

9
stakeholder (James & Montgomery, 2017). By allowing them to make informed decisions
about their offerings based on their ethical and sustainability practices, they can put pressure
on the companies to act in a responsible manner (Egels-Zandén, Hulthén & Wulff, 2014;
Ortiz Martinez & Crowther, 2008).

Mol (2015), and Dingwerth and Eichinger (2010) have presented transparency as a powerful
tool for stakeholders in order to hold authoritative companies accountable for their actions.
The demand for transparency can from this perspective be perceived as somewhat
problematic from a company view, as it in itself shifts the power from the companies to the
stakeholders and consumers. Some large corporations have further criticised and questioned
the public demand for transparency, as they argue that their information regarding supply
chain is too valuable to be disclosed (Doorey, 2011).

2.2.2 The dimensions of supply chain transparency

2.2.2.1 Degree of disclosure

As Fashion Revolution (2017, p.7) describes it: “If you can’t see it, you can’t fix it”. The
degree of disclosure, hence the amount and type of information regarding the supply chain
is an essential element in achieving legitimacy as it allows external actors to monitor
sustainability claims (Laudal, 2010). By making information regarding the supply chain
public, it helps stakeholders, non-governmental organizations, unions, communities and the
workers themselves to reinforce and uphold ethical standards (Fashion Revolution, 2017).
The transparency in terms of disclosure thereby facilitates the understanding from all
involved parts of what went wrong, who is responsible and how to fix it, and thereby
reinforcing the accountability of companies.

Companies can further gain competitive advantages when increasing their level of
information disclosure. By applying the “we have nothing to hide” approach toward
openness can be a way of differentiating themselves from the companies that perform badly
in this matter. On the other side, some companies are only acting transparent on their more
successful examples of their activities, called CSR-washing. This has lead stakeholders to

10
react with scepticism to solely rely on the information disclosed by the company (Egels-
Zandén, Hulthén & Wulff, 2015).

2.2.2.2 Degree of assessment

The idea of companies not having full knowledge about their supply chain might seem
improbable, but the truth is that many companies have a very limited visibility of their supply
chain, and further having a poor understanding of how to acquire the information, what it
means and how to report it (Marshall et al., 2016). This is because many organizations have
highly complex and globalized supply chains, which cross many national borders and can
pass through hundreds of workers during the process (Marshall et al., 2016). Degree of
assessment refers to the degree of control, knowledge and insight the company has regarding
their supply chain, thus whether or not they have a good grasp of the information (Marshall
et al., 2016). A high degree of assessment increases the companies’ ability to track suppliers,
which decreases the risk for including unauthorized suppliers that do not comply with CSR
standards in the chain, and thereby harming the reputation of a company (Fashion
Revolution, 2017).

Additionally, it includes if this information is reported accurately and appropriately. Within

this term we also chose to include third-party standards and independent auditors, such as
SA8000, the leading social certification standard for factories and organizations (Social
Accountability International, 2018), and Fairtrade certification (Fairtrade, 2018).

The degree of disclosure and the degree of assessment will hereby also be referred to as the
two dimensions of supply chain transparency.

2.2.2.3 The supply chain transparency matrix

In order for a company to select a supply chain disclosure strategy, they first need to account
for previous discussed elements such as: their knowledge and insight in the supply chain,
what type of information to disclose, and how much of it they want to reveal. Marshall et
al. (2016) included these aspects when they developed their transparency matrix, in which
the two dimensions of transparency, degree of disclosure and degree of assessment, mirrors

11
four typical supply chain disclosure strategies. The strategies in this are based on the amount
and depth of evaluation and the level of public data disclosure.

Figure 1: The supply chain transparency matrix by Marshall et al. (2016)

The companies that fall under the Transparent strategy strive for the greatest amount of
disclosure of information regarding the supply chain, often both internally within the
company as well as externally towards stakeholders. In general, these companies tend to
perceive disclosure as a core competence and capability. An excellent example of this is the
apparel company Everlane, who has adopted what they call “radical transparency” that
includes total disclosure of their factories and costs (Everlane, 2018). Everlane has taken the
general notion of transparency and transformed it into the foundation of their business
model. The category Secret Companies include strategies in which the companies have high
degree of assessment, thus knowledge about their supply chain activities, but only disclose
little or no information to external actors. These companies view their supply chain
information as intellectual property that creates competitive advantage and therefore
deliberately only conceal supply chain information. Examples of activities that can be secret
could be development of new products manufacturing processes, sources of supply, recipies,

12
technical specifications and customers and suppliers. Coca Cola has for example never
revealed their secret recipe, and the same goes for KFC’s fried chicken batter. The Distracting
Companies on the other hand report information to external actors, but in a way that
unpurposely or purposely distracting. The companies which unpurposely distract are simply
not able to understand or evaluate what information is of relevance and report excessively
on other or all kinds of practices. On the other hand, some companies deliberately distract
external actors and stakeholders from discovering questionable practices by focusing
excessively on self selected practices, while deemphasizing others. This strategy includes
CSR-washing practices, which are common in many industries among them the one of
apparel (Fashion Revolution, 2017). The last category is the Withheld Companies, that entirely
withhold information from external actors and stakeholders. The reasons for them not to
disclose information regarding the supply chain could either be that they simply do not have
the information and thereby unknowingly withhold information, or that they intentionally
avoid disclosing information because it includes harmful practices that would have a negative
effect on the company's reputation.

2.2.3 Transparency in apparel supply chains

In 2013 the apparel factory at Rana Plaza collapse resulting in the death of 1133 workers,
and further 2500 injured (Centre for Policy Dialogue, 2013). The factories operating in the
building produced apparel for many well-known global brands, and although labels from
among others Primark and Walmart were found in the ruins it took the companies weeks to
determine their connections to the factories (Reinecke & Donaghey, 2015: Fashion
Revolution, 2017). In the aftermath of the terrible incident the apparel industry has
undergone public scrutiny for not ensuring the safety and health, and many researchers even
claim that it could have been avoided if the supply chain would have been transparent (Burke,
2015). Later it was found out that the top floors of the factory were built without proper
planning permission, and was not even intended to be used as a factory housing for
thousands of workers and their machines (Reinecke & Donaghey, 2015). The incident
marked the beginning of a global movement, which today have resulted in an increased
awareness among consumers of the supply chains of apparel and its related social issues.
Dickson (2013) stated that transparency is a crucial aspect of CSR-efforts in the apparel

13
industry in terms of the public reporting of policies and practices related to the human rights.
Transparency is even being perceived as the solution to many of the complex problems
facing apparel organizations today (Egels-Zandén, Hulthén & Wulff, 2014).

Fashion Revolution, a global movement working for greater transparency, sustainability and
ethics in the fashion industry, yearly releases reports with transparency evaluations of the 100
of the biggest global fashion and apparel brand and retailers. They include “disclosure of
brands’ policies, procedures, goals and commitments, performance, progress and real-world
impacts on workers, communities and the environment” (Fashion Revolution, 2017, p. 13).
They further specify that transparency in the apparel industry includes knowing who stitched
what, who dyed the fabric and who farmed the cotton. The organization have identified five
key areas of transparency in the apparel industry that are evaluated in the report: (1) policy
and commitments, meaning whether or not the company actually have policies, how they are
put into practice and future goals; (2) governance, including naming the people responsible
for social and environmental impact, how to contact them, and how the company
incorporates human rights into buying and sourcing practices; (3) traceability, mainly relating
to suppliers, if they are public or not and how much details that are provided; (4) know,
show and fix; which refers to how the company assess the implementation of its supplier
policies, how they fix problems at supplier facilities, and how themselves and workers are
able to report grievances; and lastly (5) spotlight issues; which includes questions like how
companies can ensure that workers are being paid a living wage, what companies are doing
to support workers freedom of association, and how the company are reducing consumption
of resources.

14
2.3 Consumer behaviour
A consumer is explained to be anyone who has recognized a need or desire and is involved
in the process of selecting, purchasing or disposes a product aiming to seek satisfaction.
Consumer behaviour becomes relevant to study in order to better understand consumers
needs and purchasing patterns as the behaviours are often affected by multiple factors and
seen as rather complex since each consumer or consumer groups, are driven by their own
attitudes and motivations to purchase (Solomon, Bamossy, Askegaard & Hogg, 2010).

2.3.1 Theory of Reasoned Action

The theory of reasoned action (TRA) was first introduced in 1967 by Fishbein and Ajzen,
and has ever since been vividly applied as a conceptual framework to examine consumer
behaviour (Ajzen, 2011; Ajzen & Fishbein, 1980). TRA is well documented among
researchers and is considered to be one of the most popular attitude-behaviour models in
the social science literature (Solomon et al., 2010).

As the name may indicate, the theory suggests that an attitude is formed by the individual’s
belief to engage in a certain behaviour and is based on the expected outcome from
performing that particular behaviour (Ajzen & Madden, 1986). Simply put, the behaviour is
determined by the intention (Ajzen, 1991). The intention according to Ajzen (1991) will
capture the motivational factors that power a behaviour. This involves how hard a person is
willing to try, or the amount of effort he/she is planning to employ to perform the behaviour.
TRA specifies that there exist two major determinants that will predict the intention. These
are attitudes toward the behaviour, and the pressure of subjective norms (Fishbein & Ajzen,
1975). Ajzen and Fishbein (1980) define attitudes toward the behaviour as the person's own
personal favourableness or non-favourableness toward that behaviour, whereas subjective
norms are conveyed as a person's perception of important social influences that matters to
him/her such as family and friends (Ajzen & Fishbein, 1980; Ajzen & Madden, 1986). TRA
is working from the approach of rationality, meaning that people make reasonable use of the
information that they are proposed to. Therefore, the behaviour is explained to be a result
of a conscious decision (Fishbein & Ajzen, 1975). The TRA has during decades been applied

15
to many different studies in many different contexts, and some of these will be discussed
under 2.5 Research model and hypotheses.

Figure 2: Theory of Reasoned Action by Fishbein and Ajzen

2.3.2 Attitudes and social responsibility

In accordance with the TRA, attitudes are important as they form the basis of consumer
behavior (Keller, 1993; Ajzen & Fishbein, 1980), however, attitudes are for many reasons
quite difficult to fully comprehend. For example, consumers can have the same attitude
towards something but base it on various of different reasons (Keller, 1993; Ajzen &
Fishbein, 1980; Ajzen, 1991). Research has for example shown that attitudes are influenced
by a company’s competence (Castaldo, Premazzi & Zerbini, 2010), transparency of supply
chain activities (Kang & Hustvedt, 2014) and disclosing CSR-efforts (Diddi & Niehm, 2017;
Hillenbrand, Money & Ghobadian, 2013).

According to the Theory of Reasoned Action, attitudes are influenced by behavioural beliefs
of likely consequences from a given behaviour (Ajzen, 1991). Values along with beliefs,
knowledge and personal characteristics serve a base for the formation of attitudes (Ajzen &
Fishbein, 1980). However, attitudes are seen as more specific when they correspond in a
context to a given behaviour (Ajzen & Fishbein, 1980). Scholarly research pertaining to the
apparel industry with focus on social issues found that beliefs regarding ethical issues can be
impacted by a range of factors that influence consumers attitudes in various extent (Dickson,
2001; Kozar & Hiller Connell, 2013). For example, Dickson (1999) examined consumers
concern and attitudes toward labour issues in the apparel industry and found that consumers

16
are predominately searching for solutions to problems such as abandonment of child labour.
Other research suggests that social practices in the supply chain such as insufficient and
socially unacceptable working conditions also impact the attitudes towards a brand, resulting
in a decreased purchase intention (Kozar & Hiller Connell, 2013).

Education has been referred to as another potential factor that impacts attitudes of
consumers (Shim, 1995; Dickson, 2000). Similarly, Dickson (2000) stated that increased
knowledge regarding social issues will most likely lead to increased concern among
consumers and thereof impact their attitude. It has therefore been suggested that an
increased disclosure of responsible business practices in the supply chain should be taken
into consideration by companies in order to positively influence consumer attitudes, and
thereby purchase intention (Dickson, 2000).

Correspondingly, literature suggest that consumers feel good when they purchase a brand
that is associated with social responsibility behaviour (Poortinga & Vlek, 2004). This was
further supported by Kim, Littrell and Ogle (1999) who stated a positive relationship
between attitudes and apparel purchase from socially responsible companies. A similar
research conducted by Yan, Ogle and Hyllengard (2010) proposed a positive relation
between consumers attitudes as a response to a firm's aggressive advertising of social
responsibility efforts.

2.3.3 Trust and social responsibility

Morgan and Hunt (1994) explained trust as confidence in the other party’s reliability and
integrity. More recent definitions have been conducted by other scholars such as Chaudhuri
and Holbrook (2001) where they explain trust as a consumer’s belief that a company acts in
concern of consumers and that the company keeps what it promises (Erdem & Swait, 2004).
This includes a consumers’ belief that the company acts responsibly, legally, ethically and
favourable toward the society’s best interest (Pavlou & Fygenson, 2006).

Trust is essential as it is suggested to play an important role as a mediator to and foundation

of any consumer-company relationship. Nevertheless, trust becomes imperative as previous
scholars have found that trust can predict positive outcomes such as consumer willingness

17
to act, loyalty and overall market performance (Erdem & Swait, 2004; Chaudhuri &
Holbrook, 2001). Even more important, many previous studies have concluded the positive
direct effect of trust on specifically the purchase intention (Chiu, Huang & Yen, 2010; Gefen,
Karahanna & Straub, 2003; Kang & Hustvedt, 2014; Setiawan & Aachyar, 2012).

Even though trust has a widely acknowledged effect on consumer behavior, limited research
has address potential factors that will help predict what is actually constructing trust (Kang
& Hustvedt, 2014). Nevertheless, trust is explained to be built in a slow pace over a period
of time through consumers set of beliefs focusing primarily on honesty, benevolence and
competence, which are all concepts included in social responsibility (Doney & Cannon,
1997). Middlemiss (2003) therefore suggested that a company should start by being credible,
transparent and honest in order to reach a trustworthy reputation. Hillenbrand et al. (2013)
further demonstrated that when a company is engaging in social responsibility, it will yield
trust and positive intent of consumers toward the company. Disclosure of information will
influence consumers trust positively as the consumers receive access to information to
evaluate by themselves (Kang & Hustvedt, 2014). Attempts to form trust from apparel
companies can thus be seen in their increased engagement in CSR, and efforts and initiatives
related to social responsibility (Kang & Hustvedt, 2014).

2.3.4 Purchase intention and social responsibility

The decision making of today’s consumers are driven by a sense of moral obligation towards
society and the environment (Shaw et al., 2006). The new sustainable and ethical consumer
therefore differs from others as they are considerably more motivated to purchase from
companies that promote fair working conditions in order to decrease the risk of having an
indirect negative impact on society (Lundblad & Davies, 2016). Dickson (2001), similarly
found that consumers purchasing decisions are increasingly affected by ethical concerns such
as the risk of negative social impacts. In this regard, apparel companies that are involved in
sustainability and CSR can offer an ethical choice for these conscious consumers, yet still
meet the consumption demand of consumers and their need for identity construction
(McNeill & Moore, 2015). The potential explanations for consumers increased motivation
to base their purchase decisions on CSR-related business practices is widely discussed.

18
Consumers that aim to act sustainable or ethical in their fashion purchasing behaviour are
for example inclined to be driven by end goals such as self-expression, group conformity,
aesthetic satisfaction (Lundblad & Davies, 2016), ethical obligations (Shaw et al., 2006) or
avoiding feelings such as guilt (Ha-Brookshire & Hodges, 2009).

In research it is sometimes somewhat difficult to study the actual purchase behavior, which
is why scholars often base their research on the purchase intention. In accordance with the
TRA the behavior is determined by the intention (Ajzen, 1991). The purchase intention is
therefore the likelihood that a consumer will plan or be willing to purchase a product in the
future (Wu, Yeh & Hsiao, 2011). Even though this is not a direct evidence that an actual
purchase will be performed, purchase intention is very useful in consumer research to predict
the buying decision.

2.4 Generation Y - the conscious generation

Generation y includes the people born between 1980-2000 (Goldman Sachs, n.d), and is a
large and powerful group of consumers that just recently entered adulthood and thus having
a long future of consumer decisions (Hwang, Lee & Diddi, 2014). The generation y is
considered sophisticated, informative, technology-wise, and is further characterized by being
selective about who they listen to. Generation y consumers want to experience the world for
themselves and make their own judgement toward a company’s operations (Jin Ma, Littrell
& Niehm, 2012; Hwang et al., 2014). The information empowerment that characterize the
generation, in combination with a high level of awareness and being action oriented, makes
them sceptical towards companies that claim to be concerned about environmental and social
issues, this including businesses with transparent supply chains (Bhaduri & Ha-Brookshire,
2011; Hwang et al., 2014). Especially as generation y is seen as a rather sceptical segment,
research has found them to respond better to the presentation of a company’s investment in
environmental and social causes than of those who evade engagement in these matters.
However, this has only been proven to function as long as the consumers is perceiving
companies to be authentic (Bhaduri & Ha-Brookshire, 2011) and when this is the case,
consumers are according to Straughan and Roberts (1999) more willing to engage in

19
responsible behaviours as they experience an increased control to improve environmental
and social issues.

2.5 Research model and hypotheses

The TRA model has as aforementioned been used in many different contexts in order to
understand consumer behaviors, including the CSR and transparency context. Hillenbrand
et al. (2013) for example based their research model on TRA, in a study with the purpose to
examine how social responsibility affected the trust of employees and consumers. Another
study by Yan et al. (2010) modified the TRA model in order to study how a brand with a
very explicit communication of CSR efforts affected consumer purchase intention. Both
studies suggested that CSR efforts directly influences the consumer behavior in different
ways. Kang and Hustvedt (2014) further used the TRA in developing their research model.
The study aimed to identify how trust is built, by studying the relationships between
transparency, social responsibility, trust, attitude, word-of-mouth intention and purchase
intention. As this model is based on Theory of Reasoned Action that has been proven to be
especially useful in social science literature, our confidence of selecting this model as a base
in order to answer the purpose is therefore increased.

As this study seek to measure how the two dimensions of supply chain transparency degree
of disclosure and degree of assessment affect trust, general attitude and purchase intention
these two dimensions will serve as the independent variables from which we want to examine
its direct effects on other variables.

Figure 3: Research model

20
The literature review concludes the relevance of studying how consumers perceive and react
to CSR within supply chains and the transparency of such claims. Based on the literature we
would expect that CSR efforts and perceived transparency in fact do affect consumers
attitudes and trust towards companies. The perceived CSR-efforts and the transparency of it
impact the trust according to Doney and Cannon (1997), Middlemiss (2003), Hillenbrand et
al. (2013), and Kang and Hustvedt (2014). Therefore, we derive the following hypotheses:

H1: High degree of disclosure will have a direct positive effect on trust towards the
company.

H3: High degree of assessment will have a direct positive effect on trust towards the
company.

The perceived social responsibility and the transparency of it impacts the attitude according
to Dickson (1999;2000;2001), Shim (1995), Kang and Hustvedt (2014), Diddi and Niehm,
(2017) and Hillenbrand et al. (2013). Therefore, we derive the following hypotheses:

H2: High degree of disclosure will have a direct positive effect on general attitude
towards the company.

H4: High degree of assessment will have a direct positive effect on general attitude
towards the company.

