Transportation (2021) 48:3013–3035

https://doi.org/10.1007/s11116-020-10153-5

A novel perspective to enhance the role of TPB in predicting

green travel: the moderation of affective‑cognitive
congruence of attitudes

Xuemei Fu1

Published online: 3 January 2021

© The Author(s), under exclusive licence to Springer Science+Business Media, LLC part of Springer Nature 2021

Abstract
Successful and sustained interventions to obtain travel behavior change would be achieved
based on a thorough understanding of individual’s decision-making process on travel. To
narrow the gap between the theory of planned behavior (TPB) and green travel behavior,
this study extends classical TPB by accommodating the moderating effect of affective-
cognitive congruence of attitudes. Based on a cluster analysis utilizing affective and cog-
nitive attitudes towards private driving, four groups are obtained, each characterized by
different extent of feelings/emotions for private driving and of evaluations/beliefs about
the consequences on environment due to car-use. A multi-group structural equation model
analysis explicitly confirms the moderating role of affective-cognitive congruence of atti-
tudes, given that the structural relations between overall attitude, subjective norm, active
and passive PBC, green travel intention, as well as green travel behavior significantly differ
depending on the extent and direction of the congruence between affective and cognitive
attitudes. It is expected that the empirical findings might be useful for transport administra-
tors to maximize the effects of their limited resources and funds.

Keywords The theory of planned behavior · Moderation · Green travel · Affective-

cognitive congruence · Attitudes

Introduction

During the past several decades, the process of urbanization in China has been undergoing
at an unprecedented rate, and the number of private cars in cities significantly increases. The
average congestion delay index of 400 cities in China during the first half year of 2016 is 1.58
(Report on urban traffic in China in the first half year of 2016), indicating that almost one third
one-third of the total travel time are caused by traffic congestion. Besides, transportation sec-
tion, responsible for 23% of the global ­CO2 emissions due to fossil fuel combustion in 2012
(Change 2014), has been extensively recognized as one of the main causes for environmental

* Xuemei Fu
[email protected]
1
School of Management, Shandong University, No. 27, South Shanda Road,
Licheng District, Jinan 250100, China

13
Vol.:(0123456789)
3014 Transportation (2021) 48:3013–3035

pollution. Being the largest developing country, China has become one leading market for new
cars and the biggest C ­ O2 producer worldwide (Annual report on the environmental manage-
ment of motor vehicles in China 2016). In this situation, the widespread of the environmen-
tally- and socially- sustainable green travel modes (including walk, bicycle, and public trans-
port in the current study) seems to necessarily required since they are able to satisfy people’s
mobility needs while produce limited emissions and noise (Dudow 1998; Jeon et al. 2010).
Considerable efforts have been made on this object by transportation administrators, while the
outcome is unsatisfactory.
Successful and sustained interventions to obtain travel behavior change would be achieved
based on a thorough understanding of individual’s decision-making process on travel. The
theory of planned behavior (TPB) has been extensively applied for behavioral interpretation
due to its explicit definition of constructs, standardized measurements, and explanatory power
(Skar et al. 2008). In spite of these strengths, empirical results and interventions utilizing the
TPB are still inconclusive. One popular approach to improve the TPB is to accommodate
moderators in the conceptual framework, among which the affective-cognitive congruence of
attitude has got growing interest, i.e., the degree to which individual’s affective attitude about
particular object/behavior is consistent with his/her cognitive attitudes about the same object/
behavior. Norman (1975) showed that greater affective-cognitive congruence is associated
with stronger attitude-behavior relationship. Theoretically, a moderator refers to a variable that
significantly influences the strength and/or direction of the relationship between the other two
variables (Baron and Kenny 1986). In this regard, it is expected that the structural relations
among TPB predictors, intentions and the behavior of green travel would vary depending on
the affective-cognitive congruency of attitudes.
The prominent objective of this study is to systematically examine whether affective-cog-
nitive congruence of attitudes may enhance the predictive power of the TPB in green travel
behavior (i.e., the use of green transport modes), by moderating the relationships among TPB
predictors (i.e., overall attitude, subjective norm, and perceived behavioral control), green
travel intention and behavior. A two-step procedure is developed to empirically examine the
conception. Specifically, a cluster analysis based on the combination of affective and cognitive
attitudes is first performed to classify the sample population, which is designed to generate
groups with diverse extent of affective-cognitive congruence. Then, a multi-group structural
equation model (SEM) is implemented for inter-group comparison to finally identify the mod-
erating role of affective-cognitive attitude congruence. The findings are expected to be useful
for transport departments in promoting green travel modes and reducing car-use in an effective
way.
The remaining parts of this study are structured as follows. The next section provides a
theoretical background by reviewing previous studies related to the TPB, the affective-cog-
nitve congruence of attitudes and its moderating effect, based on which several hypotheses
are established. Then, sample data, indicators, and factors, are described in detail followed by
a presentation and discussion about results of the two-step analysis to examine the role played
by the affective-cognitve congruence of attitudes. The final section summarizes main conclu-
sions and policy implications.

13
Transportation (2021) 48:3013–3035 3015

Literature review and hypotheses

The success of interventions directed towards promoting green travel behavior and reduc-
ing private driving largely depend on understanding the motivations and obstacles underly-
ing behavioral decision.

The theory of planned behavior

Originated in psychology, the TPB has been applied in a variety of fields. The basic
idea underlying TPB is that people are likely to carry out a particular behavior if they
hold positive beliefs about the likely consequences and experiences associated with
the behavior (i.e., behavioral belief), about the normative recommendations and actual
behavior of important others (i.e., normative beliefs), and about the presence/lack of
capabilities, resources, and opportunities that promote/impede the performance of the
behavior (i.e., control belief) (Ajzen 1985; Conner et al. 1999). Generally speaking,
being agreed as an immediate antecedent of the target behavior, behavioral intention is
determined by three factors including attitude, subjective norm, and perceived behav-
ioral control (PBC), and PBC is also assumed to have a direct impact on behavior.
A comprehensive review suggests that this theory has been used to explain a wide
range of transport-related behavior, such as use of car/bike/public transport mode (Sig-
urdardottir et al. 2013; Zhang et al. 2016), adoption intention of clean cars (Zhang
et al. 2014), acceptance of sustainable transportation system (Khoo and Ong 2015),
and road violations (Castanier et al. 2013), etc. In spite of the availability of the TPB,
additional elements are usually incorporated to enhance its behavioral representation
and explanatory power. Donald et al. (2014) extend the TPB by including moral norm,
descriptive norm, as well as environmental concern within the domain of transport
mode choice. Fu and Juan (2017) develop a comprehensive model integrating the
TPB and the customer satisfaction theory to explore the public transit use, with the
direct impact of perceived service quality on both behavioral intention and behavior
confirmed, whilst PBC possesses no significant influence on behavior. To predict the
changes in bus ridership after the implementation of a bus pass program beyond that
accounted for by the original TPB factors, a set of variables including environmental
values, descriptive norm, car-use moral norm, problem awareness and felt responsibil-
ity are added (Heath and Gifford 2002).
More importantly, it is commonly acknowledged that certain variables might mod-
erate the relationships determined by the TPB, which “offers the possibility of extend-
ing our understanding of the basis of social behavior” (Conner et al. 2003). Within the
transport research domain, habit and PBC are the two most frequently examined mod-
erators. In the case of travel mode choice, the decision to use specific mode is likely
to be made automatically if related mode-use habit is strong (Klöckner and Matthies
2004). That is, the deliberate decision-making process as defined by the TPB is not
likely to happen since the structural relationships between TPB predictors, behavioral
intention, and the final choice behavior are significantly decreased. But if the habits
are weak, cognitive factors within the TPB would play significant roles in predicting
the mode choice decision. By the same token, since an attitude could be considered
as a judgement concerning one’s orientation towards particular subject, it is plausible
that an individual with attitude towards specific object characterized by high affective-
cognitive congruence might be labeled as ‘attitude schematic’ (Chaiken and Baldwin

13
3016 Transportation (2021) 48:3013–3035

1981). For these people, they are more resistant to counter-schematic information, and
manifest greater cross-situational consistency in behaviors (Markus 1977; Kuiper and
Rogers 1979), i.e., the moderating effect of affective-cognitive congruence of attitudes
introduced in what follows.

Affective‑cognitive congruence of attitudes and its moderating effect

Attitude is an important factor in explaining human behavior, and how to measure atti-
tude is an interesting topic, which has been extensively discussed. Hoffmann et al. (2017)
emphasized the importance to identify the complex structure of attitude in measurement
methods in obtaining effective interventions. According to the TPB, behavioral beliefs
would produces a favorable or unfavorable attitude towards the behavior (Ajzen 1985),
indicating that both affective (being operationalized with affective word pairs like good vs
bad or pleasant vs, unpleasant) and cognitive (primarily utilitarian aspect with behavioral
beliefs) dimensions should be included to characterize attitude, which has been extensively
mentioned (Yang and Yoo 2004). In social psychology, affective attitude refers to expres-
sions of feelings and emotions towards a psychological object or behavior, while cognitive
attitude is personal expressions of knowledge- and experience- based behavioral beliefs
about the object (Rosenberg 1960, 1968).
Generally, the congruence of attitudes is determined based on the discrepancy (or dis-
tinction) between affective and cognitive attitudes (Rosenberg 1968). Two approaches have
been primarily adopted to identify the affective-cognitive congruence of attitudes, both
based on the similarity between scores on affective attitudes and on cognitive attitudes
(Norman 1975; Schleicher et al. 2004). In the first approach, subjects are rank-ordered in
terms of their affective attitude scores, and then in terms of their cognitive attitude scores.
The level of affective-cognitive attitude for each subject is defined as the absolute value of
the difference between his/her positions in the two ranks. A higher absolute value of the dif-
ference corresponds to a lower affective-cognitive congruence of attitude, and vice versa.
Subjects are further classified into two groups by means of a median-split, i.e., a high and
a low congruency groups. The second approach was following the criteria in Lavine et al.
(1998) and directly split the sample into two groups. Specifically, the congruent group is
comprised of subjects whose scores of affective and cognitive attitudes are both above the
zero-mean (i.e., positive affective/cognitive attitudes), or below the zero-mean (i.e., nega-
tive affective/cognitive attitudes) simultaneously, whereas the incongruent group refers to
subjects whose scores are above the zero-mean in one dimension, and below the zero-mean
in the other, i.e., positive affective/negative cognitive attitudes, or negative affective/posi-
tive cognitive attitudes.
Congruence between affective and cognitive dimensions of attitudes contributes to iden-
tify attitude strength and stability, i.e., the degree to which attitude about specific object
is influential and durable (Petty and Krosnick 1995). More importantly, the congruence
between affective and cognitive components of attitudes is generally considered to be an
important moderator within the framework of TPB. It is argued by Sanchezgarcia and
Batistafoguet (2008) that situations not taking into account the moderation effect of the
congruence between the affective and cognitive components of attitudes when applying the
TPB to predict particular behavior, could only obtain a kind of average main effects. To put
it simply, the variance in the behavioral intention explained by the TPB when accounting
for the moderating influence of congruence of affective-cognitive attitudes is believed and
identified to be larger than that when the moderation effect is neglected.