We can further expect that trust have an effect on the behavioral intention and thereby the
actual behaviour in terms of purchase intention. Erdem and Swait (2004), Kang and Hustvedt
(2014), Chaudhuri and Holbrook (2001), Chiu et al. (2010), Gefen et al. (2003), and Setiawan
and Achyar (2012) all suggests this relationship. This leads us to the following:

H5: Trust will have a positive direct effect on purchase intention

21
Additionally, we can expect that general attitude have a effect on the behavioral intention
and thereby the actual behaviour in terms of Purchase Intention as suggested by the TRA
(e.g. Ajzen & Fishbein, 1980) and Kim et al. (1999). This leads us to the following:

H6: Positive general attitude will have a positive direct effect on purchase intention.

22
3. Methodology

In our research, we aimed toward a structured, clear and concise methodological approach
in order to make sure that we were able to reach our purpose and to answer our research
questions. Thus, we presented a visual overview for the reader to easily interpret the various
stages that we followed.

Figure 4: Research overview

3.1 Research philosophy

3.1.1 Epistemological considerations

Research in general always contains sets of important assumptions about how one perceives
the world, and these assumptions construct the research strategy as well as its methods
(Saunders, Lewis & Thornhill, 2009). Epistemological issues are one kind of assumptions
that regards what should be considered acceptable knowledge within a discipline. It further
questions whether or not the social world can be studied guided by the same principles,
procedures and ethos as the natural sciences (Bryman & Bell, 2011). Positivism is one of

23
these views that adheres that only factual knowledge gained through observations, including
measurements is trustworthy. In this study, we gained our factual knowledge through
quantitative observations of reality that we statistically measured. We adopt this positivistic
position in which we affirm the importance of imitating the natural sciences, and thereby the
application of such methods on the study of social reality (Bryman & Bell, 2011).

The central issue in the continuous discussions regarding epistemological considerations is

that some scholars and research traditions reject the idea that the same principles that are
applied in natural sciences can be used in the study of social reality (Bryman & Bell, 2011).
The alternative philosophical approach that can be applied in social sciences is the
interpretivism that contrasts positivism. Interpretivism suggests that the subject matter of
social sciences, that is people and institutions, is radically different from that of natural
sciences. It further requires the social scientist to comprehend to subjective meanings of
social behavior (Bryman & Bell, 2011). When adopting a positivistic position in this study,
we were instead limited to data collection and interpretation in an objective way which we
believed to be suitable for our research (Saunders et al., 2009).

As interpretivism strives for the understanding of human behavior, positivism aims towards
explaining it, which was in line with the purpose of this study which was to determine how
disclosure of CSR-efforts in terms of the two dimensions of supply chain transparency,
degree of disclosure and degree of assessment, affect consumer perceptions and intentions
in the context of the apparel consumption. In accordance with positivism, we used existing
theory in our development of hypotheses that were tested (Saunders et al., 2009).

3.1.2 Ontological considerations

Ontological questions are in social sciences concerned with social entities, and whether or
not they should be considered objective, thus external to social actors, or social
constructions, that are created by the social actors. The two different positions within this
consideration are objectivism and constructivism, and in this research, we will adopt the
former of the two (Saunders et al., 2009). Objectivism implies that social phenomena and
their meanings exists independent of social actors, and that they further prevail in a separate

24
construct from actors. In line with our research, objectivism approaches the social
phenomena as existing whether or not we chose to study it, and in accordance to this line of
reasoning we believe that our objects and their meanings exists independent of us as
researchers or other actors. Constructivism on the other hand emphasizes that social
phenomena and their meanings are continually being accomplished by social actors, and
further that they are in a constant state of revision (Saunders et al., 2009). We take an
objective stance in our study in concern of how we approach our respondents and the
research. However, we also acknowledge that there are indeed certain elements in the studies
of ethical controversies that are more contextual in the sense of that respondents often are
affected by social biases and subjective norms. As we do consider this constructivist element,
we suggest that an objectivist ontological approach is most suitable for the purpose of our
study.

3.2 Research approach

3.2.1 Deduction

Deductive theory is the most common perception of the relationship between theory and
research in social sciences (Bryman & Bell, 2011). With a foundation in what is known about
a particular research area and the theoretical considerations within it, a researcher deducts
one or more hypotheses that will be subjected to empirical scrutiny. In this research approach
the theory and the hypotheses inducted from it drives the data gathering, meaning that we
hold a theory and make predictions of its consequences based on it. The contrasting
approach of induction instead implies that theory is the outcome of the research, and that
the process includes drawing generalizable inferences from observations (Bryman & Bell,
2011). By having a deductive approach, we instead aim to serve the purpose of this study, as
we strive for drawing more specific inferences from observations based on general theory.
Reasons for why a mainly deductive approach is suitable for our study is that there already
exist many theories and literature on the subject, and we therefore have a substantial
knowledge foundation to base our research on. The hypotheses and research model in this
study are derived from a vast literature review, and based on previous research, which is a
clear deductive inference.

25
However, as research is often not conclusively deductive, our research includes certain
inductive inferences. For example, inductive elements in our study are the dimensions of
transparency, which have not yet been utilized in the way we propose them to be, but instead
originally were proposed to fill a managerial purpose. Furthermore, does the process of
induction include revision of theory in which the researcher infers the implications of his or
her findings for the original theory, which is an important component of our research as we
eventually establish a revised theory and model (Bryman & Bell, 2011). On this note Bryman
and Bell (2011) suggests that instances when the data does not fit the original hypotheses is
a reason to avert the logical linear deductive process, thus including the revised theory that
is ought to be the outcome of our research. This clearly demonstrates that there exist some
overlapping of induction and deduction in this study.

3.2.2 Quantitative research

The method chosen for this empirical study is a mono-method which implies that the
primary data collection will be of quantitative nature and follows a deductive approach
(Saunders et al., 2009). As we in our study aim toward determining the causal relationships
between variables, a quantitative study can further be deemed as appropriate as causality is a
strong concern in most quantitative research. This is further in line with a positivistic
philosophy, as the natural sciences often proceeds in this way (Bryman & Bell, 2011). A
quantitative research enabled us to collect large amount of numerical data which we
interpreted statistically (Saunders et al., 2009; Babin & Zikmund, 2016).

Further, a quantitative approach enabled the variables within the modified research model
to be accurately measured. General attitude, trust and purchase intention in relation to the
dimensions of supply chain transparency were analysed through quantitative measurements
by using Likert scales in order to measure how much the respondents agreed or disagreed
with the presented questions (Saunders et al., 2009). We deemed this to be more appropriate
since a qualitative research commonly uses interviews or focus groups and categorizes its
data with small samples in its discovery-oriented approach (Saunders et al., 2009). A
qualitative method is further based on subjective judgements where hypotheses are more
likely to be hindered based on small sample sizes (Babin & Zikmund, 2016). Since our

26
quantitative research allowed us to collect a larger sample size, we were therefore able to
support or reject our hypotheses (Babin & Zikmund, 2016).

Each approach, qualitative or quantitative have both advantages and disadvantages where
the missing of statistical aspect in the qualitative research is seen as the major difference
between these two (Babin & Zikmund, 2016). One major drawback from choosing a
quantitative research was the lack of gaining knowledge on consumers deeper feelings and
thoughts regarding the two dimensions of supply chain transparency. However, we felt more
agile to withdraw evidences in terms of numerical values in order to draw a generalized
conclusion. This allowed us to accurately and statistically analyse our findings where we
compiled our discoveries in numbers and diagrams.

3.3 Research design

The methodology chapter has until now explained our philosophical position of positivism,
our objectivism as well as the deductive inferences that will guide this research. Based on
these considerations, a research design will be built that will guide us towards the purpose of
the study. Research design includes the related strategies, choices and time horizons and is
build upon the previous assumptions. As our research aims towards examining causal
relationships between variables, namely how the independent variables that is the two
dimensions of supply chain transparency affects the dependent variables of trust, general
attitude and purchase intention, the research design is termed explanatory research (Saunders
et al., 2009). In explanatory research designs, one seeks to explain the relationship between
variables and to do this through studying a specific situation or problem. Explanatory
research takes a foundation in the research hypotheses that specify the nature and direction
of the studied variables, much like our study (Babin & Zikmund, 2016).

The research strategy on the other hand is guided by the hypotheses and objective of the
study, the extent of existing knowledge, the amount of time and other resources as well as
previously discussed philosophical standpoints (Saunders et al., 2009). A common strategy
that usually is associated with a deductive approach is the survey strategy, that also is
particularly popular in business research (Bryman & Bell, 2011). As the researchers in
quantitative studies usually are concerned with generalizability (Bryman & Bell, 2011), and

27
because it often is a goal in explanatory studies (Babin & Zikmund, 2016), a survey is suitable
as it enables us to collect a large amount of data from a sizeable population, and when
sampling is used it could generate representative findings (Saunders et al., 2009). Further, our
hypotheses are founded on the premise that two independent variables affect other variables
and as it in this case is essential to be able to compare the effects, the survey strategy is further
beneficial as the results from it will be standardised and thereby allowing easy comparison
(Saunders et al., 2009). In the same line of arguing, quantitative researchers are generally also
concerned with the ability to measure, and the survey strategy allows this in a suitable way
(Bryman & Bell, 2011). By using such a strategy, we were later able to utilize the explanatory
tool of structural equation modelling (SEM) (Babin & Zikmund, 2016) which provided us
with advantages in terms of that we were able to clearly see that the independent variables
of the dimensions of supply chain transparency did not have the initially proposed
relationship.

As we were faced with time constraints and lacked adequate resources, a survey strategy is a
pragmatic solution in which we were able to collect a great amount of data in an economical
way (Saunders et al., 2009). The given constraints also pressured us into doing a cross-
sectional research which means that we will study this phenomena at a particular time, and
not during a longer period (Saunders et al., 2009).

However, it should also be mentioned that our research strategy included some elements that
are usually related to experimental strategies. This study includes assessing the causal links
between variables, something that is very typical for experimental strategies (Saunders et al.,
2009). The element that can be perceived as experimental is the manipulation of
independent variables of the scenarios that would pose as the foundation of the survey
(Saunders et al., 2009). This was based on the fact that we wanted to exclude all other
potential factors and elements that could potentially influence the dependent variables, and
create a sort of isolated environment where only the factors related to the transparency of
the supply chain could be influencing.

28
3.4 Data collection method

3.4.1 Secondary data

Secondary data can be explained as data that has already been collected for another purpose
than of the problem we have at hand (Saunders et al., 2009; Babin & Zikmund, 2016). Since
this research is following a deductive reasoning, existing theory and literature were required
in terms of secondary marketing research. We collected secondary data from literature in
forms of books relevant to the field of study and internet as a speedy research tool in order
to access relevant articles and journals.

We looked for journals and literature in Primo which is a search engine provided by the
Jönköpings University. We also used databases such as Emerald and Springer Business and
Management journals in our search process. In our critical evaluation of the articles chosen
we made sure to look at the number of citations, the year of publish and only accept peer-
review articles in order to avoid unreliable sources and outdated information. In order to
reassure accuracy and validity of our secondary data, we did cross-checks by checking the
data from various sources (Babin & Zikmund, 2016).

3.4.2 Primary data

Naturally, we were also in need of collecting primary data in order to find empirical evidence
to be able to fulfill our purpose. To collect the data, we choose an online questionnaire. The
online analysing tool Qualtrics was selected for this, and the data collected was then
converted in to the statistical tool SPSS for further analysis. In order to collect our primary
data properly, we had to determine who to sample, what is the appropriate survey design and
how to analyse our findings. This will be further explained below.

29
3.5 Sampling

3.5.1 Non-probability sample

A sample is defined by Babin and Zikmund (2016) as a subset of some larger population that
can be measured or observed in order to understand the larger picture of the population.
There exist two sampling methods, probability sampling (i.e. representative sampling) and
non-probability (i.e. judgemental sampling) where the latter was chosen for our
research (Saunders et al., 2009).

Our study is an explanatory and quantitative research, and these are more often likely to
make sense of probability sample techniques. However, in this report convenient sampling
was used which belongs under the non-probability sample techniques (Saunders et al., 2009).
This is due to the fact that the population is unknown, and a probability sampling is therefore
impossible to conduct. Moreover, convenient sampling was appropriate to collect a larger
sample size in a opportune way after considering circumstances and factors that influenced
our execution such as time constraint and inadequate resources. Utilizing the internet was
further seen as a cost-effective procedure which lead us to the final decision of conducting
an online survey.

It can can be problematic to do a convenient sampling drawn from social networks from a
generalizability point of view (Babin & Zikmund, 2016), however, we were still allowed to
generalise although not on statistical ground (Saunders et al., 2009).

3.5.2 Sample selection

Generation y is the market segment that we found suitable to sample from in regard to our
field of study, due to their informative, selective and technology-wise characteristics in
combination with their increased awareness of apparel companies’ social responsibility
efforts (Bhaduri & Ha-Brookshire, 2011; Jin Ma et al., 2012; Hwang et al., 2014). Based on
these suitable characteristics, we believed that our sample selection could be reached via the
internet because this is where generation y’s information is mostly obtained and exchanged.
We further had access to our wide contact network online including friends and relatives that

30
were included in this particularly segment. Therefore, confidence in attaining larger response
rates from generation y as our sample selection was increased.

The collection of data was hereby conducted online where our respondents were invited to
voluntary participate and share our online survey which was distributed on social media
networks.

3.6 Survey method

3.6.1 Self-completion questionnaire

A survey strategy allows us to collect quantitative data which was later analysed by using
descriptive and inferential statistics (Saunders et al., 2009). Additionally, a survey strategy
further provided us with increased control over the research process and when we wanted
the sampling to be used (Saunders et al., 2009). Questionnaires are normally found in the
explanatory research (Saunders et al., 2009) and we decided to use a self-administrated
questionnaire mediated via the internet in an attempt to be as objective as possible and letting
the respondents finish their answers in their own pace. This in order to not infer with our
respondents and therefore minimizing the consequence of subjectively infer and cause
interviewer biases.

We send out the survey questionnaire in a well timely manner in order to make sure we had
enough time to (1) collect as many responses required (2) to have the time to conduct, analyse
and interpret our findings to demonstrate the relationships between variables.

3.6.2 Scenario design

Previous research discussed in the theoretical framework on consumer perceptions on social

responsibility and transparency have almost exclusively been done on existing brands that
consumer have previous experience with. As we in this study prefer a subjective insight in
the effects of the dimensions of supply chain transparency, we believed it to be appropriate
to construct scenarios including no other elements that could influence other than the ones

31
relevant to the study. Within these we could exclude potential influential factors such as: well
known brands that the respondents likely already have a relation to, clothing preferences,
CSR cues related to environmental factors, price points and other visual elements that might
be more or less appealing. Manipulating the independent variables like this has a experimental
element to it which was deemed to be suitable for the purpose. With a foundation in the
Supply Chain Transparency Matrix of Marshall et al. (2016) where four different types of
supply chain transparency strategies can be found, we created four scenarios that portrayed
these in regard to high versus low degree of disclosure and degree of assessment:

Figure 5: Scenarios

Each of the strategies were constructed in accordance to the five key areas of apparel industry
transparency identified by Fashion Revolution (2017) as brought up in section 2.2.4
transparency in apparel supply chains: (1) policy and commitments, (2) governance, (3)
traceability, (4) know, show and fix, and (5) spotlight issues. Most of the information included
in the scenarios were found on existing apparel companies web pages, and was therefore
grounded in real life situations. The scenarios were constructed to mirror an online purchase
situation of a plain white T-shirt without brand and price. The scenarios can be found in
Appendix 1.

32
3.6.3 Questionnaire design

An online survey method provided us as researchers with beneficial elements such as faster
responses, taking advantage of the internet distribution facilities, the ease of collecting larger
sample sizes and hindered us from spending a considerable amount of monetary resources
(Babin & Zikmund, 2016). Applying Qualtrics as the fundamental survey tool provided us
with an overall neat and appealing design of the survey, applying already existing templates
and make use of filter questions, forced answers and include pictures together with other
functionalities that eased our distribution and data analysis process.

In the beginning of the questionnaire we choose to include an overall introduction covering

what the participants can expect from participating in the survey such as the purpose of the
survey, how much estimated time the survey will take, guaranteed anonymity and how the
data collected will be handled. To further increase engagement we offered ourselves to
donate 3 SEK (0,3 Euro) for each survey completed to the United Nations Refugee Agency.
It was clearly stated that participating is voluntarily yet important for us in our attempt to
reach our purpose where we relied upon the respondents’ contribution. For those who
choose to participate, we included demographic questions to withdraw fundamental
background information of the respondents. We choose (1) gender, (2) year of birth, (3)
nationality, (4) highest achieved education, and lastly (5) occupation as the most appropriate
items to include. The demographic questions easily reassured that our sample was in line
with our purpose and excluded the ones that were not consistent with our target sample of
generation y. Additionally, we included three questions that provided us with an indication
of the respondents’ initial positions on social practices in the apparel industry (see table 1 for
all questions).

We are aware that poorly designed questions are influencing the accuracy and can lead to
misinterpretations and hence result in bias such as inadequate responses and
misunderstandings (Babin & Zikmund, 2016). We therefore avoided to include
abbreviations, leading questions, long complex questions and words with double meanings
(Babin & Zikmund, 2016). Additionally, we included definition of terms that followed on
each page throughout the survey where we described what a supply chain and social practices

33
refers to in order to make sure that respondents could easily interpret and answer the
questions. Further, for the ease of understanding of our respondents, brackets where clear
instructions such as “tick” or “scroll” were included to minimize confusion.

All scenarios were initially presented to the respondents in order for them to evaluate the
information, thereafter each scenario was presented one at the time containing the same set
of questions. In regard to the quantitative research design, we used a structured questionnaire
only containing closed-end questions to be able to measure our variables and to be able to
code these in a proper manner which is expected for the use of SPSS (Babin & Zikmund,
2016). We choose to apply 7-point likert scale ranging from 1= strongly disagree to 7=
strongly agree, as our response format. We choose 7-point likert scale since this has been
frequently applied in similar studies as ours (e.g. Hwang et al., 2014).

As our research included two previously untested variables namely, degree of assessment and
degree of disclosure. We therefore had to create these statements by carefully study closely
related studies in their formatting to be properly measured. Additional questions covering
trust, general attitude and purchase intention were withdrawn from already existing studies
that had empirical proof which assured that they are measuring what was intended to be
measured. Statements belonging to general attitude had to be combined from two different
studies. This is due to the fact that we wanted to include an item that accounted for the fact
that the question regards apparel specifically, see item G3 (table 1). Additionally, we added a
question regarding trust in the specific setting of the supply chain (TS1), however this was
not included into any measurement. A full overview of the questionnaire can be found in
appendix 2.

34
 Table 1: Instruments
Construct Coding Source

Demographics Qualtrics Library

What is your gender? Y1 (2016)
What is your year of birth? Y2
What is your country of origin? Y3
What is your current employment status? Y4

Initial attitudes to social practices in the clothing industry Authors of this study
I am knowledgeable about issues affecting workers in clothing S1 (2018)
businesses.
I am concerned with issues affecting workers in clothing S2
businesses.
I think it is easy to determine if clothing businesses are socially S3
responsible.

Degree of disclosure Authors of this study

I believe the brand provides me with a lot of information regarding D1 (2018)
the social practices in their supply chain.
I believe the brand wants to ease my purchase decision by providing D2
me with plenty of information regarding the social practices in their
supply chain.
I believe the information regarding the brand’s social practices in the D3
supply chain is complete.

Degree of assessment Authors of this study

I believe the brand has a good knowledge about the social practices A1 (2018)
in their supply chain.
I believe the brand has the ability to control the social practices in A2
their supply chain.

Trust Newell & Goldsmith

I would trust the brand. T1 (2001)
I believe that the brand is honest. T2
I believe what the brand tells me. T3

Trust in supply chain Authors of this study

I trust that the brand is making efforts to improve social practices in TS1 (2018)
their supply chain.

General Attitude Sengupta & Johar

I think the brand is a good brand. G1 (2002)
My opinion of the brand is very favorable. G2
I would feel good if I wore this brand. G3 Chaudhuri &
Holbrook (2001)

Purchase Intention Krishnamurthy &

I would likely consider the brand the next time I thought about P1 Sivaraman (2002)
buying a T-shirt.
I would likely suggest the T-shirt to a friend. P2
I would likely check reviews regarding the T-shirt. * P3

* Item was deleted due to low correlations

35
3.7 Analysis of data
We applied the analytical software of SPSS and AMOS 25.0 to be used in our analysis of
quantitative data. Our data set was firstly imported from Qualtrics to SPSS. SPSS is a
commonly used software program to perform various of data analyses (Saunders et al., 2009).
For example, descriptive, means, standard deviations and Cronbach's alpha of the
measurements were calculated in the program, and thereafter testing the assumptions for
moving along to the structural equation modelling in AMOS. The SPSS module AMOS 25.0
is a sophisticated statistical software that is used for structural equation modelling (SEM).
SEM is a multivariate statistical analysis technique and an explanatory tool usually described
as a combination of factor analysis and multiple regression analysis (Babin & Zikmund,
2016). As we included the two new dimensions of supply chain transparency, and the thesis
has an explanatory direction, SEM was seen as a vital to perform as it enabled us to test
theory and assessed its fit with reality in an accurate way. This further allowed us to evaluate
the relationships between the variables simultaneously, meaning that it calculated all of our
variables at the same time in our research model instead of treating the coefficients separately
(Babin & Zikmund, 2016). Therefore, we were able to test our hypotheses and confirm
relationships among the observed and unobserved variables in a more precise way and
moving beyond regression that only account for the variances of dependent variables (Babin
& Zikmund, 2016). AMOS provided us with fit indexes which gave us the ability to assess
the validity and reliability of the research model, an overview of how well the research model
fit the theory as well as addressing potential errors that could cause problems in the model
(Hair et al., 1998).

We followed a two-step procedure suggested by Anderson and Gerbing (1988):

(1) established a measurement model and tested it through confirmatory factor analysis
(CFA) in the measurement model, CFA functioned as a confirmatory technique driven by
theoretical relationships between our observed and unobserved variables. CFA tested how
well our measurable items represented the five constructs, and was used as a tool to confirm
or reject our measurement theory (Hair et al., 1998).

36
(2) Built a structural model to test the linear relationships among constructs in order to test
our hypotheses. The structural model represented the theory on how our constructs were
related to other constructs and further tested the causal relationships (Hair et al., 1998).

SEM became evidential in its beneficial application as it gave signals that some relationships
in the model may not have be accounted for. Therefore, we investigated new relationships
that had theoretical support and could hereby introduce a new more powerful revised model
which was better in the explanation of the purpose.

3.8 Reliability and validity

3.8.1 Reliability

Reliability is to which extent our data collection techniques and analysis procedures will yield
consistent results if it was to be repeated (Saunders et al., 2009). It is an indicator of a
measurements internal consistency and is used to represent a compound measures
homogeneity (Babin & Zikmund, 2016). In order to assure that we had a reliable scale and
internal consistency, we applied the widely used coefficient alpha (i.e. Cronbach's alpha)
which was available by using SPSS. We could assure good reliability as our scales coefficients
were exceeding 0,7 (Babin & Zikmund, 2016) See 4.2.2 Reliability for details. By using AMOS
and running SEM, we did also perform specific reliability tests such as composite reliability
and AVE (Average Variance Extracted). A more detailed explanation of this can be found
in 4.4.2.2.

We are aware of the fact that questions regarding sustainability related to ethical standpoints
often can be a subject to social desirability biases and this type of participant bias have the
potential to threaten the reliability (Bryman & Bell, 2011; Saunders et al., 2009). To
counteract this, we took measures that assured complete anonymity, and would further argue
that an online questionnaire would significantly reduce the social desirability bias as it can be
made without any involvement of the researchers.

37
3.8.2 Validity

As a good measurement should be both precise and accurate, validity is “the accuracy of a
measure or the extent to which a score truthfully represents a concept” (Babin & Zikmund,
2016, p. 281). Validity addresses weather a measure is measuring what it is supposed to
measure, and by testing the content validity, criterion validity and construct validity we were
able to assure it (Babin & Zikmund, 2016).

Content validity in this study refers to our measurement items belonging to the constructs,
and that they were actually measuring what we intended them to measure (Babin & Zikmund,
2016). Content validity was assured through the extensive literature review where we
established the questions included in the questionnaire from previous studies that have
examined similar research topics as ours. By reaching content validity, we were able to
demonstrate that our our questionnaire included questions that measured our theoretical
constructs (Babin & Zikmund, 2016). Criterion validity on the other hand, provided us with
the ability to find out and draw conclusions about if our measures could predict an outcome
or correlate with measures of similar concepts or established criteria (Babin & Zikmund,
2016). By running correlation analysis using SPSS, this enabled us to examine the various
correlations. Finally, construct validity was further investigated and is suggested to be existing
when a measure reliably measures and represents a unique concept and consists of several
components such as content validity, criterion validity and fit validity (Babin & Zikmund,
2016). Construct validity was ensured after running Confirmatory Factor Analysis in AMOS.
Additionally, since we used AMOS and performed SEM, we were able to reach convergent
validity, a more explicit explanation of this can be found in 4.4.2.2.

3.8.3 Pilot study

Before we send out the questionnaire we made use of a third party to reassure that the
questions and response alternatives were in line with an acceptable survey layout. We did two
pilot-tests on five respondents in order to receive feedback on potential problems which lead
us to the advantage of improving our questionnaire (Babin & Zikmund, 2016). The pilot-
tests also allowed us to test the reliability of the previously untested measurements for degree
of disclosure and degree of assessment, as well as general attitude that were combined by

38
different measures. It further allowed us to understand how the scenarios were perceived,
the formulation of the questions and the amount of time necessary to complete the survey.
The reason two pilot tests were made is because the structure and presentation of the
scenarios needed to be altered, and after the second pilot study there seemed to be little to
none misunderstanding of them.

3.9 Ethical considerations

Literature suggest that researchers must take ethical considerations in to account in various
of steps while performing a research project (Saunders et al., 2009). Throughout the study,
we aimed toward a subjective working approach in our critically examination of reviewing
previous studies in terms of literature review in order to demonstrate a reliable foundation
of this study.

In the stage of collecting primary data from external parties of this study, we placed extra
effort to ensure that ethical guidelines were met in concern of our respondents. By using an
online self-administrated questionnaire, we believed that the respondents would be sited in
a situation that serve for more comfort, less stress and no need to feel embarrassed by a third
part influencing their responses while conducting the survey (Saunders, et al., 2009).
Additionally, anonymity of the participations was guaranteed and clearly stated in the
beginning of the survey along with an explanation of the purpose of the study. The survey
was further voluntary to perform, and the participant had their own choice to withdraw from
the survey whenever they wanted to. Furthermore, we made sure to treat the respondent’s
responses under confidentiality and explained how their participations contribution will be
used as findings for our purpose of the thesis. Based on the set of questions included in the
demographic part, we as researchers had little or no insight on who had completed the survey
or not. Further, as we were demonstrating that a donating 0,3 Euros (3 SEK) per completed
survey will go to the United Nations Refugee Agency, this indicated that no monetary gain
was included for us as researchers (Saunders, et al., 2009).

39
4. Empirical findings and analysis
The data collected from the online questionnaire was immediately imported into the
quantitative research software tool SPSS. As a first step the responses were examined for
errors and missing values, where only the fully completed surveys were included into the
final data set. This resulted in approximately 60 responses being removed due to not finishing
the survey, and additionally three responses that were clear cases of response bias in form of
deliberate falsification as the time to complete the questionnaire were well under the required
time (1-2 minutes in contrast to the others needing a minimum of 10) and included answering
only 1 or 7 throughout the questionnaire (Babin & Zikmund, 2016). After the data had been
cleared 108 respondents remained and these are the ones that were included in the following
analysis. Firstly, demographics were covered in order to establish respondent information,
and a brief estimation of the respondent’s position on social issues in the clothing industry
were concluded. Before moving on towards further analysis, the reliability of the constructs
was determined by observing the correlation matrix and additionally by measuring
Cronbach’s Alpha. After establishing reliable constructs, the descriptive information was
analyzed including means and standard deviations in order to compare the different
scenarios. This was followed by hypotheses testing by correlation and multiple regression
analysis in SPSS, as well as testing the assumptions for SEM.

From SPSS, the data was imported into AMOS in order to test the proposed model by the
use of structural equation modelling. This type of technique allows us to examine a series of
separate multiple regression equations simultaneously (Hair et al., 1998). By following a two-
step procedure, a measurement model was first established, and after achieving an adequate
model fit, a structural model was built to further test our hypotheses (Anderson & Gerbing,
1988).

40
4.1 Sample demographics
Table 2: Sample demographics

Demographics Frequency (%)

Gender

Male 37 34,3
Female 71 65,7

Country of origin

Sweden 78 72,2
Germany 6 5,6
U.K 5 4,6
Netherlands 3 2,8
Belgium 2 1,9
France 2 1,9
Brazil 1 0,9
Denmark 1 0,9
Greece 1 0,9
Italy 1 0,9
Latvia 1 0,9
Norway 1 0,9
Philippines 1 0,9
Poland 1 0,9
Rwanda 1 0,9
Slovenia 1 0,9
Somalia 1 0,9
United States of America 1 0,9

Employment Status

Full time employment 51 47,2

Part time employment 6 5,6
Self-employed 6 5,6
Unemployed/Looking for work 3 2,8
Student 40 37
Retired 1 0,9
Other 1 0,9

Highest achieved level of education

High school degree 13 12

Some college but no degree 6 5,6
Associate degree 1 0,9
Bachelor’s degree 60 55,6
Master’s degree 28 25,9

Total Respondents 108 100

Mean year of birth 1990, Median 1991, SD of 3,9 years.

41
Our respondents included 65,7% females and 34,3% males. As the questionnaire initially
removed respondents not included in our generation y age range, all 108 respondents were
born between 1980-2000. The mean year of birth was 1990, while the median year was 1991,
and the standard deviation was 3,9 years. This implies that the age distribution was quite
representative for the study.

The majority of the respondents originated from Sweden (72,2%) where Germany (5,6%)
and United Kingdom (4,6%) had the second and third most respondents. 18 countries were
represented among the respondents. Most respondents were either full time employed
(47,2%) or students (37%) and the majority had completed a bachelor’s degree (55,6%) or
master’s (25,9%) as the highest achieved level of education.

4.2 Descriptive

4.2.1 Initial positions on social practices in the apparel industry

The survey initially covered three statements that aimed towards estimating personal beliefs
on the topic: The respondents perceived (1) knowledge and (2) concern of social issues in
the apparel industry, and additionally if they deemed it to be (3) easy to determine if apparel
companies are socially responsible. The respondents were asked to answer on a 7 point likert
scale (1= strongly disagree and 7= strongly agree).

The statement that the respondents agreed with the strongest was “I am concerned with social
issues in the clothing business “, that had the highest mean value of 5,07 and the lowest standard
deviation of 1,344. 26 respondents answered 4 or below, suggesting that 75,9% agree to some
extent that they are concerned.

The statement “I am knowledgeable about social issues in the clothing business” reached an almost as
high mean value ot 4,94 (SD 1,497), which means that the respondents overall believe that
they are somewhat knowledgeable about these issues. Only 27 out of the 108 respondents
answered 4 or below, meaning that 75% of them agreed to some extent. These two measures
imply that the the concern for social issues is slightly higher than the perceived knowledge.

42
The third statement “I think it is easy to determine if clothing businesses are socially responsible”,
resulted in the lowest mean of only 3,26, showing that they indeed found it difficult to
determine if clothing companies are socially responsible. Out of the 108 respondents only
25 answered above 4, meaning that only 23,2% of the respondents believe it to be easy to
determine if clothing businesses are socially responsible.

Table 3: Initial positions on social practices in the supply chain

S1 - Knowledge S2 - Concern S3 - Easiness of determination

Mean 4,94 5,07 3,26

Std. Deviation 1,497 1,344 1,468

On a scale 1 Strongly disagree - 7 Strongly Agree.

A correlation analysis of the variables was made in order to examine the relationships
between them. The only correlation that yielded a significant result was S1 - Knowledge and
S2 - Concern, which showed a quite large correlation (Correlations between R=0,5-1 are
considered large according to Pallant, 2005). The correlation suggests that as the perceived
knowledge regarding social issues in the supply chain increases, the concern regarding the
same does too. However, none of these variables had any significant correlation with the
ability to determine whether or not a company is socially responsible. S3 - easiness of
determination and knowledge showed a small correlation (R=0,10-0,29 is considered a small
correlation according to Pallant, 2007), with a significance of 0,116, suggesting that as the
perceived knowledge increases, respondents deem it to be easier to determine if clothing
businesses are socially responsible. However, this relationship was not statistically significant.

43
 Table 4: Correlations on initial positions

S1 - Knowledge S2 - Concern S3 - Easiness of determination

S1 - Knowledge 1

S2 - Concern 0,606** 1

S3 - Easiness of determination 0,134 -0.043 1

** Correlation is significant at the 0.01 level (2-tailed).

4.2.2 Reliability

Initially a correlation matrix was made in order to detect potential low correlations between
the constructs. All factors scored adequate in relation to each other (See appendix 3) with
the exception of item P3 “I would likely check reviews regarding the T-shirt”. This item had quite
low correlations to the other items, and was therefore tested further in order to determine
its potential exclusion. Cronbach’s Alpha was established for the constructs, all of which
reached above 0,8 which is deemed appropriate. The item P3 was evidently not reliable as it
decreased the total coefficient to 0,808, as opposed to 0,853 when only including P1 and P2
into the measurement. P3 could be perceived as a deviant item, as some respondents might
never look at online reviews at all, nevertheless if online reviews even exist for clothing. This
item could be argued to be useful when measuring purchase intention of for example
electronic devices, but not as useful regarding apparel. The P3 variable was therefore
removed from further analysis.

Table 5: Measurement reliability - Cronbach’s Alpha

Degree of Degree of Trust General Purchase

Disclosure Assessment T1, T2 & T3 Attitude Intention
D1, D2 & D3 A1 & A2 G1, G2 & G3 P1 & P2

Cronbach’s ,0,899 ,0,809 ,0,889 ,0,908 ,0,853

Alpha

44
4.2.3 Scenario comparison

After establishing reliable constructs the means and standard deviations for each scenario
were calculated. The responses from each scenario seemed to vary quite a lot at first glance,
which were in accordance to our predictions. As the generation y in particular is further
described as the conscious generation characterized with information empowerment and a
high level of awareness, we could further expect the generation y to be able to distinguish
companies that have a genuine intention of disclosing of its social practices. Therefore, we
would expect the transparency strategy of scenario 2, the transparent strategy, with high
degree of assessment/high degree of disclosure to score highest on all of the constructs.
Scenario 1, the secret strategy, that has high degree of assessment/low degree of disclosure,
and scenario 4, the distracting strategy, that has low degree of assessment/high degree of
disclosure would presumably score somewhat similar. Lastly, we would expect the generation
y to score low on scenario 3, the withheld strategy. The results from the descriptive analysis
confirmed most of these suspicions. However, there were also certain results that provides
an interesting insight into the minds of the generation y.

Figure 6: Radar chart of scenario means

45
 Table 6: Scenario comparison

Scenario Disclosure Assessment Trust General Attitude Purchase Intention

Mean SD Mean SD Mean SD Mean SD Mean SD

1 3,42 1,19 3,86 1,30 3,92 1,28 3,95 1,19 3,69 1,42

2 5,94 1,04 5,71 1,10 5,69 0,96 5,61 1,07 5,44 1,20

3 1,68 1,00 2,31 1,19 3,04 1,16 2,94 1,09 2,73 1,15

4 4,60 1,39 4,38 1,51 4,44 1,43 4,44 1,39 4,31 1,38

Scenario 1, the secret strategy in blue (high degree of assessment/low degree of disclosure),
had all mean scores between 3,42-3,95. It further had the largest standard deviation in
purchase intention out of the four scenarios, meaning that the respondent where least in
agreement of this construct. The respondents did however perceive the degree of assessment
to be higher in relation to the degree of disclosure, suggesting that the respondents to some
extent believed the company claims about being in control of the supply chain.

Scenario 2, the transparent strategy in green (high degree of assessment/high degree of

disclosure), scored the highest overall including trust, general attitude and purchase intention.
Its mean scores ranged between 5,44-5,94. This had the lowest standard deviation on degree
of assessment, trust and general attitude, which concluded that generation y seemed to agree
the most upon the fact that this brand evoked the most positive attitudes.

Scenario 3, the withheld strategy in yellow (low degree of assessment/low degree of

disclosure) scored the lowest mean scores overall (1,68-3,04), and had the lowest standard
deviation on both degree of disclosure and purchase intention. This suggests that the
respondents were in highest agreement of these constructs. However, even though this
scenario gave no information whatsoever about the social practices in the supply chain, or
the supply chain at all for that matter, yet it scored a surprising mean of 2,31 of degree of

46
assessment compared to the degree of disclosure of 1,68. Even more surprisingly the trust
scored a mean of 3,04, which is comparatively high. This suggests that even though there is
no information provided about the supply chain, the respondents appeared to believe the
company were in somewhat control, and further have somewhat trust in the company.

Scenario 4, the distracting strategy in red (low degree of assessment/high degree of

disclosure) had mean scores between 4,31-4,60. Its degree of disclosure was slightly higher
compared to the degree of assessment, much like it was intended to be perceived. However,
the scenario appeared to have a higher degree of assessment than scenario 1 even though no
information indicated that. This suggests that respondents indeed have issues with
determining if companies are socially responsible, and further that it is difficult to understand
the difference between real transparency and CSR-washing. This indicates that it is not the
actual information provided in this type of communication that affects the perceived degree
of assessment, but also the amount of related information.

4.3 Testing the assumptions for SEM

Before inputting the data and initializing the model, the data have to be tested in order to
assure that it is not violating any assumptions of multivariate data analysis. According to Hair
et al. (1998) these assumptions include normality in terms of distribution, homoscedasticity
and linearity. We further assess the outliers from a multivariate perspective, and check for
collinearity issues. The data was tested in SPSS.

A correlation analysis was made in order to briefly examine the strength and direction of any
potential linear relationship between the constructs (Pallant, 2007). By testing the Pearson
correlation, we were able to examine if there existed statistical evidence for a linear
relationship between two continuous variables where correlation is significant at the 0.01
level (2-tailed) and N=432 (108x4 scenarios). The Pearson correlation revealed only positive
and large significant correlations between all of the constructs, thus supporting H1-H6.
Correlations between 0,5-1 are considered large (Pallant, 2007). See appendix 4 for full
correlation matrix between the constructs.

47
In order to assess linearity a standard multiple regression analysis was made in order to
further explore the interrelationships between the variables in the proposed model (Pallant,
2007). Initially the assumption of sample size needed to be fulfilled before conducting a
multiple regression analysis. The sample size is recommended to be N < 50 +8m
(m=number of independent variables) which for the sake of this research would be fulfilled
as N=432 (Tabachnick & Fidell, 2001). For each regression analysis, a test for collinearity
problems was made in accordance to Hair et al. (1998). They suggest that correlation values
above 0,9 can indicate a multicollinearity problem, and further that it should be considered
to be a problem if a condition index above the threshold value of 15 (or 30) accounts for
>0,90 of the proportion of variance for two or more variables. Additionally, comparisons
with conclusions drawn from the variance inflation factor (VIF) and tolerance values should
be made, where VIF should not exceed 10, and tolerance should exceed 0,1. This test was
performed initially in all regression analyses. The normal probability plot and residuals
scatterplot were inspected for detecting potential violations of assumptions. The results
showed no signs of multicollinearity issues, the normal probability plots suggested no major
deviations from normality and the scatterplots had an adequate distribution which suggested
no violations of assumptions. The R values from the regressions showed significant variance
in the dependent variables, and the significant beta values all supported the relationships in
hypotheses H1-H6. See appendix 5 for all regression results. Additionally, we explored
linearity in the form of a curve estimation as an extra precaution where all types of possible
relationships were explored. The linear relationships were strong (the F-values), and
significant in all cases, and therefore deemed to be sufficiently linear to be tested in SEM.
See appendix 6 for curve estimation results.