13
Transportation (2021) 48:3013–3035 3017

According to previous studies (see, e.g., Rosenberg 1960, 1968), people with high affec-
tive-cognitive attitude congruence are likely to have “well-thought-out” attitudes, reflecting
a relatively stable underlying disposition towards the attitude object, whereas those with
attitudes characterized by a low congruence are likely to have “poorly articulated” atti-
tudes, reflecting an unstable disposition. In this regard, individuals with high affective-cog-
nitive congruence of attitude are likely to exhibit more stable attitudes, who are expected
to possess stronger attitude-behavior relations than those with low affective-cognitive con-
gruence (Norman 1975; Rosenberg 1968). Skår et al. (2008) found the affective-cognitive
congruence of attitudes moderates both the PBC-behavior and intention-behavior relation-
ships when examining undergraduates’ physical activity, where higher level of congruence
enhances predictive power of the TPB. Further, a meta-analysis of Cooke and Sheeran
(2004) confirmed the moderating effects of affective-cognitive congruence on cognition
(including attitude, subjective norm, and PBC)-intention/behavior associations.
Regarding the green travel behavior, it is of significant importance to know how the two
dimensions of attitudes, i.e., the emotion-based affective component and the knowledge-
based cognitive component, are integrated when people are faced with green travel deci-
sion (Beige and Axhausen 2017), as well as to know how they affect the psychological
decision-making process defined by the TPB. As far as we know, the subject has not been
researched yet, which is expected to be practical for promoting green transport modes.

Conceptual framework and hypotheses

A great number of studies have been conducted to investigate the psychological determi-
nants of green travel behavior, including walking, cycling, riding public transport, as well
as reducing driving. Jia et al. (2018) used binary logistic model to study the influences of
attitudinal and low-carbon factors on the choice of public transportation, where low-carbon
knowledge and instrumental attitude are two strongest predictors. Ru et al. (2018) found
that the interaction between experiential attitude and norms (i.e., the composite variable by
multiplying the two single variables) are positively associated with green travel intention,
whereas the interaction effects between instrumental attitude and norms on green travel
intention are negative. More importantly, substantial attention has been paid to accom-
modate the self-selection effect in order to better predict green travel behavior. That is,
people who prefer walking and biking are likely to self-select into neighborhood where
these modes are well supported by the infrastructures and land-use. Ramezani et al. (2018)
suggested that certain attitudes such as sensitivity to peer pressure regarding the value
attributed to sustainable transportation and transit preference could play significant role
in increasing sustainable mode choice even in the neighborhoods with specific physical
restrictions.
Previous studies have shown that the decision of green travel is a complex process influ-
enced by multiple factors. Nevertheless, it is almost impossible that influences of all factors
could be figured out in single research. This study attempts to enhance the predictive power
of TPB in green travel behavior by accommodating the moderating effect of the affective-
cognitive congruence of attitudes. Based on the above literatures, three sets of hypotheses
are established, which will be empirically examined following the conceptual framework in
Fig. 1.

13
3018 Transportation (2021) 48:3013–3035

Overall attitude H1a Intention

H1a
H2: Moderating
Subjective norm

H1a H1b

Active perceived H1a

behavioral control H2: Moderating
H1b

Passive perceived
behavioral control
H1b Green travel

Fig. 1  Conceptual framework about the TPB to predict green travel

H1a TPB predictors including overall attitude to reduce driving, subjective norm about
green travel modes, as well as PBC over green travel modes are assumed to positively
influence people’s intention to use green travel modes.

H1b Both PBC over green travel modes and intention to use green travel modes are
assumed to have significantly positive influences on people’s green travel behavior.

H2 All relationships in H
­ 1a and H
­ 1b are assumed to differ between groups with different
affective-cognitive congruence of attitudes.

Sample data and variables

Survey and sample description

An on-site field survey was conducted in Zhenjiang City, Jiangsu Province of China from
June 22nd to 26th, 2015, to collect information about people’s attitude, opinions, and
behavior specific to green travel. Three locations including the high-speed railway station,
old railway station, as well as coach station, were chosen since a relatively large number
of people were easily accessible there and the sample variety could be better ensured. Par-
ticipants were randomly contacted by trained interviewers. They were asked to fill in the
survey using provided cellphones, which would cost them for about five minutes. Finally,
2941 valid questionnaires were obtained and used for analysis. Table 1 presents the statisti-
cal description of their socio-economic characteristics.
The selected surveying sites suggest that people from other cities are also potential
respondents. Actually, participants mainly come from Zhenjiang and several nearby cit-
ies, e.g., Shanghai, Wuxi, and Suzhou, which are all located in the Yangtze Delta Region.

13
Transportation (2021) 48:3013–3035 3019

Table 1  Statistical description of the sample characteristics

Socio-economic characteristics Percentage Socio-economic characteristics Percentage

Age Gender
Below 26 36.8 Male 62.2
26–35 30.5 Female 37.8
36–45 19.4
Above 45 13.3
Income (¥/month) Educational status
Less than 2001 23.4 High school or below 26.8
2001–4000 32.7 Other colleges 26.6
4001–6000 26.7 Undergraduate or above 46.6
More than 6000 17.3