Potential multivariate influential outliers were detected by testing the Cook’s distance. The
measure is a summary of the influence in all other residuals of a single case, and is considered
to be the single most representative measure of influence of overall fit (Hair et al.,1998).
Values above 1 indicates a substantial influence. The following relationships between
dependent and independent variables were examined: degree of disclosure and degree of
assessment on trust, degree of disclosure and degree of assessment on general attitude, and
trust and general attitude on purchase intention. In the latter of the proposed relationships
one outlier was found, however none were found in the others. This indicates that we only

48
have one case that has a substantial influence on the other residuals, thus no reason to take
action. The scatterplots of Cook’s distance can be found in the appendix 7.

Next, the normality was examined. Normality is the most fundamental assumption in
multivariate analysis, which refers to the shape of the data distribution for an individual
matric variable and its correspondence to the normal distribution. Therefore, the normal
probability plot, kurtosis and skewness were examined (Hair et al.,1998). The normal
probability plots suggested no major deviations from normality (appendix 5). The values for
skewness and kurtosis should preferably be within the +-1 range, but can also be accepted
within +-2 (George & Mallery, 2010). There were no indications of skewness issues, however
the items related to degree of disclosure and assessment showed somewhat devient kurtosis
values (−1,359, −1,179, −1,308, −1,204, −1,048). The kurtosis refers to the outliers of the
distribution of data and the deviant values are likely an effect from these. However, as all of
the items are measured on a 7-point likert scale and the respondents are asked about their
personal opinions and attitudes, there is really no justified way in which we could remove
potential outliers anyway. As +-2 still indicates an adequate distribution, we conclude that
distribution of the data is acceptable for conducting SEM. See appendix 8 for kurtosis and
skewness results.

As a third step the homoscedasticity was examined graphically as proposed by Hair et al.
(1998). Homoscedasticity is the assumption related primarily to the dependence relationships
between variables, and refers to dependent variables exhibiting equal levels of variance across
a range of predictor variables. Scatterplots for each relationship was performed which
indicated that the assumption is supported. See appendix 9 for homoscedasticity results.

In conclusion, we found no violations of assumptions in our data and continued to perform

SEM.

49
4.4 Structural Equation Modelling
The structural equation modelling followed Anderson and Gerbing (1988) recommended
two step procedures including (1) the construction of a measurement model and (2) the
construction of the structural model (Hair et al., 1998). Both the measurement model and
structural model was first examined for offending estimates (Hair et al., 1998): (1) Negative
error variances or insignificant error variances for any construct, (2) standardized coefficients
exceeding or very close to 1.0, or (3) very large standard errors associated with any estimated
coefficient. Following this, the model fit indexes were examined and thereafter the regression
weights and their significance.

As the results from the initial structural model will suggest, the hypothesized relationships
between the variables were not entirely correctly assessed. Therefore, a revised model was
tested as well, which provided us with valuable insights.

4.4.1 Model fit indexes

Due to the fact that SEM lacks one perfect statistical test to demonstrate the strength of a
model’s predictors, a series of goodness-of-fit measures have been developed by previous
researchers and should be combined in order to assert an appropriate model fit. We chose
to include the fit indices suggested by Hooper, Coughlan and Mullen (2008).

50
 Table 7: Model fit indexes with thresholds

Model Fit Threshold

Measurements

Chi-Square (x²) Chi-square is the minimum value of the discrepancy.

Degrees of Freedom >0 Degrees of freedom >0 indicates an overidentified model, which is
Significance level >0.05 the goal in SEM (Hair et al., 1998).
It is prefered that the significance level is above 0,05 as it indicates
that the actual and predicted input matrices are not statistically
different (Hair et al., 1998).

Relative Chi-Square 1-5 The relative Chi-square is the Chi-square divided by the degrees of
(CMIN/DF) freedom in order to make it less dependent on the sample size
(Arbuckle, 2010). In general values between 1-5 are deemed
acceptable (Hair et al., 1998)

Goodness-of-fit index >0.90 GFI represents the overall degree of fit and values above 0,9
(GFI) indicate well fitting models (Hooper et al., 2008). GFI represents
the squared residuals from prediction compared with the actual
data, not adjusted for the degrees of freedom (Hair et al., 1998).

Adjusted Goodness- >0.90 Values above 0,9 indicate well fitting models (Hair et al., 1998).
of-fix index (AGFI) AGFI is an extension of GFI where AGFI adjust GFI based upon
degrees of freedom (Hair et al., 1998).

Standardized Root < .0,08 The average of the residuals between the observed and estimated
Mean Square Residuals variances and covariances based on the standard residuals (Hooper
(SRMR) et al., 2008). The value should be as close to 0 as possible, thus
values <,08 are considered good (Hu & Bentler, 1999).

Root Mean Square <0.08 The RMSEA indicates how well the model with unknown but
Error Approximation optimally chosen parameter estimates would fit the populations
(RMSEA) covariance matrix (Byrne, 1998). RMSEA values below 0,08 are
seen as acceptable (Hair et al.,1998).

Tucker-Lewis Index > 0.95 TLI values above 0,95 is recommended (Hooper et al., 2008). TLI
(TLI) is a comparison of the proposed model and the null model (Hair et
al., 1998).

Normed Fit Index >0.95 NFI values above 0,95 is recommended (Hooper et al., 2008). NFI
(NFI) compare χ2 of the proposed model and χ2 of the null model (Hair
et al., 1998).

Comparative Fit Index >0,95 CFI values above 0,95 is recommended (Hooper et al., 2008). CFI
(CFI) is a revised form of NFI which takes in to account sample size even
when it is small.

51
4.4.2 Measurement model

The measurement model was built in AMOS by including five latent variables of degree of
disclosure, degree of assessment, trust, general attitude and purchase intention together with
their indicators. The measurement model was validated through confirmatory factor analysis
(CFA) which is a special form of factor analysis that through SEM can play a confirmatory
role instead of the conventional exploratory one. Compared to factor analysis, CFA allows
us to control the specification of indicators for each construct, and additionally test
statistically the goodness-of-fit for the proposed model (Hair et al., 1998). The CFA was
conducted through a maximum likelihood estimation. Through CFA, AMOS calculated the
standardized estimates and squared multiple correlations for every item in the measurement
model. We were hereby enabled to evaluate the fitness of our latent measurement model and
assess the reliability and validity of all of the constructs. The CFA also enabled us to allocate
potential measurements errors and to evaluate if our data fit the hypothesized measurement
model.

A check for offending estimates was conducted before assessing the measurement model fit
(Hair et al., 1998). No negative error variances were found, no standardized coefficients were
exceeding 1,0, and no very large standard errors were found (See appendix 10).

4.4.2.1 Measurement model fit

While performing CFA it is allowed to draw theoretically justified covariances between the
errors in order to obtain a better model fit. Based on the modification indices suggested by
AMOS, covariances were drawn between two errors within the same latent variable at degree
of disclosure, trust and general attitude. These covariances are all logically and theoretically
found, as they would be expected to be covaried as they are within the same construct
measuring the same thing (Arbuckle, 2010). See appendix 10 for details. Starting by drawing
the strongest covariance and then retesting the model fit a total of three times, the final model
fit was deemed adequate.

52
 Table 8: Measurement model fit

Model Fit Measurements Threshold Obtained value

Chi-Square (x²) 137,634

Degrees of Freedom >0 52
Significance level (P-value) >0.05 0,000

Relative Chi-Square (CMIN/DF ) 1-5 2,647

Goodness-of-fit index (GFI) >0.90 0,955

Adjusted Goodness-of-fix index (AGFI) >0.90 0,922

Standardized Root Mean Square Residuals (SRMR) <0,08 0,0196

Root Mean Square Error Approximation (RMSEA) <0.08 0,062

Tucker-Lewis Index (TLI) >0.95 0,976

Normed Fit Index (NFI) >0.95 0,974

Comparative Fit Index (CFI) >0,95 0,984

One of the most traditional model fit measurements is the Chi-Square P-value which should be >0,05
in order to establish insignificant differences between actual and predicted matrices and thereby test
the exact fit of the model (Hair et al., 1998; Arbuckle, 1998). As the value from the proposed model
is significant at ,000, this has potential to be problematic. However, the measurement is highly
sensitive to sample sizes and many researches claims the measurements unsuitability in many regards
(Hair et al., 1998; Jöreskog, 1969). Partly, this is because it is generally acknowledged that many
models are still very useful approximations even though they don't exactly fit the population
(Arbuckle, 1998). Many researchers further agree on the fact that the sample size has such a large
impact on the Chi-Square significance, that its often misleading and not credible, and further will only
be accepted is the sample in not too big (Arbuckle, 1998; Gulliksen & Tukey, 1958; Cochran, 1952).
The Chi-square significance of a too big sample, above 200, will nearly always be rejected (Hair et al.,
1998; Bentler & Bonnet, 1980; Jöreskog & Sörbom, 1993). Therefore, many researchers use other
measurements, such as the Relative Chi-Square (CMIN/DF) that minimizes the impact of sample
size (Hooper et al., 2008). Additionally, a RMSEA below 0,05 indicate a close fit, in comparison with

53
the Chi-Square P-value that tests the exact fit. The Relative Chi-Square is well within acceptable range,
the RMSEA accepts a close fit, and the rest of the measurements shows acceptable values. Therefore,
we deem the model fit of the measurement model to be accepted, and continue to establish validity
and reliability of the model.

Table 9: Measurement model results

CFA SMC t value Cronbachs Composite AVE

Loading Alpha Reliability

Degree of 0,899 0,8892 0,728

disclosure

D1 0,891 0,793 24,008**

D2 0,806 0,65 20,116**

D3 0,861 0,741 -

Degree of 0,809 0,8101 0,681

assessment

A1 0,845 0,713 19,974**

A2 0,805 0,648 -

Trust 0,889 0,7603 0,748

T1 0,924 0,854 20,574**

T2 0,872 0,76 20,444**

T3 0,795 0,633 -

General Attitude 0,908 0,9164 0,785

G1 0,862 0,743 25,43**

G2 0,892 0,796 -

G3 0,904 0,816 28,365**

Purchase Intention 0,853 0,8559 0,748

P1 0,867 0,752 23,481**

P2 0,863 0,745 -

P3 Deleted Deleted Deleted

Significant ** p<0,001

54
4.4.2.2 Reliability and validity of the measurement model
Internal reliability was established by determining the Cronbach Alpha coefficient which
assures internal consistency within the items in our scales. The Cronbach Alpha coefficients
exceeded 0,7 for all constructs in the measurement model and therefore demonstrated that
our measuring items were holding together in the measurement of each construct (Hair et
al., 1998). Further, composite reliability (CR) on each latent construct were in a range
between 0,76 to 0,89 and therefore exceeding the rule of CR>0,5 in order for this reliability
to be reached (Fornell & Larcker, 1981). Further, for a measurement model to be reliable the
average percentage of variation (also known as Average Variance Extracted, AVE) that can
be explained by the measurement items of the constructs should exceed or equal to 0,5. As
shown in table 9 AVE for our constructs exceeded this requirement (Sabri, 2016).

Since all the fit indexes were meeting the recommended level of acceptance, we achieved
construct validity of our measurement model (Sabri, 2016). Convergent validity was evident
as the measurement items in the model were statistically significant, every CFA loadings were
larger than 0,7 and SMC (Squared multiple correlation) for each item reached above 0,5
(Fornell & Larcker, 1981). The aforementioned AVE exceeded 0,5 which also further
functions as a verification that convergent validity was achieved (Sabri, 2016). By reaching
validity of the measurement model it can be concluded that the measurement items within it
have the ability to measure what it was supposed to measure (Sabri, 2016). See appendix 11
for calculations of AVE and CR.

4.4.3 Structural model

After the measurement model showed acceptable results, the structural model was built in
AMOS. In the structural model, we are able to test the relationships between the independent
variables (exogenous) to the dependent variables (endogenous). Degree of disclosure and
degree of assessment functioned as exogenous latent variables, and general attitude, trust and
purchase intention functioned as endogenous latent variables. In accordance to Hair et al.
(1998) an examination of offending estimates was made, and concluded that no negative
error variances were found, no standardized coefficients were exceeding 1,0, and no very
large standard errors were found (See appendix 12).

55
4.4.3.1 Structural model fit

Table 10: Structural model fit

Model Fit Measurements Threshold Obtained value

Chi-Square (x²) 145,689

Degrees of Freedom >0 54
Significance level (P-value) >0.05 0,000

Relative Chi-Square (CMIN/DF ) 1-5 2,698

Goodness-of-fit index (GFI) >0.90 0,952

Adjusted Goodness-of-fix index (AGFI) >0.90 ,0929

Standardized Root Mean Square Residuals (SRMR) <0,08 0,0204

Root Mean Square Error Approximation (RMSEA) <0.08 0,063

Tucker-Lewis Index (TLI) >0.95 0,975

Normed Fit Index (NFI) >0.95 0,973

Comparative Fit Index (CFI) >0,95 0,983

The structural model demonstrates that all specific item loadings reached above 0,7. Just as
the measurement model, the structural model fit is deemed to be acceptable, and we
therefore continued to examine the relationships between the variables by using standardized
regression weights. The results from the structural model fit is visible in table 10.

56
4.4.3.2 Hypotheses testing
All relationships significant at 0.01 level are marked with **. The degree of disclosure showed
some effect on trust (y=0,189) and general attitude (y=0,252), however none of these were
significant (p>0,01) in contrast to the results from the multiple linear regression. The
following hypotheses could therefore not be supported when testing the all of the
relationships simultaneously.

H1: High degree of disclosure will have a direct positive effect on trust towards the
brand

H2: High degree of disclosure will have a direct positive effect on general attitude
towards the brand

The degree of assessment on the other hand, showed a significant effect on trust (y=0,715**)
and general attitude (y=0,659**). This suggests that degree of assessment has the highest,
and only significant, effect on trust and general attitude compared to degree of disclosure.
Following hypotheses is therefore fully supported:

H3: High degree of assessment will have a direct positive effect on trust towards the
brand

H4: High degree of assessment will have a direct positive effect on general attitude
towards the brand

There was an indication of a slight effect of trust on purchase intention (y=0,117, p>0,01)
but as it was not significant it could not be fully supported in contrast to the results from the
multiple linear regression.

H5: Trust will have a positive direct effect on purchase intention

Lastly the relationship between general attitude and purchase intention was examined, which
showed that the former indeed has a rather large effect on the latter (0,837**). Therefore,
H6 is fully supported.

H6: Positive general attitude will have a positive direct effect on purchase intention

57
 Figure 7: Structural model results

The structural model showed insignificant results for some of the hypothesised relationships,
and thereby a critical evaluation of the model is deemed necessary. As the model fit from
both the measurement model and the structural model is good we have reason to believe
that there are relationships within the model that has not been correctly assessed. An
additional literature review was made in order to seek possible explanations for the
insignificant results. As the correlation and regression analysis supported the relationships
for H1, H2 and H5, but the results from the SEM suggests otherwise, we see the potential
for revising the theory related to these hypotheses and their constructs. The next section will
therefore cover an additional literature review and a revised model.

4.4.4 Revised model

4.4.4.1 Revised theory

An additional review of theory was conducted in order to understand if there potentially
could be other relationships between the variables. As the degree of disclosure did not yield
any significant effects on the constructs of trust and general attitude, the search was initially
focused on this construct. As evident in the results from questionnaire question K3 “I think
it is easy to determine if a company is socially responsible”, the respondents agreed upon the

58
fact that it is indeed difficult to gain an understanding for when companies are sincere with
social responsibility or not. This suggests a difficulty in distinguishing a company’s actual
efforts, insights and knowledge from what they communicate, i.e. their disclosure.

As the terms degree of disclosure and degree of assessment not yet have been examined
through a consumer perspective, following theories are derived from studies regarding
related subjects such as the general notion of CSR communication.

The data from the regression and correlation analysis suggests that supply chain transparency
in general has a positive effect on attitudes, trust and behavioral intentions, regardless if
consumers can understand whether it is sincere or not. This can potentially be explained by
the fact that claiming transparency efforts (Eisingerich & Bell, 2008) or the general
communication of CSR (Pope & Wæraas, 2016) often is perceived as enough evidence that
the company is responsible, and thereby have a positive influence on consumers.
Furthermore, providing accessible information regarding CSR in itself acts as a cue
suggesting that the company has nothing to hide, and consequently consumers perceive the
company as more trustworthy (Foscht, Lin & Eisingerich, 2018). This translates into that the
degree of disclosure alone could function as a signal that the company is responsible, and
thereby positively affecting consumers. Additionally, a high degree of disclosure can
potentially make the decision-making process easier, attributed to the fact that consumers
have cognitive limitations, and when they are confronted with a large amount of information
will not process all relevant information, and therefore make a less informed decision
(Foscht, Lin & Eisingerich, 2018). Conclusively, it is possible that as the amount of
information increases, consumers take a shortcut in their cognitive process and interprets
the amount of information into meaning that the company has a good knowledge and insight
regarding the social practices in their supply chain, i.e. the degree of assessment. This leads
us to derive the idea that disclosure has a direct effect on the perceived assessment.

Consumers find the general communication of CSR as a sufficient indicator to base their
attitudes and trust on, yet studies show that consumers are extremely sceptical about CSR
communication (Pope & Wæraas, 2016). This is indeed a contradiction to aforementioned
theories, yet as a phenomenon have been earlier acknowledged by for example Pope &

59
Wæraas (2016). Studies further suggest that consumers make use of quite particular indicators
when judging CSR messages: by their specificity, scientific plausibility and consistency
(Mayser & Zick 1993; Wagner, Lutz & Weitz, 2009), the duration of the proposed CSR
activity (Vanhamme & Grobben, 2009), if the CSR message is claimed by a third-party
organisation (Yoon, Gurhan & Schwarz, 2006) and many other quite specific elements (Pope
& Wæraas, 2016). This indicates that even though consumers place faith in CSR messages
overall, they are yet aware of the indicators providing them with evidence of the CSR efforts
and utilizes these then evaluating the message and company. In the concept of this study this
would translate into meaning that the degree of assessment is what consumers will judge
upon rather than the amount of information i.e. the degree of disclosure. Therefore, we can
presume that consumers are more likely to base their general attitude and trust on assessment
rather than disclosure.

In conclusion we propose that degree of disclosure only has a direct effect on degree of
assessment, and that only degree of assessment has a direct effect on trust and general
attitude.

4.4.4.2 Revised model

Figure 8: Revised model

The revised model (figure 8) includes following changes in comparison to the originally
proposed model: (1) Degree of disclosure have a direct effect on degree of assessment, (2)
Degree of disclosure have no direct effects on trust and general attitude, but instead an

60
indirect effect mediated through degree of assessment. The model was built and tested in
AMOS in the same way as the original.

Before assessing the model fit an examination of offending estimates was made, and
concluded that no negative error variances were found, no standardized coefficients were
exceeding 1,0, and no very large standard errors were found (see appendix 13).

4.4.4.3 Revised model fit

Table 11: Revised structural model fit

Model Fit Measurements Threshold Obtained value

Chi-Square (x²) 177,833

Degrees of Freedom >0 57
Significance level (P-value) >0.05 0,000

Relative Chi-Square (CMIN/DF ) 1-5 3,12

Goodness-of-fit index (GFI) >0.90 0,942

Adjusted Goodness-of-fix index (AGFI) >0.90 0,907

Standardized Root Mean Square Residuals (SRMR) <0,08 ,0253

Root Mean Square Error Approximation (RMSEA) <0.08 0,07

Tucker-Lewis Index (TLI) >0.95 0,969

Normed Fit Index (NFI) >0.95 0,967

Comparative Fit Index (CFI) >0,95 0,977

The revised model was built and tested for fit indices following the same procedure as the
initial model in order to confirm an acceptable fit. Similar to the original structural model,
the revised model has acceptable fit indexes, and is therefore used for further testing.

61
4.4.4.4 Testing the revised model

As model fit was deemed acceptable, the relationships within the revised model was tested.
In addition to regression weights and its related significance, we further tested for squared
multiple correlations (R²) which indicates the percentage on total variation of Y explained by
its independent variable (Hair et al., 1998).

Figure 9: Structural revised model results

The revised model as shown in figure 9, still demonstrates that the specific item loadings
reached above 0,7. Further, all relationships were significant at 0.01 level (**), except one
being significant at 0.05 (*). We theorized that degree of disclosure would have a direct effect
on degree of assessment, and this relationship was concluded to be correctly assessed as the
relationship was very strong (y=0,927*) and significant. This means that for every standard
deviation (SD) degree of disclosure increases, degree of assessment increase with 0,927 SD.
The squared multiple correlation for degree of assessment suggested that 85,9% of its
variance is accounted for by degree of disclosure, suggesting that it had a large impact.

We further theorised that degree of assessment was the only independent variable affecting
trust and general attitude, which relates to previous hypothesis H3: high degree of assessment
will have a direct positive effect on trust towards the company, and H4: high degree of

62
assessment will have a direct positive effect on general attitude towards the company. These
hypotheses were earlier accepted in the original model, and the results from the revised
model indicates the same thing (y=0,944**, y=0,947**). The effect suggests that when degree
of assessment increases 1 SD, trust increases by 0,944 SD, and general attitude by 0,947 SD.
This concludes that in the transparency context, degree of disclosure affects the perceived
degree of assessment which in turn has an influence on trust and general attitude. The degree
of disclosure and degree of assessment are therefore not covaried, but instead the latter is
dependent on the former. The SMC further suggests that degree of assessment explains
89,7% of the variance in general attitude, and 89,1% of trust.

As for the relationship between trust and purchase intention, H5: Trust will have a positive
direct effect on Purchase Intention, the original model yielded no significant effects. In the
revised model however, the relationship was quite weak, but significant at a 0,05 level
(y=0,175, p=0,038). This concludes that by every SD trust increases, purchase intention
increases by 0,175 SD. The reason for why this relationship was not significant before is
likely because of this construct in comparison to the model being slightly unstable, and would
probably have yielded a stronger effect with a larger sample. This in combinations with all
previous theory being in unison of the influence of trust on purchase intention, we therefore
chose to accept H5.

The relationship between general attitude and purchase intention relates to H6: Positive
general attitude will have a positive direct effect on purchase intention, and was in the original
model deemed acceptable. In the revised model, it is indeed concluded that general attitude
has a quite strong effect on purchase intention (Y=0,783*), meaning that by every SD general
attitude increases, purchase intention increases by 0,783 SD. This means that within the
sample, the general attitude had a much larger effect on purchase intention than trust. The
total variance of purchase intention explained by trust and general attitude is 88,8%. The
results from the revised model indicates that it is a suitable model to explain the relationships
between the constructs.

63
4.5 Hypotheses conclusion

Table 12: Hypotheses conclusion

Hypotheses

H1 High degree of disclosure will have a direct positive effect on trust towards the Rejected
brand.

H2 High degree of disclosure will have a direct positive effect on general attitude Rejected
towards the brand.

H3 High degree of assessment will have a direct positive effect on trust towards the Accepted
brand.

H4 High degree of assessment will have a direct positive effect on general attitude Accepted
towards the brand.

H5 Trust will have a positive direct effect on purchase intention Accepted

H6 Positive general attitude will have a positive direct effect on purchase intention. Accepted

64
5. Discussion

5.1 Initial positions on social practices

The respondents showed that they were indeed concerned with social issues in the supply
chain with a mean score of 5,07 and 75,9% of them agreed to be concerned to some extent.
The perceived knowledge about set issues was almost equally high with a mean of 4,94, and
75% agreeing to be knowledgeable some extent. These relatively high means was expected
as the generation y is considered to be a conscious generation characterized by information
empowerment (Bhaduri & Ha-Brookshire, 2011). However, it could be argued that as they
are a generation so aware, the mean scores should be even higher. The mean scores are for
example closer to the neutral alternative 4, then to 7 strongly agree. Most worrisome is the
fact that the other part of the respondents, 24,1% were not concerned with social issues in
the supply chain, and 25% believed themselves not to be knowledgeable about set issues.
For the conscious generation y, this is a bit surprising to say the least, as we can assume these
25% without knowledge still consumes clothing. However, studies by for example Hill and
Lee (2012) similarly suggested that there is an absence of holistic assessment of generation
y’s knowledge of CSR in the apparel industry, and another study by Hiller Connell and Kozar
(2009) suggested that only 62% of their american college respondents were knowledgeable
about social issues in foreign clothing manufacturing. Additionally, the latter study also
resulted in 83% of the respondents being concerned about the same issue.

The correlation analysis confirmed the perceived knowledge to be related to the concern,
which confirmed Dickson’s (2000) statement that a greater knowledge regarding social issues
most likely will lead to an increased concern among consumers and thereof impact their
attitude.

The data from the statement regarding the perceived easiness of determination however
provided us with quite clear evidence of the state of social responsibility in apparel supply
chains. The mean score was 3,26 and only 23,2% agreed to some extent that thought it was
easy to determine if clothing businesses are socially responsible. That the majority of the
respondents were concerned and knowledgeable about the related issues, but still perceived

65
it to be difficult to determine even more explains how apparel companies today are able to
deceive stakeholders into believing they are socially responsible. This supports the findings
by Gwilt and Rissanen (2011) which suggested that even though increasing awareness of
ethical issues, consumers are still often confused about CSR related issues. This could be due
to the fact that the different CSR relates cues within the apparel industry are many (eco,
organic, fair trade and recycled for example), and therefore it perceived as very complex
(Thomas, 2008). As CSR-washing being a prevailing phenomenon in the fashion industry
this difficulty to determine which companies are honest and sincere in their CSR efforts has
further led to an increased scepticism, not only among generation y (Bhaduri & Ha-
Brookshire, 2011), but also in general towards CSR-communication (Pope & Wæraas, 2016).

5.2 Scenario comparison

Much of the literature included in the theoretical framework suggested that transparency of
social practices in general would increase trust, influence the general attitude and indirectly
the purchase intention. For example, Lundblad and Davies (2016) and Dickson (2001) found
that consumers purchasing decisions are increasingly affected by ethical concerns, and
thereby the transparency of claims of acting ethically. For the case of generation y, Bhaduri
and Ha-Brookshire (2011) suggested that they are particularly sceptical towards companies
that claim to be concerned about environmental and social issues, and as such we could
expect that the trust and general attitude, would decrease as the level of transparency did,
and further that they would be able to differentiate between genuine and disingenuous supply
chain transparency such as CSR-washing strategies.

Two of the scenarios were evidently easier to assess than the others, as Scenario 2, the
transparent strategy, and Scenario 3, the withheld strategy, had the lowest standard deviations
on all constructs. This was however very expected as they both had clear high/high and
low/low on degree of assessment and disclosure. The other two scenarios on the other hand,
Scenario 1, the secret strategy, and Scenario 4, the distracting strategy, that reflected high/low
and low/high degree of assessment and degree of disclosure, had the highest standard
deviations indicating that the respondents agreed less on these. This we believe to reflect the
conflict between relying on a brand that claims to be in control and abiding fair practices

66
without explicitly disclosing the full information (scenario 1), and relying on a brand
disclosing a lot of meaningless vaguely related information (scenario 4). The higher standard
deviations can for example be a result of some people being able to assess what information
is substantial, and what is distracting. Previous discussion concluded the difficulties for
consumers to correctly assess if businesses CSR claims in general. In the light of the scenario
comparison, this conclusion can be further augmented as researchers such as Peattie and
Crane (2005) claim that this confusion further can cause misconceptions about the about the
CSR activities in the apparel industry which is a likely explanation for this result.

Scenario 4, the distracting strategy, was based on classical CSR-washing claims and we thus
expected the respondents of generation y to be sceptical enough to recognize this which they
did in comparison to Scenario 2, the transparent company. This means that even if they are
able to separate sincere social responsibility from disingenuous CSR-washing, they still
indirectly base their attitude, trust and behaviour on how much information the company
share, and not just that they claim to be in control and have insight in their supply chain as
in scenario 1, the withheld strategy. This have so far in the discussion been attributed to that
it is difficult for consumers to determine what is substantial information regarding the supply
chain and what is not. Another potential explanation is provided by Pope and Wæraas (2016)
that suggests that consumers usually do not inspect CSR performance even when it is
accessible to them. For example, a survey by Penn Schoen Berland (2010) showed that 75%
of the respondents claimed to be more likely to perform a purchase after reading CSR related
information on a corporate website, but only 13% had sought such information. Instead,
consumers merely interpret the CSR related messages as cues for the company acting
responsible (Pope & Wæraas, 2016). That consumers in general do not take advantage of
extensive CSR-related information could therefore suggests that they might not engage with
the content related to supply chain transparency, and thereby rather passively judge the
information based on the amount of information, i.e. the degree of disclosure. This naturally
diminish their opportunity to make informed decisions on such matters, and additionally
provides an explanation for why CSR-washing is possible.

67
5.3 Proposed model

5.3.1 The relationship between degree of assessment and degree of

disclosure

In the initial model development, we aimed for testing the different dimensions of supply
chain transparency as covaried and affecting the other variables as separate constructs, so
that we could assess their impact on one another, and on other variables. The dimensions
originate from managerial research on how to establish a suitable transparency strategy for
supply chain disclosure (Marshall et al., 2016), and it is therefore of interest also to examine
the influence of these on consumer behaviour. To the best of our knowledge, this way of
dividing transparency into dimensions have not been examined from a consumer perspective
before, therefore we had reason for examining and hypothesizing these dimensions as being
separate and equal to each other. However, as the data was analyzed it became evident that
the relationship between the variables were not as initially proposed. This was evident as
none of the relationships related to the degree of disclosure was significant in the SEM,
which was further augmented when assessing the results from question S3, concluding that
only 23,2% of the respondents believed it to some extent being easy to determine whether
clothing companies are socially responsible or not.

The revised theory section provided us with evidence that supported a different relationship
between disclosure and assessment, namely that the more information that is disclosed,
regardless of its quality, affects the perceived degree of assessment. This follows a logical trail
of thought, as well as underlining the difficulties of determining CSR efforts as genuine or
not. This could however also be an effect of not only difficulties and confusion, but also a
result of laziness meaning that people do not actually care enough to read the information
or simply judge the social responsibility of the company solely based on the CSR cue in itself.
Pope and Wæraas (2016) for example suggested that although consumers have access to CSR
information, many of them are not utilizing it to make informed decisions. Instead, CSR
communication in general is valued in itself, regardless of the CSR performance or the quality
of the information disclosed. As a great contradiction, it seems that although the amount of
information displayed plays an important role in assessing companies CSR efforts,

68
consumers do not base their evaluation mainly on this dimension. Instead, our research
showed that trust, attitudes and behavior intention was based on the perceived degree of
assessment of the company. Previous research on general communication of CSR efforts
similarly suggest that consumers base their judgements of CSR messages on specific
indicators such as scientific plausibility, consistency, the duration of the proposed CSR
activity and third-party involvement (Mayser & Zick 1993; Wagner et al., 2009; Vanhamme
& Grobben 2009; Yoon et al., 2006). This contradiction might seem circumstantial, but have
been acknowledged by Pope and Wæraas (2016) in the concept of CSR communication in
general. We attribute this to the fact that generation y is sceptical and aware of CSR-washing
techniques, and therefore have some sense of what kind of evidence that suggests this.

The contradiction of the dimensions of transparency could be an indicator that generation y

has a very good conceptual knowledge of CSR efforts, its communication and CSR-washing
strategies, but fail to understand it in practice, either for reasons of difficulties of
interpretation and/or of laziness. In conclusion, we believe the relationships established in
this study to be credible and well-founded in our data as well as through studies on similar
concepts.

5.3.2 Transparency as predictor of trust

It is evident from the results of regression and correlation analysis that supply chain
transparency in general has a positive effect on generation y trust which has previously been
suggested by for example Kang and Hustvedt (2014). They reasoned that transparency will
yield favourable characteristics of trust in the eye of the consumers as it invites consumers
to access information to base their evaluations on.

The initial structural model in SEM demonstrated an insignificant relationship between

degree of disclosure on trust and a significant relationship between degree of assessment on
trust. Due to the introduction of the revised model we argue that, as the amount of
information regarding assessment increases, generation y is more likely to perceive the
company as more knowledgeable and in control over their supply chain activities and
therefore an increase in trust will follow.

69
Since the amount of information disclosed is not by itself a good predictor of trust, this can
be a result of many different things. For example, even though generation y favour the
accessibility of information as suggested by for example Jin Ma et al. (2012) and Hwang et
al. (2014), signals of generation y sceptics toward apparel companies CSR efforts becomes
evidential when 25% expressed their belief that it is indeed difficult to judge when company’s
efforts are truly sincere. This indicates that generation y is more likely to look for certain
elements and at the quality of the information that will affect trust rather than the amount
of information itself. Our findings somewhat correspond with Doney and Cannon (1997)
study where they suggested that a company’s competence is one motivational factor that will
influence consumers trust.

During our scenario comparison where generation y’s trust yielded the highest mean scores
on scenario 2, the transparent strategy and scenario 4, the distracting strategy, both
containing high disclosure of information although scenario 4 included clear inferences of
CSR-washing. This would most likely have given a lower mean score of trust if this had been
perfectly perceived and therefore their expressed concern that it is difficult to judge the
company's sincerity further becomes evident. Additionally, as generation y previously
signalled that they are somewhat knowledgeable about social issues, we argue that it may be
the fact that a lack of knowledge could exist within the segment which makes it difficult to
evaluate what is considered “good” in apparel companies voluntary CSR efforts, and what is
sincere. In this line of arguing and since we included a range of different CSR related efforts
in the scenarios, this may have caused some confusion among generation y to know what
criteria to judge upon. Therefore, we argue that the amount of information could have
allowed generation y to take a favourable shortcut in their interpretation and perceived the
high disclosure scenarios as more trustworthy. This argument corresponds well with Foscht,
Lin & Eisingerich’s (2018) study, who specifically outlined transparency to act as a cue for
consumers where they perceive a transparent company as more trustworthy.

Further, it could also be the case that some were able to see through the amount of
information looking at certain motivational elements included in the scenarios that
influenced trust. However, as we did not specify any measurable CSR efforts, it becomes

70
difficult to determine if there were any particular information that influenced the trust more
than others. The question of specifying what builds trust was also discusses by Kang and
Hustvedt (2014), and is further supported by previous studies such as Hill and Lee (2012)
who underlined the lack of assessing generation y holistic preferences of CSR attributes.

Summarizing, Middlemiss (2003), Hillenbrand et al. (2013) and Kang and Hustvedt (2014)
suggested that perceived social responsibility and the transparency of it will impact
consumers trust toward a company. The research conducted in this study supports their
findings, however as we divide the notion of transparency into degree of assessment and
disclosure our research brings another dimension into the concept. We can conclude that
transparency is in fact a predictor of trust, although this is to great extent based on the
perceived assessment which in turn is affected by the degree of disclosure.

5.3.3 Transparency as predictor of general attitude

In the theoretical framework Dickson (1999; 2000; 2001), Shim (1995), Kang and Hustvedt
(2014), Diddi and Niehm, (2017) and Hillenbrand et al. (2013) proposed a positive
relationship between perceived responsibility and the transparency on consumers attitudes.
We support their reasoning and argue that generation y in general is in favour of transparency
of CSR and that the positive relationship exists and set the foundation to form positive
attitudes. This argument especially became evident in our regression and correlation analysis.

The initial structural model in SEM was demonstrating an insignificant relationship between
the degree of disclosure on general attitude while it was suggesting for a positive significant
relationship between degree of assessment on consumers general attitudes. A positive
significant relationship between the latter relationship was also the case in the revised
structural model. Based on the revised structural model’s result, we argue that when the
amount of information regarding assessment increases, generation y is more likely to perceive
the company as more knowledgeable and in control of their supply chain activities and
therefore an increase in general attitude will follow.

71
Ajzen (1991) and Keller (1993) outlined the complexity of attitudes as consumer can have a
unified positive attitude toward the phenomena of transparency and CSR communication,
however the formation of these attitudes can be based on different reasons. With a
foundation in the revised structural model, it becomes plausible that generation y is looking
for specific criterions that will affect their attitudes as the amount of degree of assessment
had a direct effect on general attitude. The difficulty to outline exactly what will favour
generation y attitudes further corresponds with Hill and Lee (2012) who demonstrated the
limited understanding of generation y’s attitudes of CSR attributes. By taken the
characteristics that symbolized generation y in to account, generation y can be argued to
more actively search for action oriented elements that minimize specific issues of social
impact. This was for instance explained by Dickson (2001) who suggested that consumers
have more favourable attitudes toward activities that will prevent child labour. Lundblad and
Davies, (2016) also proposed more action oriented activities as they found that activities that
demonstrated improvements of insufficient working places of labour will generate positive
attitudes among consumers.

Since generation y is favouring CSR communication in general, exactly what was causing
positive attitudes in our study is difficult to pinpoint as we included various of CSR efforts
of social practices in the scenarios in the questionnaire. Based on the scenario comparison,
it was however clear that generation y favoured scenario 2, the transparent strategy followed
by scenario 4, the distracted strategy where both scenarios had the highest degree of
disclosure. Scenario 4 which included clear CSR-washing elements had high mean scores,
and on this note Pope & Wæraas (2016) suggested that its likely that some may view a
company as responsible after only looking at the amount of information which could be one
reason in our case. We further find an additional explanation in Foscht, Lin & Eisingerich
(2018) study who suggested that accessible information regarding CSR in itself acts as a cue
suggesting that the company has nothing to hide. Therefore, some consumers of generation
y may interpret these types of apparel companies CSR communication as more favourable
as they can easily be perceived as more in control of their CSR efforts based on the amount
of assessment they choose to disclose.

72
In sum, it becomes evidential that transparency as a predictor of general attitude follows
similar lines of argument as in the prediction of trust. We therefore argue that transparency
is a predictor of general attitude when it finds its arbitration through the perceived degree of
assessment that is affected by the degree of disclosure.

73
5.3.4 Proposed model

As many times already discussed, previous research on transparency concluded that it as a

general notion related to CSR should have positive effects on consumer perceptions,
intentions and attitudes. We initially proposed a model including these variables as being
dependent on the two dimensions of transparency: degree of assessment and degree of
disclosure. This model was supposed to reflect the relationships between the variables as
perceived by the theoretical framework, and predicted the dimensions of transparency to be
equal as they together would capture the full notion of transparency. Much like the proposed
model by Kang and Hustvedt (2014) transparency affected trust and general attitude, and
mediated through these also the purchase intention.