At the time when the survey was conducted, 1 dollar was equal to about ¥6.22

Such case is expected not to significantly bias the findings since the primary objective is
to enhance the TPB in understanding green travel with the moderation of affective-cogni-
tive congruence of attitudes incorporated; instead of predicting the aggregated tendency of
population’s behavior, in which situation the sample representativeness for region of inter-
est is necessarily required.
There are several points to be explained about the background of these cities. First, it is
known that most of the cities within this area have highly-developed economy and well-
provided urban infrastructures. Hence, it is reasonable to believe that nearly all participants
live in high-density cities with direct and convenient access to public transport, excellent
walking and bicycling environment. Hence, the built environment is considered to not con-
siderably vary across participants. Second, average car ownership of these cities is at the
level of about 0.25 per capita. Assuming three persons in a family, it is expected that most
participants (representing his/her family) have access to or could at least, afford to buy a
car due to the highly-developed economy. In addition, scholars have argued that it is the
perception of transport modes that affects people’s travel behavior to a larger extent, rather
than the mere existence of those modes (see Kuhnimhof et al. 2006; Pronello and Camusso
2011).
Next, given the annoying traffic environment that mainly results from the increasing
number of private cars, a wide range of administrative measures have been taken by local
transport authorities. For example, only a limited number of license plates are provided
annually by way of auctioning in Shanghai in order to impose strict restrictions on new
cars. It should be mentioned that a new license plate in Shanghai cost almost ¥100,000,
almost equaling to the money on an ordinary family car. Particular measures are also
implemented in all these cities to promote public transport, such as well-planned/increas-
ing routes and stops to provide convenient access, multiple concessionary fares, keeping
improving services and facilities.
From a theoretical perspective, it has been extensively confirmed in existing literatures
that subjective environmental concern significantly contributes to people’s intention to
travel by environmentally-friendly modes and their actual behavior (e.g., Yang et al. 2020;
Prillwitz and Barr 2011). As being obtained from a statistical examination on sample data
(not presented for brevity), people in the above cities have individually developed a cer-
tain degree of environmental consciousness due to a realization of the worsening living,

13
3020 Transportation (2021) 48:3013–3035

traveling, and social environment surrounding them, which are believed to at least make
them consider the negative influences on environment due to private driving, or even
further make them reduce driving. Nevertheless, there are currently limited administra-
tive strategies or publicizing advertisements implemented to enhance people’s subjective
environmental concern. In this regard, effective administrative efforts and policy instru-
ments are necessarily required to improve people’s environmental concern in order to
greatly encourage them to reduce driving, as well as to use green travel modes. All the
above elements jointly justify the importance to account for the role of environmental con-
cern, which, in the current study, is developed to as the cognitive dimension of attitude
towards driving, i.e., functional knowledge and belief about the consequences on environ-
ment due to private driving, or simply environmental concern. That is, the accommodation
of environmental concern helps to assess whether congruence of attitudes towards driv-
ing between affective attitude (i.e., car affection) and cognitive attitude (i.e., environmental
concern due to car-use), as show in Fig. 2, with other factors would have any influence on
green travel behavior, based on which effective measures and strategies could be specifi-
cally designed. As far as we know, this has never been done in previous studies.

Factors and indicators

The survey was designed with three sections including personal socio-economic informa-
tion, environmental protection regarding garbage disposal, and green travel predisposition.
For brevity, only survey questions used are introduced here. The factors included in the
model framework are defined as follows. Measurement indicators specific to each factor
and results of the indicator-factor loadings based on the exploratory factor analysis (EFA)
are presented in Table 2. Several aspects should be explained in advance.
Above all, PBC is classified into two dimensions based on their measurement indica-
tors, i.e., active and passive PBC. In the present study, PBC over both car and non-car
modes (i.e., bicycle and public transport) are considered. Specifically, PBC over public
transport and bicycle could be regarded as the motivator for the choice decision of green
travel modes, which pulls the active green travel behavior in view of the favorable condi-
tions. In contrast, PBC over car-use is the obstacle resulting from restrictions and traffic

Affective attitude
(Positive feelings and emotions towards private driving)

Group-2 33.2% Group-4 17.0%

+
Positive/Incongruent Negative/Congruent
Cognitive attitude Cognitive attitude
(With knowledge and beliefs about (Lack of knowledge and beliefs
+ -
the consequences on environment about the consequences on
due to private driving) environment due to private driving)
Group 3 23.2% Group-1 26.6%
-
Positive/Congruent Negative/Incongruent

Affective attitude
(Negative feelings and emotions towards private driving)

Fig. 2  Graphical representation of the cluster analysis results

13
 Table 2  Results of the EFA
Factor Items Factor loading Cronbach’s alpha Variance
explained
(%)

Affective attitude (AFF) AFF1 Driving to work is convenient and makes my schedule flexible. 0.768 0.686 61.67
AFF2 Driving to work represents social status and enhances my happiness. 0.758
AFF3 Driving to work saves a lot time and improves the quality of my life. 0.828
Cognitive attitude (COG) COG1 Do you think the use of cars influences the environment? 0.563 0.555 52.99
Transportation (2021) 48:3013–3035

COG2 Reducing driving is beneficial for improving urban traffic congestion and environment 0.799
pollution.
COG3 Regarding environmental protection, green travel modes are more effective than other 0.796
modes.
Overall attitude ATT​ It is required and responsible for each resident to reduce the use of private cars. 0.772 – –
Subjective norm (SN) SN1 People around me are used to travel by green modes. 0.882 0.714 77.56
SN2 People around me hope that I travel by green modes. 0.882
Passive PBC (PPBC) PPBC1 The restriction on car purchase and the higher use cost make me inclined for green 0.848 0.609 71.91
travel.
PPBC2 The problems caused by private cars (e.g., congestion, parking, and accidents) make me 0.848
inclined for green travel.
Active PBC (APBC) APBC1 The promotion of public bicycles makes me inclined for green travel. 0.867 0.670 75.22
APBC2 The development of public transport makes me inclined for green travel. 0.867
Intention (INT) INT1 I intend to travel by green modes under most situations. 0.825 0.777 69.21
INT2 It is highly likely that I will travel by green modes. 0.856
INT3 I have a strong willingness to travel by green modes. 0.815
Green travel (GT) GT1 Whether or not commute by green modes. 0.767 0.247 58.76
GT2 I usually travel by green modes. 0.767