As evident in our research the proposed relationship between the degree of disclosure and
degree of assessment was slightly different than originally suggested. Instead, degree of
disclosure affects degree of assessment which acts as a single predictor of trust, general
attitude and purchase intention. At this point we have presented possible reasons and
support of this relationship, thus perceive it to be a valid and logical theory for how
transparency of social practices in the supply chain affects intentions, attitudes and trust.
Therefore, we propose the following model of how supply chain transparency of social
practices influences consumers:

Figure 10: Proposed model of how supply chain transparency of social practices influence
consumers.

74
6. Conclusion

6.1 Purpose and research questions

The purpose of this study was to determine how disclosure of CSR-efforts related to social
practices in terms of the two dimensions of supply chain transparency, degree of disclosure
and degree of assessment, affect generation y’s perceptions and intentions in the context of
the apparel consumption. We further aimed towards theory development in terms of
establishing a model explaining how these dimensions predicts outcomes in terms of attitude,
trust and intention. In order to fulfill the purpose two research questions were developed:

RQ1. What is the relationship the dimensions of supply chain transparency, degree
of disclosure and degree of assessment, in the minds of generation y?

Our study concludes that both degree of assessment as well as the degree of disclosure plays
vital parts in the consumer evaluation of supply chain transparency and the social
responsibility within it. The dimensions were originally theorized to both have a direct effect
towards the other variables of trust and general attitude, but the data instead revealed that
the degree of disclosure had a mediating effect through degree of assessment. Generation y
perceives a high degree of disclosure as a cue that the company is acting socially responsible
and therefore perceive it as the company has a good control and knowledge of their supply
chains, regardless of the quality of such information. Degree of assessment on the other hand
seems to be more difficult to evaluate, which further provides an explanation for why CSR-
washing is possible today.

RQ2. How do the dimensions of supply chain transparency, degree of disclosure and
degree of assessment, affect the intention, attitudes and trust towards a brand among
generation y in an apparel context?
This study suggest that generation y’s perception of brands transparency efforts plays an
important role in building trust, general attitude and indirect the purchase intention. The
analysis of the data further evidenced that the dimensions of supply chain transparency do
not have a completely equal relationship to the other variables of trust, general attitude and

75
purchase intention. Instead, the degree of disclosure, i.e. how much information the
company discloses, directly affects the perceived degree of assessment, i.e. the perceived
control and insight of the supply chain that the company possess. A high degree of disclosure
is not in itself influential on trust, general attitudes and purchase intention, but is instead
mediated through degree of assessment. Generation y is more inclined to base their trust,
general attitude and purchase intention towards a company on the perceived degree of
assessment, hence how much control and knowledge the company has of their supply chain.
As a generation known to be aware yet sceptical of CSR efforts this research concludes their
role as a conscious generation.

6.2 Implications

6.2.1 Theoretical implications

First and foremost, this study contributes to theory by extending the traditional Theory of
Reasoned Action (Fishbein & Ajzen, 1975) into including the dimensions of supply chain
transparency, trust and general attitude as predictors of purchase intention. The developed
model provides theory with a valid explanation of how the dimensions of supply chain
transparency interacts, and affect the consumers of generation y.

An abundance of research has been made on the effects of CSR efforts on consumer behavior (E.g.
Marin et al., 2009; Groza et al., 2011; Diddi & Niehm, 2017; Dillman, 2000), and some on
the more specific notion of business transparency affecting consumer perceptions of
companies (Carter & Rogers, 2008; Dubbink et al., 2008; Augustine, 2012). This study
confirms these theories, and further contributes to these with establishing the importance of
transparency of CSR efforts particularly regarding disclosing information about the social
practices in the supply chain.

Most of the research on transparency concludes it as being a tool for companies to present
their CSR claims to stakeholders in a “best practice” manner in order to abide the demands
from stakeholders (Marshall et al., 2015, Reynolds & Yuthas, 2008; Mol, 2015; Doorey,
2011). The dimensions of supply chain transparency were for example originally proposed

76
to function as conceptualizations of supply chain transparency strategies (Marshall et al.,
2015). To the best of our knowledge, no research has been made on how these strategies are
perceived by consumers, let alone how the dimensions of supply chain transparency are
perceived. This research is therefore the first of its kind to apply theory from business
management about supply chain transparency strategies on theories of consumer behavior
in order to understand the practical implications in terms consumer perceptions. This
provides further depth into the field of CSR-transparency, in particular how supply chain
transparency of social practices influences consumers of generation y.

The study can further serve as a stepping stone in research on generation y as it concludes
that they are susceptive to large amounts of such information, but mainly base their attitudes,
trust and behavior on the perceived company assessment of the supply chain. Furthermore,
this provides an insight to why successible CSR-washing is possible today, as it explains how
the dimensions of supply chain transparency relates to each other and is processed by
generation y.

6.2.2 Managerial implications

The greatest managerial implication of this research is that it improves the understanding of
the behavioral intention of generation y in the context of socially sustainable consumption.
The research concludes that there in fact is potential for companies to persuade or influence
the consumers of generation y with an effective supply chain transparency strategy. Many
TRA-based models include variables that is very difficult for a brand to influence, such as
perceived behavioural control as suggested by Ajzen (1985). Our proposed model however
suggests that companies can directly affect the consumers behavioral intentions by utilizing
the dimensions of supply chain transparency in an appropriate way.

For apparel marketers who aims at communicating their CSR efforts within the supply chain
to generation y, there are several conclusions from this study to account for. First the
majority (75,9%) of the respondents are concerned with social issues in the clothing business,
additionally the majority (75%) believed themselves to be knowledgeable to a certain extent
about such issues. Although some studies suggest that this does not correspond with their

77
actual behavior at a later stage, it still means that there really is no way for brands to
completely avoid addressing such elements (Carrigan & Attalla, 2001). Next, only 23,2% of
the respondents believed it to be easy to determine whether or not clothing businesses are
socially responsible. As a marketer, one therefore has to facilitate the transfer of information
regarding social responsibility in the supply chain into a way that it is clear and accessible for
the consumers to interpret it. The study further concludes that generation y bases their
behavior on the perceived control, knowledge and insight the company seem to have of their
supply chain, even though this is enhanced by a large amount of information. The
communicated supply chain transparency should therefore contain specific evidence and
substantial information in order to maximize the trust, attitudes and intention of generation
y. The amount of information that is disclosed further augments the perceived assessment
of the company, however this should not be utilized as a tool in itself as such strategy falls
under the CSR-washing category and thereby risks being scrutinized by the sceptical
generation y, or by organisations such as Fashion Revolution (2018) or Transparency
International (2018).

Lastly, without being too cynical, most of the companies who not yet have a transparent
supply chain probably do not disclose such information as it would be perceived as
inadequate or because they do not have enough insight in it. It can therefore be difficult for
companies to discuss improving social practices and labour conditions, as it admits that there
is something that needs to be improved. This research can therefore serve as a stepping stone
for these companies, as it reassures them that a step towards supply chain transparency even
when admitting issues within it will benefit their relationship with the consumer.

6.2.3 Societal implications

Although this research provides theoretical and managerial implications, in this case it is the
societal implications that should be made a focal point. The main reason for studying how
transparency affects consumers should not be how to increase sales towards generation y,
but instead how the people directly affected by the social practices in the supply chain could
benefit from it. Transparency regarding social practices down the supply chain is impacting
the society at large since making information regarding the supply chain public helps

78
stakeholders, non-governmental organizations, unions, communities and the workers
themselves to reinforce and uphold ethical standards (Fashion Revolution, 2017).

As the study concluded only positive outcomes for the brands that indicated high scores on
both of the dimensions of supply chain transparency, marketers and decision makers who
aims for a high level of transparency also will have to address their assessment of the supply
chain in order to appear credible. The implications solely from this fact means that brands
will have to take substantial action towards improving and maintaining the social practices in
the supply chain, possibly improving the lives and working conditions for thousands of
people.

Consequently, an increased supply chain transparency allows consumers to make informed

decisions about their apparel consumption. Instead of letting consumers guess if a company
is acting favourable or not in toward labour exploitation, it is more efficient if brands choose
or are obligated to engage in acceptable and common third-party standards and independent
auditors, such as SA8000, the leading social certification standard for factories and
organizations (Social Accountability International, 2018), and Fairtrade certification
(Fairtrade, 2018) to assure human welfare.

6.3 Future research

As the notion of transparency has increased in its application and importance during the last
decade, the topical field of transparency could therefore be argued to just have started its
journey and that it is likely that more and more apparel companies are willing to adopt some
sort of a transparency strategy. Nevertheless, transparency in the context of CSR has the
potential of generating positive characteristics and stimulate the consumer-company
relationships in a favourable way. Therefore, additional research in this topical field is to be
recommended in order to contribute to the limited amount of research and suggest for
further improvements of apparel companies.

Dimensions of supply chain transparency: We propose a research model and introduced

two new dimension of supply chain transparency in the already existing attitude and
behavioural model of TRA. It is foremost advisable for future researchers to use already

79
pursued consumer behaviour models such as the TRA if future research wish to take place
in a similar approach as ours. Additionally, if the two dimensions of degree of disclosure and
degree of assessment are going to be exclusively examined, we stress the importance to
acknowledge the mediating role of degree of disclosure through degree of assessment instead
of treating the dimensions as equal constructs.

Degree of Assessment: Since the degree of assessment was the only construct that could
solely support a direct effect on generation y’s trust and general attitude, we argue that
generation y is seemingly looking for certain criteria to build their evaluation on, where
transparency plays a part of it. Therefore, we suggest that a more thoroughly research should
be performed to map potential criteria that generation y search for in the aspect of
transparent social supply chain efforts.

Sample Selection: We choose generation y based on their characteristics of being a segment

that is more aware and informed, and therefore we anticipated this segment to be knowledge
regarding social issues in the apparel industry. However, a suitable sample selection belonging
specifically to ethical and sustainable sample could be recommended for future research in
order to more accurately describe the effects of transparent CSR communication strategies
on this particular sample.

Examination between groups: In this thesis, respondents from all over the world were
included to participates in our online survey as to reflect how generation y is affected by the
two dimensions of supply chain transparency. Therefore, we suggest for future researchers
to examine the potential effect between gender differences, educational backgrounds or if
the country of origin etc. may play a considerable role in the perception of transparent
communication strategies. Additionally, in order to more certainly draw conclusions
regarding generation y, a comparative study should be made on other generations.

80
6.4 Limitations

We limited the primary data collection to be in the form of an online self-administrated

questionnaire. Hereby we made sure that our own self reflections of how the respondents
would perceive the proposed scenarios were limited and the potential of interviewer bias was
non-existing. It is plausible that allowing for all nationalities to participate, a unified sample
selection with the same or similar geographical background could have generated more
accurate understandings. However, due to time-constraints we were not able to reach a larger
sample. If we would have had more time, this could have had an additional impact of our
results with the possibility of only including one or more country of origin to be examined.

The degree of disclosure and degree of assessment has never before been empirically
examined from a consumer perspective to our knowledge. Therefore, the lack of supportive
research in order to find evidentially proven measurement items that would be perfectly
applicable for these dimensions were difficult. Main limitation is therefore our use of
secondary data in the construction of these measurements items. However, by the use of
pilot studies, studying as closely related studies as ours with belonging constructs, and based
on the results provided from running SEM in AMOS, it did not however prevent us from
using these measurements in our study.

As explained in the methodology chapter, in the construction of the questionnaire we made

sure to mainly include real life cases in the setup of our scenarios. However, we only choose
to include elements that were positively perceived, such as bar codes, list of fabrics and
accessibility of CSR reports. As transparency refers to disclosure of information which means
to act with an open approach including the disclosure of both positive and sensitive business
practices, we did not choose to include negative elements in this study which most likely
could have provided us with extensive findings. However, this was not in the scoop of this
study.

We further acknowledge the existence of a potential behaviour-intention gap in our study.

Therefore, we argue that it is somewhat difficult to perfectly outline what portrays consumers
in this study since some apparel consumers belonging to generation y may wish to be seen

81
as sustainable, but in real life, the desire of being fashionable is likely to be outweighed by
other factors (McNeill, & Moore, 2015) such as price, trends and brand image (Solomon et
al., 2010).

82
References
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision
Processes, 50(2), 179-211.

Ajzen, I. (2011). The theory of planned behaviour: Reactions and reflections. Psychology &
Health, 26(9), 1113-1127.

Ajzen, I. & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood
Cliffs, N.J.: Prentice-Hall.

Ajzen, I., & Madden, T. J. (1986). Prediction of goal directed behavior: Attitudes,
intentions, and perceived behavioral control. Journal of Experimental Social Psychology, 22, 453-
474.

Anderson, J., & Gerbing, D. (1988). Structural Equation Modeling in Practice: A Review
and Recommended Two-Step Approach. Psychological Bulletin, 103(3), 411-423.

Aras, G., & Crowther, D.(2011). Governance and social responsibility: International perspectives.
New York: Palgrave Macmillan.

Arbuckle, J. L. (2010). IBM SPSS® Amos™ 19 User’s Guide. Chicago, Illinois: SPSS Inc.

Augustine, D. (2012). Good Practice in Corporate Governance. Business & Society, 51(4),
659-676.

Awaysheh, A. & Klassen, R.D. (2010). The impact of supply chain structure on the use of
supplier socially responsible practices. International Journal of Operations & Production
Management, 30(12), 1246-1268,

Babin, B., & Zikmund, W. (2016). Exploring marketing research (11th ed.). Boston:
Cengage Learnin

Ball, C. (2009). What Is Transparency? Public Integrity, 11(4), 293-308.

Barnes, L., & Lea-Greenwood, G. (2006). Fast fashioning the supply chain: Shaping the
research agenda. Journal of Fashion Marketing and Management: An International Journal, 10(3),
259-271.

Bentler, P.M. & Bonnet, D.C. (1980). Significance Tests and Goodness of Fit in the
Analysis of Covariance Structures. Psychological Bulletin, 88(3), 588-606.

Bhaduri, G. & Ha-Brookshire, J.E. (2011). “Do transparent business practices pay?
Exploration of transparency and consumer purchase intention”. Clothing and Textiles Research
Journal, 29(2), 135-49.

83
Boiral, O., Heras- Saizarbitoria, I., & Testa, F. (2017). SA8000 as CSR- Washing? The
Role of Stakeholder Pressures. Corporate Social Responsibility and Environmental Management,
24(1), 57-70.

Bryman, A. & Bell, E. (2011). Business Research Methods. 3rd Edition., United Kingdom:
Oxford University Press.

Burchell, J., & Cook, J. (2006). It's good to talk? Examining attitudes towards corporate
social responsibility dialogue and engagement processes. Business Ethics: A European Review,
15(2), 154-170.

Burke, J. (2015, April 22). Bangladesh garment workers suffer poor conditions two years
after reform vows. The Guardian. Retrieved May 18, 2018, from
https://www.theguardian.com/world/2015/apr/22/garment-workers-in-bangladesh-still-
suffering-two-years-after-factory-collapse

Byrne, B. M. (1998). Structural Equation Modeling with LISREL, PRELIS and SIMPLIS: Basic
Concepts, Applications and Programming. Mahwah, New Jersey: Lawrence Erlbaum Associates.

Carrigan, M. & Attalla, A. (2001). The myth of the ethical consumer – do ethics matter in
purchase behaviour?. Journal of Consumer Marketing, 18(7), 560-578.

Carter, C. R., & Rogers, D. S. (2008). A framework of sustainable supply chain

management: Moving toward new theory. International Journal of Physical Distribution &
Logistics Management, 38(5), 360-387.

Castaldo, S., Premazzi, K., & Zerbini, F. (2010). The meaning(s) of trust. A content analysis
on the diverse conceptualizations of trust in scholarly research on business relationships.
Journal of Business Ethics, 96(4), 657–668.

Centre for Policy Dialogue. (2013). 100 days of Rana Plaza Tragedy: A report on commitments and
delivery. Retrieved May 18, 2018, from
http://cpd.org.bd/wp-content/uploads/2013/08/100-Days-of-Rana-Plaza-Tragedy-A-
Report-on-Commitments-and-Delivery.pdf

Chaudhuri, A., & 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), 81–93.

Chiu, C.M., Huang, H.Y., & Yen. C.H. (2010). Antecedents of trust in online auctions.
Electronic Commerce Research and Applications, 9(2), 148-159.

Cochran, W. G. (1952). The χ2 test of goodness of fit. Annals of Mathematical Statistics, 23:
315–345.

Cramer, J. (2008). Organising corporate social responsibility in international product

chains. Journal of Cleaner Production, 16(3), 395-400.

84
Delmas, M. A., & Burbano, V. C. (2011). The drivers of greenwashing. California
Management Review, 54, 64–87.

Dickson, M. A. (1999). US consumers' knowledge of and concern with apparel sweatshops.

Journal of Fashion Marketing and Management: An International Journal, 3(1), 44-55.

Dickson, M. A. (2000). Personal Values, Beliefs, Knowledge, and Attitudes Relating to

Intentions to Purchase Apparel from Socially Responsible Businesses. Clothing and Textiles
Research Journal, 18(1), 19-30.

Dickson, M. A. (2001). Utility of no sweat labels for apparel consumers: Pro ling label users
and predicting their purchases. Journal of Consumer Affairs, 35(1), 96–119.

Dickson, M. A. (2013). Corporate responsibility in the global apparel industry: Toward an

integrated human rights-based approach. In S. Black, J. Entwistle, A. de la Haye, A.
Rocamora, R. Root, & H. Thomas (Eds.), The handbook of fashion studies (253–241). London,
England: Bloomsbury Publishing.

Diddi, S., & Niehm, L. (2017). Exploring the role of values and norms towards consumers’
intentions to patronize retail apparel brands engaged in corporate social responsibility
(CSR). Fashion and Textiles, 4(1), 1-20.

Dillman, D. A. (2000). Mail and internet surveys: The tailored design method. New York, NY:
Wiley.

Dingwerth, K. & Eichinger, M. (2010). Tamed Transparency: How Information Disclosure

under the Global Reporting Initiative Fails to Empower. Global Environmental Politics, 10(3),
74-96.

Dodd, M. D., & Supa, D. W. (2011). Understanding the e ect of corporate social
responsibility on consumer purchase intention. Public Relations Journal, 5(3), 1–19.

Doney, P.M. & Cannon, J.P. (1997). An examination of the nature of trust in buyer-seller
relationships, Journal of Marketing, 61( 2), 35-51.

Doorey, D. (2011). The Transparent Supply Chain: From Resistance to Implementation at

Nike and Levi-Strauss. Journal of Business Ethics, 103(4), 587-603.

Du, S., Bhattacharya, C.B., & Sen, S. (2010). Maximizing Business Returns to Corporate
Social Responsibility (CSR): The Role of CSR Communication. (Report). International Journal
of Management Reviews, 12, 8.

Du, X. (2015). How the market values greenwashing? Evidence from China.
Journal of Business Ethics, 128, 547–574.

Dubbink, W., Graafland, J., & Liedekerke, L. (2008). CSR, Transparency and the Role of
Intermediate Organisations. Journal of Business Ethics, 82(2), 391-406.

85
Egels-Zandén, N., Hulthén, K., & Wulff, G. (2015). Trade-offs in supply chain
transparency: The case of Nudie Jeans Co. Journal of Cleaner Production, 107, 95-104.

Eisingerich, A. B., & Bell, S. J. (2008). Customer education increases trust.(Intelligence).

MIT Sloan Management Review, 50(1), 10-11.

Erdem, T., & Swait, J. (2004). Brand credibility, brand consideration, and choice. Journal of
Consumer Research, 31(1), 191–198.