(1) Most of the indicators are measured on a 5-point Likert scale, i.e., 1-strongly disagree, 2-disagree, 3-neutral, 4-agree, and 5-strongly agree, except for COG1 (1-no influ-
ence, 2-certain influence, and 3-seriously pollute environment) and GT1 (0-no, 1-yes). (2) The measurement scales are reversed with regard to the original scales in the survey
to ensure that the greater number indicates a higher level of approval, which is consistent with human’s common sense
3021

13
3022 Transportation (2021) 48:3013–3035

conditions, which prevents private driving and is expected to push people to turn to green
travel modes passively. Briefly, this action is performed with the main purpose to obtain
more targeted strategies, i.e., choosing between restrictions on car-use and improvement
of non-car journey. Furthermore, active and passive PBC could be treated as approximates
of the traditional perceived autonomy and perceived capacity as two distinct components
of PBC (Chan and Fishbein 1993), respectively, where the former refers to the degree of
control over green travel modes, while the latter is mostly characterized as the judgement
of difficulty of private driving (Trafimow et al. 2002; Fishbein and Ajzen 2010). Both com-
ponents have been confirmed to predict intention and behavior independently (Povey et al.
2000; Trafimow and Trafimow 1998).
Second, the affective component of attitude is generally regarded as containing the
encoded sensations, emotions, and feelings with the attitude object, whereas the cognitive
component contains the encoded attributes and beliefs about the object. Following this def-
inition, in the current study, affective attitude is defined as the private driving affection, i.e.,
appreciating the satisfaction and feelings provided by driving, and cognitive attitude refers
to thoughts and evaluations about the behavioral consequences on environment from travel
i.e., benefits from green modes and damages from private driving, or briefly environmental
concern.
Finally, originally put forwards by (Rosenberg 1960), affective-cognitive congruence
is actually measuring the extent of correspondence between attitude-relevant feelings and
beliefs. In this regard, affective-cognitive congruence is independent of the overall atti-
tude, which measures the overall evaluative proposition to like or dislike, and is separately
included in the model.

• Affective attitude: The feelings and emotions that individuals possess with private driv-
ing.
• Cognitive attitude: The belief and knowledge held by individuals about the behavio-
ral consequences on environment from reducing private driving and increasing green
travel modes.
• Overall attitude: Individual’s opinion and judgement on private driving.
• Subjective norm: The behavior and expectations from important others with regard to
green travel.
• Passive PBC: The judgment of the difficulty to drive (i.e., “pushed control”). Specif-
ically, people are assumed to possess a high PBC resulted from the restrictions and
enforcement imposed on driving, which might make them perceive difficulty and low
capacity with driving. Hence, they are likely to be forced to turn to non-car modes in a
passive way.
• Active PBC: The perceived control over green travel brought about by the improve-
ment of transport environment with non-car modes (i.e., “pulled control”). People with
a stronger active PBC would perceive the use of green travel modes due to the user-
friendly bicycling and public transport environment, such that they are likely to be
more proactive in the green travel behavior.
• Intention: The strength of individual’s inclination for performing green travel behavior.
• Behavior: The frequency of individual’s actual green travel in real life.

The values of Cronbach’s alpha suggest a satisfying internal consistency regarding most
indicators, except for the two indicators for the behavior in question. It is probably because
of their different scales. The GT1 is actually binary measuring whether participants com-
mute by green mode, while the GT2 is measured based on a 5-point scale. Since GT1 and

13
Transportation (2021) 48:3013–3035 3023

GT2 quantify the green travel from revealed and stated dimensions, respectively, it is deter-
mined that both indicators are remained to get an elaborate construct.
In spite of the focus on green travel modes, some attitude factors and their indicators are
associated with private driving (or car-use). This is done due to particular considerations
as follows.
First of all, it is assumed that people’s attitudes towards green travel modes are likely to
be shaped by diverse meanings that go beyond yet relate to green travel modes. On the one
hand, these meanings are expected to be environmentally-relevant due to the eco-friendly
nature of green travel modes. On the other hand, the sense of satisfaction and achievement
provided by car-use has been recognized as the primary hindrance that prevents the green
travel behavior. For example, Black et al. (2001) showed that the parents’ car-use atti-
tudes including environmental awareness and car-centeredness influence children’s walk-
ing to school, while Mcmillan (2007) reported the importance of parents’ beliefs about
the convenience of driving in children’s active school commuting. In this situation, it is
believed that the intention and decision to travel by green modes is influenced by the bal-
ance between the affective car centeredness from private driving - cognitive environmental
awareness to reduce car-use and the desire for green travel modes (Black et al. 2001).
Second, Bradley (2013) argued that there is no common knowledge about the definition
of which is and which is not an “attitudinal measurement”. Tagler and Cozzarelli (2013)
also acknowledged that their measures for the two dimensions of attitudes might not be
fully distinct measures of the affective and cognitive components of attitudes. It is sug-
gested that the use of measures developed for other attitude objects would be more inform-
ative (Tagler and Cozzarelli 2013). In the transportation research, both car-related and non-
car-related attitudes have been considered and confirmed to have significant influence on
the choice behavior of green travel modes.
Finally, it should be noted that these measures in Table 2 cannot be assumed as ‘pure’
measures of affective and cognitive attitudes. Instead, the measures for affective attitudes
should be considered as more ‘affective’ than ‘cognitive’, and the measures for cognitive
attitudes as more ‘cognitive’ than ‘affective’. In other words, these measures are practically
adopted because they seem to be the best ones available that satisfy the requirement and
objective of this study.
Hence, it is expected that the specific attitudes and their indicators would contribute
to the explanatory power of TPB on green travel behavior in view of the salient com-
peting pattern between private cars and green travel modes, along with the abundant
measurements.