Everlane. (2018). About. Retrieved May 18, 2018, from https://www.everlane.com/about

Fairtrade. (2018). Standards. Retrieved May 18, 2018, from

https://www.fairtrade.net/standards.html

Fashion Revolution. (2017). Fashion transparency index. Retrieved May 18, 2018, from
https://www.fashionrevolution.org/faqs-fashion-transparency-index-2017/
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory
and research. Reading, Mass: Addison-Wesley.

Fornell, C., & Larcker, D. (1981). Evaluating Structural Equation Models with
Unobservable Variables and Measurement Error. Journal of Marketing Research, 18(1), 39.

Foscht, T., Lin, Y. & Eisingerich, A. B. (2018). Blinds up or down?: The influence of
transparency, future orientation, and CSR on sustainable and responsible behavior.
European Journal of Marketing, 52(3/4), 476-498

Gefen, D., Karahanna, E., & Straub, D.W. (2003). Trust and TAM in online shopping: An
integrated model (1). (Research Article). MIS Quarterly, 27(1), 51-90.

George, D., & Mallery, M. (2010). SPSS for Windows Step by Step: A Simple Guide and Reference
(10 ed.). Boston: Pearson.

Goldman Sachs. (n.d.). Millennials coming of age (infographic). Retrieved May 18, 2018, from
http://www.goldmansachs.com/our-thinking/pages/millennials/

Groza, M. D., Pronschinske, M. R., & Walker, M. (2011). Perceived organizational motives
and consumer responses to proactive and reactive CSR. Journal of Business Ethics, 102(4),
639–652.

Gulliksen, H., & J. W. Tukey. (1958). Reliability for the law of comparative judgment.
Psychometrika, 23, 95–110.

Gwilt, A., & Rissanen, T. (2011). Shaping sustainable fashion: Changing the way we make and
use clothes. London: Earthscan.

Ha-Brookshire, J., & Hodges, N. (2009). Socially Responsible Consumer Behavior? Clothing
and Textiles Research Journal, 27(3), 179-196.

86
Harrison, R., Newholm, T. & Shaw, D. (2005). The ethical consumer (1st ed.). London: SAGE.

Henninger, C. E., Alevizou, P. J. & Oates, C. J. (2016). What is sustainable fashion? Journal
of Fashion Marketing and Management: An International Journal, 20(4), 400–416.

Hill, J., & Lee, H. H. (2012). Young Generation Y consumers’ perceptions of sustainability in
the apparel industry. Journal of Fashion Marketing and Management, 16(4), 477–491.

Hillenbrand, C., Money, K., & Ghobadian, A. (2013). Unpacking the Mechanism by which
Corporate Responsibility Impacts Stakeholder Relationships. British Journal of Management,
24(1), 127-146.

Hiller Connell, K.Y. and Kozar, J.M. (2009). Students’ attitude-behaviour gap on issues of
labor exploitation. Proceedings of the International Textile and Apparel Association. Washington,
Bellevue.

Hobson, J. (2013). To die for? The health and safety of fast fashion. Occupational Medicine,
63(5), 317-319.

Honestby. (2018). About. Retrieved May 19, 2018, from

http://www.honestby.com/en/page/16/about.html

Hooper, D. R., Coughlan, J., & Mullen, M. (2008). Structural equation modelling:
Guidelines for determining model fit. Electronic Journal of Business Research Methods, 6(1), 53-
60.

Hu, L.T. & Bentler, P.M. (1999). Cutoff Criteria for Fit Indexes in Covariance Structure
Analysis: Conventional Criteria Versus New Alternatives. Structural Equation Modeling, 6(1),
1-55.

Hustvedt, G., & Dickson, M. (2009). Consumer likelihood of purchasing organic cotton
apparel. Journal of Fashion Marketing and Management: An International Journal, 13(1), 49-65.

Hwang, C., Lee, Y., & Diddi, S. (2014). Generation Y's moral obligation and purchase
intentions for organic, fair-trade, and recycled apparel products. International Journal of
Fashion Design, Technology and Education,1-11.

James, A., & Montgomery, B. (2017). Engaging the fashion consumer in a transparent
business model. International Journal of Fashion Design, Technology and Education, 10(3), 287-299.

Jacobs, B., & Singhal. V.R. (2017). The effect of the Rana Plaza disaster on shareholder
wealth of retailers: Implications for sourcing strategies and supply chain governance. Journal
of Operations Management, 49(51), 52-66.

Jin Ma, Y., Littrell, M., & Niehm, L. (2012). Young female consumers' intentions toward
fair trade consumption. International Journal of Retail & Distribution Management, 40(1), 41-63.

87
Jonsen, K., Galunic, C., Weeks, J. & Braga, T. (2015). Evaluating espoused values: Does
articulating values pay off? European Management Journal, 33(5), 332-340.

Jöreskog, K. (1969). A general approach to confirmatory maximum likelihood factor

analysis. Psychometrika, 34(2), 183-202.

Jöreskog, K. G., & Sörbom, D. (1993). LISREL 8: Structural Equation Modeling with the
SIMPLIS Command Language. Chicago, IL: Scientific Software International Inc.

Kang, J., & Hustvedt, G. (2014). Building Trust Between Consumers and Corporations:
The Role of Consumer Perceptions of Transparency and Social Responsibility. Journal of
Business Ethics, 125(2), 253-265.

Keller, K. (1993). Conceptualizing, Measuring, and Managing Customer-Based Brand

Equity. Journal of Marketing, 57(1), 1-22.

Kim, S., Littrell, M., & Ogle, J. (1999). The relative importance of social responsibility as a
predictor of purchase intentions for clothing. Journal of Fashion Marketing and Management,
3(3), 207-218.

Kozar, J., & Hiller Connell, K. (2013). Socially and environmentally responsible apparel
consumption: Knowledge, attitudes, and behaviors. Social Responsibility Journal, 9(2), 315-
324.

Krishnamurthy, P., & Sivaraman, A. (2002). Counterfactual thinking and advertising

responses. Journal of Consumer Research, 28(3), 650–658.

Laudal, T. (2010). An attempt to determine the CSR potential of the international clothing
business. Journal of Business Ethics, 96(1), 63–77.

Lee, H., Park, T., Moon, H.K., Yang, Y., & Kim, C. (2009). Corporate philanthropy,
attitude towards corporations, and purchase intentions: A South Korea study. Journal of
Business Research, 62(10), 939-946.

Lee, N., Choi, Y. J., Youn, C., & Lee, Y. (2012). Does Green Fashion Retailing Make
Consumers More Eco-friendly? Clothing and Textiles Research Journal, 30(1), 67-82.

Lundblad, L., & Davies, I. (2016). The values and motivations behind sustainable fashion
consumption. Journal of Consumer Behaviour, 15(2), 149-162.

Maignan, I., & Ferrell, O. (2004). Corporate social responsibility and marketing: An
integrative framework. Journal of the Academy of Marketing Science, 32(1), 3-19.

Mani, V., Agarwal, R., Gunasekaran, A., Papadopoulos, T., Dubey, R. & Childe, S.J. (2016).
Social sustainability in the supply chain: Construct development and measurement
validation. Ecological Indicators,71, 270-279.

88
Marin, L., Ruiz, S., & Rubio, A. (2009). The role of identity salience in the e ects of
corporate social responsibility on consumer behavior. Journal of Business Ethics, 84(1), 65–78.

Marquis, C. & Toffel, M. W. (2012). When do Firms Greenwash? Corporate Visibility, Civil Society
Scrutiny, and Environmental Disclosure. Harvard Business School: USA.

Marshall, D., McGrath, P., Harrigan, F., & McCarthy, L. (2016). What's your strategy for
supply chain disclosure? MIT Sloan Management Review, 57(2), 36-45.

Masson, R., Iosif, L., Mackerron, G., & Fernie, J. (2007). Managing complexity in agile
global fashion industry supply chains. The International Journal of Logistics Management, 18(2),
238-254.

Mayser, R. N., & Zick, C. D. (1993). Poisoning the well: Do environmental claims strain
consumer credulity? Advances in Consumer Research, 20(1), 698–703.

McNeill, L., & Moore, R. (2015). Sustainable fashion consumption and the fast fashion
conundrum: Fashionable consumers and attitudes to sustainability in clothing choice.
International Journal of Consumer Studies, 39(3), 212-222.

Methven O'Brien, C., & Dhanarajan, S. (2016). The corporate responsibility to respect
human rights: A status review. Accounting, Auditing & Accountability Journal, 29(4), 542-567.

Middlemiss, N. (2003). Authentic not cosmetic: CSR as brand enhancement. Journal of Brand
Management, 10(4), 353.

Mittelstaedt, J.D, Shultz, C.J, Kilbourne, W.E., Peterson, M., McDonagh, P., & Prothero,
A. (2014). Sustainability as Megatrend. Journal of Macromarketing, 34(3), 253-264.

Mol, A. (2015). Transparency and value chain sustainability. Journal of Cleaner Production, 107,
154-161.

Moosmayer, D. C., & Fuljahn, A. (2010). Consumer perceptions of cause related marketing
campaigns. Journal of Consumer Marketing, 27(6), 543–549.

Morgan, R,M., & Hunt, S.D. (1994). The commitment-trust theory of relationship
marketing. Journal of Marketing, 58(3), 20-38.

New, S. (1997). The scope of supply chain management research. Supply Chain Management:
An International Journal, 2(1), 15-22.

Newell, S. & Goldsmith, R. E. (2001). The development of a scale to measure perceived

corporate credibility. Journal of Business Research, 52(3), 235-247.

Nudie Jeans co (2018). This is nudie jeans. Retrieved May 19, 2018, from
https://www.nudiejeans.com/page/this-is-nudie-jeans

89
O'Rourke, D. (2003). Outsourcing Regulation: Analyzing Nongovernmental Systems of
Labor Standards and Monitoring. Policy Studies Journal, 31(1), 1-29.

Ortiz Martinez, E., & Crowther, D. (2008). Is disclosure the right way to comply with
stakeholders? The Shell case. Business Ethics: A European Review, 17(1), 13-22.

Pallant, J. (2007). SPSS survival manual: A step by step guide to data analysis using SPSS for Windows
(3rd ed.). Maidenhead: Open University Press.

Patagonia. (2018). Footprint. Retrieved May 19, 2018, from

http://www.patagonia.com/footprint.html

Pavlou, P.A., & Fygenson, M. (2006). Understanding and predicting electronic commerce
adoption: An extension of the theory of planned behavior. MIS Quarterly, 30(1), 115-143.

Peattie, K. & Crane, A., 2005. Green marketing:legend, myth, farce or prophesy?
Qualitative Market Research: An International Journal, 8(4), 357-370.

Pedersen, E., & Gwozdz, R. (2014). From Resistance to Opportunity-Seeking: Strategic

Responses to Institutional Pressures for Corporate Social Responsibility in the Nordic
Fashion Industry. Journal of Business Ethics, 119(2), 245-264.

Penn Schoen Berland. (2010). Corporate Social Responsibility Branding Survey, 2010. Retrieved
May 18, 2018, from http://www.burson-marsteller.swiss/what-we-do/our-
thinking/corporate-social-responsibility-branding-survey-2010/?lang=en

Perry, P., & Towers, N. (2009). Determining the antecedents for a strategy of corporate
social responsibility by small- and medium-sized enterprises in the UK fashion apparel
industry. Journal of Retailing and Consumer Services,16(5), 377-385.

Pirnea, I., & Ghenta, M. (2013). Social implications of social responsibility with the
purpose of increasing the enterprise competitiveness. Calitatea, 14(2), 214-220.

Poortinga, W., Steg, L., & Vlek, C. (2004). Values, environmental concern, and
environmental behavior: A study into household energy use. Environment and Behavior, 36(1),
70–93.

Pope, S., & Wæraas, A. (2016). CSR-Washing is Rare: A Conceptual Framework, Literature
Review, and Critique. Journal of Business Ethics, 137(1), 173-193.

Qualtrics Library (2018). Library. Retrieved May 18, 2018, from

www.jibs.eu.qualtrics.com/ControlPanel/

Reinecke, J. & Donaghey, J. (2015). After Rana Plaza: Building coalitional power for labour
rights between unions and (consumption-based) social movement organisations.
Organization, 22(5), 720–740.

90
Reynolds, M., & Yuthas, K. (2008). Moral Discourse and Corporate Social Responsibility
Reporting. Journal of Business Ethics, 78(1), 47-64.

Sabri, A. (2016). Assessing the Validity and Reliability of a Measurement Model in

Structural Equation Modeling (SEM). British Journal of Mathematics & Computer Science, 15(3),
1-8.

Saunders, M., Lewis, P. & Thornhill, A. (2009). Research methods for business students (5th ed.).
Harlow, England: Pearson Education.

Sengupta, J., & Johar, G. V. (2002). Effects of inconsistent attributes information on the
predictive value of product attributes: toward a resolution of opposing perspectives. Journal
of Consumer Research, 29(6), 39–56.

Setiawan, R., & Aachyar, A. (2012). Effects of Perceived Trust and Perceived Price on
Customers’ Intention to Buy in Online Store in Indonesia. Asean Marketing Journal, 4(1), 26-
36.

Shaw, D., Hogg, G., Wilson, E., Shiu, E., & Hassan, L. (2006). Fashion victim: The impact
of fair trade concerns on clothing choice. Journal of Strategic Marketing, 14(4), 427-440.

Shim, S. (1995). Environmentalism and consumers’ clothing disposal patterns: An

exploratory study. Clothing and Textiles research journal, 13(1), 38-48

Siano, A., Vollero, A., Conte,F., & Amabile, S. (2017). "More than words": Expanding the
taxonomy of greenwashing after the Volkswagen scandal. Journal of Business Research, 71, 27-
37.

Singh, J., Salmones Sanchez, M., & Bosque, I. (2008). Understanding Corporate Social
Responsibility and Product Perceptions in Consumer Markets: A Cross-cultural Evaluation.
Journal of Business Ethics, 80(3), 597-611.

Sirdeshmukh, D., Singh, J., & Sabol, B. (2002). Consumer Trust, Value, and Loyalty in
Relational Exchanges. Journal of Marketing, 66(1), 15-37.

Social Accountability International. (2018). SA8000 Standard. Retrieved May 18, 2018, from
http://www.sa-intl.org/index.cfm?fuseaction=Page.ViewPage&PageID=1689

Solomon, M.R., Bamossy, G., Askegaard, S., & Hogg, M. (2010). Consumer Behaviour: A
European perspective. Essex:Prentice Hall

Straughan, R., & Roberts, J. (1999). Environmental segmentation alternatives: A look at

green consumer behavior in the new millennium. Journal of Consumer Marketing, 16(6), 558-
575.

Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.). New York:
HarperCollins.

91
Thomas, S. (2008). From “green blur” to ecofashion: Fashioning an eco-lexicon. Fashion
Theory: The Journal of Dress, Body & Culture, 12(4), 525–540.

Transparency International. (2018). FAQS ON CORRUPTION. Retrieved May 18, 2018,

from https://www.transparency.org/whoweare/organisation/faqs_on_corruption

Van Der Zee, D., & Van Der Vorst, J. (2005). A Modeling Framework for Supply Chain
Simulation: Opportunities for Improved Decision Making. Decision Sciences, 36(1), 65-95.

Vanhamme, J., & Grobben, B. (2009). ‘‘Too good to be true!’’. The effectiveness of CSR
history in countering negative publicity. Journal of Business Ethics, 85(Supplement 2), 273–
283.

Waddock, S. (2008). Building a new institutional infrastructure for corporate responsibility.

The Academy of Management Perspectives, 22(3), 87-108.

Wagner, T., Lutz, R. J., & Weitz, B. A. (2009). Corporate hypocrisy: Overcoming the threat
of inconsistent corporate social responsibility perceptions. Journal of Marketing, 73, 77–91.

Wu, P., Yeh, G. Y-Y., & Hsiao, C-R. (2011). The effect of store image and service quality
on brand image and purchase intention for private label brands. Australasian Marketing
Journal, 19(1), 30-39.

Yan, R.N., Ogle, J., & Hyllegard, K. (2010). The impact of message appeal and message
source on Gen Y consumers’ attitudes and purchase intentions toward american apparel.
Journal of Marketing Communications, 16(4), 203–224.

Yoon, Y., Gurhan-Canli, Z., & Schwarz, N. (2006). The effect of corporate social
responsibility (CSR) activities on companies with bad reputations. Journal of Consumer
Psychology, 16(4), 377–390.

92
Appendix 1: Scenarios

93
94
95
96
Appendix 2: Questionnaire

97
98
99
Appendix 3: Item correlations matrix

Correlations matrix

D1 D2 D3 A1 A2 T1 T2 T3 GA1 GA2 GA3 PI1 PI2 PI3

D1 1 0,788 0,77 0,692 0,671 0,729 0,652 0,563 0,664 0,7 0,679 0,663 0,652 0,402

D2 0,788 1 0,685 0,635 0,64 0,647 0,574 0,515 0,627 0,632 0,63 0,597 0,573 0,402

D3 0,77 0,685 1 0,695 0,631 0,7 0,666 0,556 0,666 0,689 0,638 0,598 0,628 0,426

A1 0,692 0,635 0,695 1 0,68 0,665 0,651 0,628 0,731 0,686 0,657 0,611 0,603 0,418
A2 0,671 0,64 0,631 0,68 1 0,715 0,635 0,519 0,622 0,637 0,631 0,564 0,561 0,417
T1 0,729 0,647 0,7 0,665 0,715 1 0,803 0,676 0,74 0,761 0,754 0,721 0,695 0,504

T2 0,652 0,574 0,666 0,651 0,635 0,803 1 0,7 0,715 0,696 0,708 0,679 0,691 0,483

T3 0,563 0,515 0,556 0,628 0,519 0,676 0,7 1 0,645 0,683 0,663 0,62 0,585 0,441

GA1 0,664 0,627 0,666 0,731 0,622 0,74 0,715 0,645 1 0,76 0,73 0,672 0,687 0,501

GA2 0,7 0,632 0,689 0,686 0,637 0,761 0,696 0,683 0,76 1 0,812 0,725 0,721 0,499

GA3 0,679 0,63 0,638 0,657 0,631 0,754 0,708 0,663 0,73 0,812 1 0,766 0,76 0,53

PI1 0,663 0,597 0,598 0,611 0,564 0,721 0,679 0,62 0,672 0,725 0,766 1 0,748 0,492

PI2 0,652 0,573 0,628 0,603 0,561 0,695 0,691 0,585 0,687 0,721 0,76 0,748 1 0,529

PI3 0,402 0,402 0,426 0,418 0,417 0,504 0,483 0,441 0,501 0,499 0,53 0,492 0,529 1

100
Appendix 4: Pearson correlation between constructs

Disclosure Assessment Trust General Attitude Purchase Intention

Disclosure 1 ,791** ,755** ,785** ,725**
Assessment ,791** 1 ,767** ,785** ,682**
Trust ,755** ,767** 1 ,851** ,786**
General Attitude ,785** ,785** ,851** 1 ,839**
Purchase Intention ,725** ,682** ,786** ,839** 1
N=432, **Correlation is significant at the 0.01 level (2-tailed).

101
Appendix 5: Regression
Disclosure and Assessment on Trust

Assessing tolerance and VIF

Variable Tolerance VIF

Disclosure 0,375 2,667

Assessment 0,375 2,667

Dependent variable: Trust

Condition indices & decomposition of coefficient variance matrix

Number Eigenvalue Condition Index Variance Proportions

Intercept Disclosure Assessment
1 2,86 1 0,02 0,01 0,01
2 0,105 5,212 0,93 0,16 0,05
3 0,034 9,143 0,05 0,83 0,94
Underlined values exceed the .90 level
Dependent variable: Trust

102
 R square & Significance

R Square 0,648
ANOVA Sig. 0
Predictors: Assessment & Disclosure
Dependent variable: Trust

Beta values

Standardized coefficient Beta Sig.