Results and discussions

The first part will present how groups with different affective-cognitive (in)congruence of
attitudes are derived, then structural relationships between TPB predictors, green travel
intention and behavior are compared to identify whether significant differences exist across
groups, i.e., whether the affective-cognitive congruence of attitudes moderates the deci-
sion-making process of green travel behavior defined by the TPB.

13
3024 Transportation (2021) 48:3013–3035

Results of cluster analysis

Factor scores of affective and cognitive attitudes obtained from the EFA are used as
the basic variables for cluster analysis. The criteria proposed in Lavine et al. (1998),
which has been specifically described in the review section, was considered such that
four sub-groups would be sufficient. Instead of directly classifying respondents based
on the sign (+ or −) of affective and cognitive attitudes, K-means clustering method
is selected and conduced in SPSS 23.0 in order to obtain a higher homogeneity within
groups and higher heterogeneity across groups. The clustering results are recorded in
Table 3, which shows the sizes and final cluster centers (i.e., scores of affective and
cognitive attitudes located at the center) of the four groups. For example, 26.6% of the
respondents fall into group-1, and their affective and cognitive attitudes are distributed
around the point (− 0.69, − 0.97). It is satisfying that the four groups exhibit anticipated
features of affective and cognitive attitudes, respectively, which justifies the four-group
solution. The following comparisons across groups regarding their psychological deci-
sion-making processes defined by TPB, also supports the rationality of this clustering.
It is worth paying attention to results of the cluster analysis. As discussed above,
both affective and cognitive components are attitudes about private driving. On the one
hand, the higher the score of affective attitude, the more positive feelings people have
on private driving. Hence, the positive affective attitudes in groups 2 and 4 suggest that
these people enjoy the pleasure and satisfaction with driving. On the other hand, the
higher the score of cognitive attitude, the stronger belief people have about the conse-
quences on the environment from reducing the behavior of driving and increasing green
travel, or their environmental consciousness and negative evaluations on cars. For exam-
ple, the positive cognitive attitudes in groups 2 and 3 show that these members might
have sufficient knowledge and experience to evaluate the consequences on environment
due to the behavior of owing and using private cars (Nilsson and Kuller 2000). In other
words, group-2 and group-3 members are concerned with the environmental issues and
negative towards car-use, while positive cognitive attitudes towards private driving are
held by members in groups 1 and 4. Briefly, the same signs (++ or −) between scores of
affective and cognitive attitudes in this study correspond to incongruence, while differ-
ent signs (+ − or − +) are related to congruence in the current study.
It should be noticed that the use of affective-cognitive congruence and incongruence
is not sufficient to identify and elaborate differences across the four groups. Since the
moderator role of environmental concern (i.e., belief about the consequences on envi-
ronment due to performing specific behavior) has been examined in other behavioral
domains (see, e.g., Bissing-Olson et al. 2013), intrinsic nature of the cognitive attitude,
i.e., people’s belief about the behavioral consequences on environment due to (reduc-
ing) private driving, is also taken into account in the following analysis. Specifically,

Table 3  Results of the cluster analysis

Groups Group-1 Group-2 Group-3 Group-4

Proportion (%) in the sample 26.6 33.2 23.2 17.0

Affective attitude (car affection) − 0.69 0.74 − 0.90 0.87
Cognitive attitude (environmental con- − 0.97 0.67 0.79 − 0.87
cern with car-use)

13
Transportation (2021) 48:3013–3035 3025

the four groups obtained are labeled based on the congruence between their affective
and cognitive attitudes (congruent or incongruent), together with the underlying envi-
ronmental concern (positive or negative). As illustrated in Fig. 2, groups 1, 2, 3, and
4 are characterized as being negative/incongruent positive/incongruent, positive/con-
gruent, and negative/congruent, respectively. Taking group-1 as an example, a nega-
tive affective but positive cognitive attitude towards private driving are observed, i.e.,
with affective-cognitive incongruence of attitudes, while its positive cognitive attitude
towards private driving corresponds to a low environmental concern, i.e., negatively and
indifferently about the behavioral consequences on environment due to using private
cars. Hence, ‘negative/incongruent’ is designated as the unique profile of group-1.

Multi‑group statistical description

Before moving to the moderation of affective-cognitive congruence of attitudes in TPB

predicting green travel, socio-economic attributes and psychological profiles (i.e., means of
related indicators in the last six rows of Table 4) of the four groups are summarized.

Table 4  Multi-group statistical description

Characteristics Group-1 Group-2 Group-3 Group-4
Negative/Incon- Positive/Incon- Positive/Congruent Negative/
gruent gruent Congru-
ent

Gender
Male 37.7 36.9 42.7 33.1
Female 62.3 63.1 57.3 66.9
Educational status
High school or below 29.6 20.6 29.8 30.6
Other colleges 29.2 26.6 21.5 29.1
Undergraduate or above 41.2 52.8 48.7 40.3
Age
Below 26 35.3 39.8 37.2 32.7
26–35 29.5 31.4 30.0 31.2
36–45 20.8 17.2 19.1 21.8
Above 45 14.5 11.6 13.8 14.3
Income (¥/month)
Less than 2001 23.1 24.7 26.0 17.8
2001–4000 34.5 30.6 34.6 31.5
4001–6000 28.3 25.2 24.7 29.6
More than 6000 14.2 19.5 14.8 21.2
Overall attitude 3.57234 3.9414 4.0414 3.70123
Subjective norm 3.23234 3.5313 3.3912 3.501
Passive PBC 3.27234 3.71134 3.60124 3.48123
Active PBC 3.43234 3.7614 3.7614 3.62123
Intention 3.39234 3.7614 3.8214 3.52123
Green travel 2.08234 2.23134 2.33124 2.10123

13
3026 Transportation (2021) 48:3013–3035

Pairwise Pearson’s Chi square test is performed to examine whether the categori-
cal socio-economic variables significantly differ across groups. Limited differences are
obtained, except for the educational status, which depends on the degree of congruence
between affective and cognitive attitudes towards private driving. Hence, the four groups
possess a similar profile regarding gender, income level, and age distributions. Analysis
of variance (ANOVA) is used to compare the continuous factor means. Numbers in super-
script suggest from which groups the current one is significantly different about particular
factor. For example, a “234” superscript of overall attitude in group-1 shows that this factor
differs significantly between Negative/Incongruent group and other groups. All differences
are significant at p value ≤ 0.05. In contrast to socio-economic attributes, most of the factor
means greatly vary across groups. To put it another way, the affective-cognitive (in)congru-
ence of attitudes is closely associated with the level of overall attitude, subjective norm,
active and passive PBC, intention, and green travel behavior, which, in a certain respect,
supports the current analysis.

Moderation of the affective‑cognitive congruence

The moderation of affective-cognitive congruence is to be investigated by examining dif-

ferences about the structural relationships across groups. Before comparison, the invari-
ance of measurement weights (or indicator-factor loadings) must be required (Deng et al.
2005). The conceptual framework in Fig. 1 is used as the basic modeling structure, which
is applied separately to each of the four groups. The goodness-of fit statistics in Table 5
suggest an excellent fit to all the four single models (Hair et al. 2006; Bentler 1980). Then,
two sets of multi-group SEMs are estimated for the four groups, among others, Model 1 is
a free model with no constraints on coefficients and Model 2 is a restricted model with des-
ignated equal measurement weights across the four groups. The χ2-difference test between
Model 1 and 2 (i.e., p-value = 0.19) indicates no significant differences between them. As a
consequence, the indicator-factor loadings are invariant across the four groups such that the
structural relations could be directly compared (Bentler 1988).
In what follows, estimation results of the measurement model (i.e., relationships
between factors and their indicators) are not presented since sufficient reliability has been
confirmed in Table 2. Besides, these factor-indicator loadings provide limited explanations
about structural relationships between TPB predictors, green travel intention and behavior,
which will be discussed in detail, with a particular focus on the differences across groups.
Figure 3 displays the standardized estimates of the multi-group SEM. Note that (1) All the
coefficients presented (i.e., solid line) are significant at the p-value ≤ 0.10, while the dotted

Table 5  Summary of goodness- Indices Negative/ Positive/ Positive/ Negative/

of-fit statistics Incongruent Incongruent Congruent Congruent

χ2/df 1.890 1.974 1.611 2.453

GFI 0.983 0.985 0.983 0.965
AGFI 0.968 0.973 0.968 0.935
NFI 0.960 0.974 0.971 0.918
CFI 0.980 0.987 0.989 0.949
RMSEA 0.033 0.032 0.031 0.055

13
Transportation (2021) 48:3013–3035 3027

Intention Intention
Overall attitude Overall attitude
(R2=0.64) 0.144
(R2=0.55)
0.2834 0.244

Subjective norm Subjective norm

0.462 0.33134

0.424 0.493 0.374 0.51

Active perceived Active perceived
behavioral control behavioral control
0.21 0.273

Passive perceived Green travel Passive perceived Green travel

behavioral control (R2=0.40) behavioral control (R2=0.38)

(a) Group-1: Negative/Incongruent (b) Group-2: Positive/Incongruent

Intention Intention
Overall attitude 0.2014 Overall attitude
(R2=0.63) (R2=0.99)

0.171

Subjective norm Subjective norm

0.552 0.722

0.284 0.261 0.94123

Active perceived Active perceived
behavioral control behavioral control
0.552

Passive perceived Green travel Passive perceived Green travel

0.182
behavioral control (R2=0.62) behavioral control (R2=0.57)

(c) Group-3: Positive/Congruent (d) Group-4: Negative/Congruent

Fig. 3  Results of the Multi-group SEM analysis

lines represent the non-significant effects. (2) The subscripts denote the groups from which
the effect in the current group is significantly different at the p-value ≤ 0.10.
At the overall level, significant differences are explicitly observed, justifying the moder-
ating role of affective-cognitive congruence of attitudes in the green travel decision-making
process as being defined by the TPB. And it is explicitly observed that all the designated
relationships are significant only in group-3 with negative and congruent affective-cogni-
tive attitudes towards private driving. Hence, hypothesis ­H2 is empirically justified. The
different degrees of explained variances for intention and green travel behavior across
groups support the moderator role of affective-cognitive congruence from another perspec-
tive. Specifically, the percentage of explained variance in the green travel behavior is sig-
nificantly larger in the congruent groups vis-à-vis the incongruent groups. An additional
examination on the situation when the moderating effect is not considered suggests that the
percentage of explained variance in this situation roughly falls into the range of the results
obtained from the incongruent and congruent situations. According to Tobisu et al. (2008),
such finding is because only average main effects related to the congruent and incongru-
ent populations together could be obtained when the moderating effect is not specifically
accommodated.