Disclosure 0,397 0
Assessment 0,453 0
Dependent variable: Trust

Disclosure and Assessment on General Attitude

Assessing tolerance and VIF

Variable Tolerance VIF

Disclosure 0,375 2,667

Assessment 0,375 2,667

103
Dependent variable: General Attitude

Condition indices & decomposition of coefficient variance matrix

Number Eigenvalue Condition Index Variance Proportions

Intercept Disclosure Assessment
1 2,86 1 0,02 0,01 0,01
2 0,105 5,212 0,93 0,16 0,05
3 0,034 9,143 0,05 0,83 0,94
Underlined values exceed the .90 level
Dependent variable: General Attitude

104
 R square & Significance

R Square 0,689
ANOVA Sig. 0
Predictors: Assessment & Disclosure
Dependent variable: Trust

Beta values

Standardized coefficient Beta Sig.

Disclosure 0,44 0
Assessment 0,437 0
Dependent variable: General Attitude

General Attitude and Trust on Purchase Intention

Assessing tolerance and VIF

Variable Tolerance VIF

Trust 0,276 3,629

General Attitude 0,276 3,629

Dependent variable: Purchase Intention

Condition indices & decomposition of coefficient variance matrix

Number Eigenvalue Condition Index Variance Proportions

Intercept Trust General Attitude
1 2,917 1 0,01 0 0
2 0,068 6,569 0,99 0,07 0,08
3 0,016 13,565 0 0,93 0,92
Underlined values exceed the .90 level
Dependent variable: Purchase Intention

105
 R square & Significance

R Square 0,723
ANOVA Sig. 0
Predictors: Trust & General Attitude
Dependent variable: Purchase Intention

Beta values

Standardized coefficient Beta Sig.

Trust 0,263 0
General Attitude 0,615 0
Dependent variable: Purchase Intention

106
Appendix 6: Curve Estimation

Model Summary and Parameter Estimates

Dependent Variable: General Attitude
Equation Model Summary Parameter Estimates
R Square F df1 df2 Sig. Constant b1 b2 b3
Linear 0,617 692,525 1 430 0 1,81 0,647
Logarithmic 0,573 578,148 1 430 0 2,14 1,82
Inverse 0,481 399,301 1 430 0 5,686 −3,642
Quadratic 0,618 346,695 2 429 0 1,965 0,533 0,016
Cubic 0,618 230,923 3 428 0 1,763 0,778 −0,06 0,007
Compound 0,542 509,771 1 430 0 1,996 1,196
Power 0,545 514,738 1 430 0 2,134 0,522
S 0,494 419,178 1 430 0 1,79 −1,085
Growth 0,542 509,771 1 430 0 0,691 0,179
Exponential 0,542 509,771 1 430 0 1,996 0,179
Logistic 0,542 509,771 1 430 0 0,501 0,836
The independent variable is Disclosure.

Model Summary and Parameter Estimates

Dependent Variable: TrustM
Equation Model Summary Parameter Estimates
R Square F df1 df2 Sig. Constant b1 b2 b3
Linear 0,571 571,526 1 430 0 1,915 0,609 0,571
Logarithmic 0,529 482,444 1 430 0 2,228 1,709 0,529
Inverse 0,442 339,948 1 430 0 5,555 −3,411 0,442
Quadratic 0,571 286,072 2 429 0 2,065 0,498 0,015 0,571
Cubic 0,571 190,273 3 428 0 2,046 0,521 0,008 0,571
Compound 0,496 423,042 1 430 0 2,038 1,186 0,496
Power 0,498 427,171 1 430 0 2,173 0,499 0,498
S 0,451 353,493 1 430 0 1,763 −1,037 0,451
Growth 0,496 423,042 1 430 0 0,712 0,171 0,496
Exponential 0,496 423,042 1 430 0 2,038 0,171 0,496
Logistic 0,496 423,042 1 430 0 0,491 0,843 0,496
The independent variable is Disclosure.

107
Model Summary and Parameter Estimates
Dependent Variable: TrustM
Equation Model Summary Parameter Estimates
R Square F df1 df2 Sig. Constant b1 b2 b3
Linear 0,589 615,422 1 430 0 1,609 0,652
Logarithmic 0,554 534,54 1 430 0 1,721 1,986
Inverse 0,449 350,722 1 430 0 5,644 −4,161
Quadratic 0,589 307,438 2 429 0 1,496 0,725 −0,009
Cubic 0,59 205,12 3 428 0 1,819 0,373 0,093 −0,009
Compound 0,525 474,312 1 430 0 1,853 1,203
Power 0,54 504,842 1 430 0 1,851 0,589
S 0,483 401,598 1 430 0 1,801 −1,297
Growth 0,525 474,312 1 430 0 0,617 0,185
Exponential 0,525 474,312 1 430 0 1,853 0,185
Logistic 0,525 474,312 1 430 0 0,54 0,831
The independent variable is AssessmentM.

Model Summary and Parameter Estimates

Dependent Variable: GeneralAttitudeM
Equation Model Summary Parameter Estimates
R Square F df1 df2 Sig. Constant b1 b2 b3
Linear 0,616 690,264 1 430 0 1,53 0,682 0,616
Logarithmic 0,556 538,118 1 430 0 1,704 2,033 0,556
Inverse 0,427 320,876 1 430 0 5,683 −4,149 0,427
Quadratic 0,617 346,266 2 429 0 1,754 0,537 0,019 0,617
Cubic 0,62 232,312 3 428 0 2,296 −0,055 0,191 0,62
Compound 0,543 511,306 1 430 0 1,845 1,207 0,543
Power 0,531 487,695 1 430 0 1,879 0,585 0,531
S 0,446 346,213 1 430 0 1,793 −1,247 0,446
Growth 0,543 511,306 1 430 0 0,613 0,188 0,543
Exponential 0,543 511,306 1 430 0 1,845 0,188 0,543
Logistic 0,543 511,306 1 430 0 0,542 0,828 0,543
The independent variable is AssessmentM.

108
Model Summary and Parameter Estimates
Dependent Variable: PurchaseIntentionM
Equation Model Summary Parameter Estimates
R Square F df1 df2 Sig. Constant b1 b2 b3
Linear 0,704 1022,384 1 430 0 0,252 0,892 0,704
Logarithmic 0,67 874,337 1 430 0 0,011 2,959 0,67
Inverse 0,526 477,857 1 430 0 5,977 −6,659 0,526
Quadratic 0,707 516,391 2 429 0 −0,211 1,157 −0,033 0,707
Cubic 0,71 349,719 3 428 0 0,816 0,115 0,263 0,71
Compound 0,695 979,395 1 430 0 1,132 1,316 0,695
Power 0,748 1277,663 1 430 0 0,971 0,97 0,748
S 0,68 912,208 1 430 0 1,972 −2,347 0,68
Growth 0,695 979,395 1 430 0 0,124 0,275 0,695
Exponential 0,695 979,395 1 430 0 1,132 0,275 0,695
Logistic 0,695 979,395 1 430 0 0,883 0,76 0,695
The independent variable is GeneralAttitudeM.

Model Summary and Parameter Estimates

Dependent Variable: PurchaseIntentionM
Equation Model Summary Parameter Estimates
R Square F df1 df2 Sig. Constant b1 b2 b3
Linear 0,667 859,837 1 430 0 0,637 0,904 0,667
Logarithmic 0,652 804,862 1 430 0 0,594 2,798 0,652
Inverse 0,552 528,953 1 430 0 6,143 −6,349 0,552
Quadratic 0,671 436,759 2 429 0 0,054 1,282 −0,052 0,671
Cubic 0,672 291,701 3 428 0 0,627 0,666 0,138 0,672
Compound 0,63 733,139 1 430 0 1,304 1,314 0,63
Power 0,675 893,56 1 430 0 1,226 0,883 0,675
S 0,638 757,908 1 430 0 1,992 −2,118 0,638
Growth 0,63 733,139 1 430 0 0,265 0,273 0,63
Exponential 0,63 733,139 1 430 0 1,304 0,273 0,63
Logistic 0,63 733,139 1 430 0 0,767 0,761 0,63
The independent variable is AssessmentS.

109
Appendix 7: Cook’s distance
General attitude & trust on purchase intention

Degree of disclosure & degree of assessment on general attitude

110
Degree of disclosure & degree of assessment on trust

111
Appendix 8: Kurtosis and Skewness

D1 D2 D3 A1 A2 T1 T2 T3 GA1 GA2 GA3 PI1 PI2

N Valid 432 432 432 432 432 432 432 432 432 432 432 432 432

Missing 0 0 0 0 0 0 0 0 0 0 0 0 0

Skewness −0,177 −0,417 0,269 −0,078 −0,245 −0,2 −0,244 −0,319 −0,243 −0,209 −0,3 −0,282 −0,037
Std. Error of
Skewness 0,117 0,117 0,117 0,117 0,117 0,117 0,117 0,117 0,117 0,117 0,117 0,117 0,117

Kurtosis −1,359 −1,179 −1,308 −1,204 −1,048 −0,805 −0,758 −0,655 −0,643 −0,849 −0,515 −0,71 −0,97
Std. Error of
Kurtosis 0,234 0,234 0,234 0,234 0,234 0,234 0,234 0,234 0,234 0,234 0,234 0,234 0,234

112
Appendix 9: Homoscedasticity
General attitude on purchase intention

Trust on purchase intention

Degree of assessment on trust

113
Degree of assessment on general attitude

114
Degree of disclosure on general attitude

Degree of disclosure on trust

Appendix 10: Measurement model results

115
Variances: (Group number 1 - Default model)

Estimate S.E. C.R. P Label

e1 1,008 0,093 10,813 ***

e2 1,374 0,121 11,376 ***

e3 0,84 0,094 8,926 ***

e4 1,224 0,105 11,62 ***

e5 1,027 0,101 10,147 ***

e6 0,912 0,076 11,991 ***

e7 0,656 0,056 11,779 ***

e8 0,383 0,049 7,893 ***

e9 0,567 0,048 11,835 ***

e10 0,693 0,06 11,508 ***

e11 0,821 0,078 10,527 ***

e12 0,64 0,062 10,327 ***

e13 0,455 0,045 10,156 ***

116
Standardized Regression Weights: (Group number 1 - Default model)

Estimate

S1.D3 <--- DisclosureF 0,861

S1.D2 <--- DisclosureF 0,806

S1.D1 <--- DisclosureF 0,891

S1.A2 <--- AssessmentF 0,805

S1.A1 <--- AssessmentF 0,845

S1.T3 <--- TrustF 0,795

S1.T2 <--- TrustF 0,872

S1.T1 <--- TrustF 0,924

S1.G2 <--- GenAttF 0,892

S1.G1 <--- GenAttF 0,862

S1.P2 <--- PIF 0,863

S1.P1 <--- PIF 0,867

S1.G3 <--- GenAttF 0,904

117
Regression Weights: (Group number 1 - Default model)

Estimate S.E. C.R. P Label

S1.D3 <--- DisclosureF 1

S1.D2 <--- DisclosureF 0,941 0,047 20,116 ***

S1.D1 <--- DisclosureF 1,058 0,044 24,008 ***

S1.A2 <--- AssessmentF 1

S1.A1 <--- AssessmentF 1,066 0,053 19,974 ***

S1.T3 <--- TrustF 1

S1.T2 <--- TrustF 1,149 0,056 20,444 ***

S1.T1 <--- TrustF 1,196 0,058 20,574 ***

S1.G2 <--- GenAttF 1

S1.G1 <--- GenAttF 0,953 0,037 25,43 ***

S1.P2 <--- PIF 1

S1.P1 <--- PIF 0,9 0,038 23,481 ***

S1.G3 <--- GenAttF 0,958 0,034 28,365 ***

118
Appendix 11: Measurement model validity and reliability

Standardized Regression Weights: (Group number 1 - Default model)

Estimate delta^2 e AVE CR

S1.D3 <---
DisclosureF 0,861 0,741321 0,258679

S1.D2 <--- DisclosureF 0,806 0,649636 0,350364

S1.D1 <--- DisclosureF 0,891 0,793881 0,206119

2,558 2,184838 0,815162 0,7282793333 0,8892221078

S1.A2 <--- AssessmentF 0,805 0,648025 0,351975

S1.A1 <--- AssessmentF 0,845 0,714025 0,285975

1,65 1,36205 0,63795 0,681025 0,8101593537

S1.T3 <--- TrustF 0,795 0,632025 0,367975

S1.T2 <--- TrustF 0,872 0,760384 0,239616

S1.T1 <--- TrustF 0,924 0,853776 0,146224

2,591 2,246185 0,38584 0,7487283333 0,760384

S1.G2 <--- GenAttF 0,892 0,795664 0,204336

<---
S1.G1 GenAttF 0,862 0,743044 0,256956

S1.G3 <--- GenAttF 0,904 0,817216 0,182784

2,658 2,355924 0,644076 0,785308 0,9164518539

S1.P2 <--- PIF 0,863 0,744769 0,255231

<---
S1.P1 PIF 0,867 0,751689 0,248311

1,73 1,496458 0,503542 0,748229 0,8559844551

CR Formula: Ʃκ)² / [(Ʃκ)² + (Ʃ1 - κ²)]

AVE Formula: Ʃ κ² / n
(Sabri, 2016)

K = factor loading of every item

n = number of items in a model

119
Appendix 12: Structural model results

Variances: (Group number 1 - Standardized Regression Weights: (Group number 1 -

Default model) Default model)

Estimate S.E. C.R. P Label Estimate

e14 0,423 0,061 6,943 *** TrustF <--- AssessmentF 0,715

e15 0,315 0,053 5,942 *** TrustF <--- DisclosureF 0,189

e16 0,253 0,061 4,136 *** GenAttF <--- DisclosureF 0,252
e1 1 0,093 10,736 *** GenAttF <--- AssessmentF 0,659
e2 1,376 0,121 11,379 *** PIF <--- GenAttF 0,837
e3 0,852 0,095 8,996 *** PIF <--- TrustF 0,117
e4 1,223 0,106 11,57 *** S1.D3 <--- DisclosureF 0,862
e5 1,027 0,102 10,095 *** S1.D2 <--- DisclosureF 0,806
e6 0,91 0,076 11,962 *** S1.D1 <--- DisclosureF 0,889
e7 0,658 0,056 11,785 *** S1.A2 <--- AssessmentF 0,805
e8 0,38 0,049 7,827 *** S1.A1 <--- AssessmentF 0,844
e9 0,561 0,048 11,752 *** S1.T3 <--- TrustF 0,796
e10 0,698 0,06 11,539 *** S1.T2 <--- TrustF 0,871
e11 0,824 0,079 10,49 *** S1.T1 <--- TrustF 0,925
e12 0,638 0,062 10,247 *** S1.G2 <--- GenAttF 0,893
e13 0,452 0,044 10,162 *** S1.G1 <--- GenAttF 0,861

S1.P2 <--- PIF 0,863

S1.P1 <--- PIF 0,868

S1.G3 <--- GenAttF 0,904

120
Regression Weights: (Group number 1 - Default model)
Estimate S.E. C.R. P Label

TrustF <--- AssessmentF 0,598 0,17 3,51 ***

TrustF <--- DisclosureF 0,14 0,145 0,964 0,335
GenAttF <--- DisclosureF 0,22 0,162 1,36 0,174

GenAttF <--- AssessmentF 0,654 0,189 3,45 ***

PIF <--- GenAttF 0,871 0,112 7,743 ***

PIF <--- TrustF 0,144 0,13 1,114 0,265

S1.D3 <--- DisclosureF 1

S1.D2 <--- DisclosureF 0,94 0,047 20,13 ***

S1.D1 <--- DisclosureF 1,055 0,044 23,966 ***

S1.A2 <--- AssessmentF 1

S1.A1 <--- AssessmentF 1,066 0,054 19,904 ***

S1.T3 <--- TrustF 1

S1.T2 <--- TrustF 1,147 0,056 20,443 ***

S1.T1 <--- TrustF 1,196 0,058 20,579 ***

S1.G2 <--- GenAttF 1

S1.G1 <--- GenAttF 0,951 0,037 25,39 ***

S1.P2 <--- PIF 1

S1.P1 <--- PIF 0,901 0,038 23,406 ***

S1.G3 <--- GenAttF 0,957 0,034 28,464 ***

121
Appendix 13: Revised model results

Variances: (Group number 1 - Default

model)
Estimate S.E. C.R. P Label
DisclosureS 2,889 0,264 10,947 ***

e17 0,299 0,056 5,329 ***

e14 0,228 0,042 5,438 ***

e15 0,169 0,035 4,849 ***

e16 0,268 0,061 4,4 ***
e1 0,997 0,094 10,62 ***
e2 1,392 0,123 11,35 ***
e3 0,848 0,096 8,876 ***
e4 1,361 0,102 13,318 ***
e5 1,193 0,092 12,912 ***
e6 0,927 0,078 11,942 ***
e7 0,656 0,056 11,67 ***

e8 0,37 0,049 7,57 ***

e9 0,561 0,048 11,639 ***
e10 0,683 0,06 11,297 ***
e11 0,823 0,079 10,461 ***

e12 0,639 0,062 10,229 ***

e13 0,448 0,045 9,936 ***

122
Standardized Regression Weights: (Group number
1 - Default model)
Estimate

AssessmentS <--- DisclosureS 0,927

TrustS <--- AssessmentS 0,944

GenAttS <--- AssessmentS 0,947

PIS <--- GenAttS 0,783

PIS <--- TrustS 0,175

S1.D3 <--- DisclosureS 0,862
S1.D2 <--- DisclosureS 0,803

S1.D1 <--- DisclosureS 0,89

S1.A2 <--- AssessmentS 0,78

S1.A1 <--- AssessmentS 0,817

S1.T3 <--- TrustS 0,792

S1.T2 <--- TrustS 0,872

S1.T1 <--- TrustS 0,927

S1.G2 <--- GenAttS 0,893

S1.G1 <--- GenAttS 0,864

S1.P2 <--- PIS 0,862

S1.P1 <--- PIS 0,867

S1.G3 <--- GenAttS 0,905

123
Regression Weights: (Group number 1 - Default model)

Estimate S.E. C.R. P Label

AssessmentS <--- DisclosureS 0,792 0,044 17,925 ***

TrustS <--- AssessmentS 0,811 0,048 16,887 ***

GenAttS <--- AssessmentS 0,969 0,049 19,603 ***

PIS <--- GenAttS 0,814 0,091 8,975 ***

PIS <--- TrustS 0,216 0,105 2,068 0,039

S1.D3 <--- DisclosureS 1 23,33 ***

S1.D2 <--- DisclosureS 0,936 0,047 19,954 ***

S1.D1 <--- DisclosureS 1,055 0,044 23,886 ***

S1.A2 <--- AssessmentS 1 23,911 ***

S1.A1 <--- AssessmentS 1,064 0,057 18,732 ***

S1.T3 <--- TrustS 1 19,79 ***

S1.T2 <--- TrustS 1,154 0,057 20,182 ***

S1.T1 <--- TrustS 1,205 0,059 20,461 ***

S1.G2 <--- GenAttS 1 20,604 ***

S1.G1 <--- GenAttS 0,954 0,037 25,479 ***

S1.P2 <--- PIS 1 25,237 ***

S1.P1 <--- PIS 0,901 0,039 23,343 ***

S1.G3 <--- GenAttS 0,958 0,034 28,389 ***

AssessmentS <--- DisclosureS 0,792 0,044 17,925 ***

124