13
3028 Transportation (2021) 48:3013–3035

Due to the considerably moderating effect of affective-cognitive congruence of atti-

tudes, Hypotheses ­H1a and ­H1b are partly supported since some effects are not significant in
particular models, which are discussed and explained in what follows:

(1) First of all, effects of the four predictors (i.e., overall attitude, subjective norm, active
and positive PBC) on green travel intention are evaluated.

Overall attitude The positive effect of overall attitude on green travel intention is sig-
nificant in the case where a strong environmental concern is found, regardless of the (in)
congruence between affective and cognitive attitudes. That is, only when people have suffi-
cient knowledge and belief about the negative consequences on environment due to private
driving behavior, will his/her attitude about reducing car-use lead to the specific intention
to use green travel modes. Furthermore, the attitude-intention relation is more likely to be
moderated by the cognitive attitude since deviations are found to exist between groups with
opposite environmental concern.
Subjective norm The greater effects of subjective norm on intention in groups with
incongruence than that in groups with congruence reveal that this relation is moderated
by the congruence of attitude. In other words, when people’s affective attitudes about the
pleasure and satisfaction with private driving are not consistent with their cognitive atti-
tudes about the negative consequences on environment due to private driving, they are
more likely to be influenced by important others to develop green travel intention. If a
person has both positive cognitive-and-affective attitudes towards private driving, how the
people around him/her behave or what they wish is found to have no influence on his/her
green travel intention.
Perceived behavioral control Both passive and active PBC positively enhance the green
travel intention. Specifically, individuals who feel with more control over green travel
behavior as a result of either being “pulled” by the improvement in public transport or
being “pushed” by policy regulations and real obstacles with private driving, are more
likely to develop a stronger intention for green travel. Nevertheless, this effect significantly
varies among groups. The effect of active PBC on green travel intention in group-2 with
positive affective yet negative cognitive attitudes towards private driving (a strong environ-
mental concern), is significantly smaller than that in other groups, while the effect of pas-
sive PBC on green travel intention in group-4 with negative affective but positive cognitive
attitudes towards private driving (a weak environmental concern), is significantly larger
than those in other groups. Taking the latter finding for illustration, for individuals with
positive affective-and-cognitive attitudes towards private driving, once they feel having
reduced control over driving because of either policy restrictions or real difficulties, there
is a great chance that they would have stronger inclinations for green travel.

(2) Next, the direct influences of PBC and intention on green travel behavior are elaborated.

Perceived behavioral control It is worth noting that passive PBC has a significantly
positive influence on the target behavior only in the 3rd group, i.e., with negative affective-
and-cognitive attitudes towards private driving (a strong environmental concern). From the
perspective of transportation management, government regulations and restrictions on car
purchasing, driving, as well as parking will do play a role in promoting the use of green

13
Transportation (2021) 48:3013–3035 3029

travel modes only for those owning both negative affective-and-cognitive attitudes towards
private driving. A similar pattern is also observed with respect to the effect of active PBC
on green travel behavior. Practically, with the improvement on services and facilities of
public transport and cycling, green travel behavior of individuals with negative affective-
and-cognitive attitudes towards private driving is to be the most encouraged among all
groups.
Green travel intention Only in group-4, the green travel intention has no significant
influence on green travel behavior. As expected, for people with positive affective-and-cog-
nitive attitudes towards private driving, even they have a strong intention for green travel, it
is unlikely that this intention will be actually implemented.
Notably, estimation results in Fig. 3 simply presents the direct relationships between
variables. It would be useful to identify which factor has the greatest influence on the green
travel behavior in question, where both direct and indirect effects should be considered.
For example, the total effect of passive PBC on green travel behavior is the sum of its
direct effect on green travel behavior and its indirect effect through green travel intention.
Hence, the most prominent predictor in each group could be explicitly known. Notably,
total effects of all TPB predictors and intention in the Negative/Congruent group are not
available in Table 6, since none of the direct determinant of green travel behavior has sig-
nificant influence, as shown in Fig. 3.
For the two groups with affective-cognitive incongruence of attitudes, namely, groups
1 and 2, it is the behavioral intention that has the largest effect on green travel behav-
ior, followed by active PBC, passive PBC, subjective norm, and finally overall attitude.
As for group 3 with affective-cognitive congruence of negative attitudes, active PBC
plays the most important part in explaining green travel behavior. Consistent with the
above findings, measures and strategies to improve the travelling environment with green
modes, which are expected to enhance people’s control over the public transport, walk-
ing, and bicycling, will be more effective in facilitating the green travel behavior of those
with affective-cognitive congruence of negative attitudes towards private driving. Since
intention has no significant influence on green travel behavior in the 4th group, overall
attitude and subjective norm which affect green travel behavior completely and indirectly
through behavioral intention, have no effects the target behavior, either. In this regard, it
is considered to be rather difficult for transport administrators to promote the green travel
behavior of population who genuinely enjoy driving and at the same time are lack of
environmental concern, i.e., not conscious about the negative consequences on environ-
ment due to private driving.

Table 6  Standardized total Negative/ Positive/ Positive/ Negative/

effects on green travel behavior Incongruent Incongru- Congruent Congru-
in the four groups ent ent

Overall attitude – 0.07 0.05 –

Subjective norm 0.14 0.12 0.04 –
Passive PBC 0.21 0.19 0.25 –
Active PBC 0.44 0.40 0.69 –
Intention 0.49 0.51 0.26 –

13
3030 Transportation (2021) 48:3013–3035

Conclusions and implications

Main conclusions

According to previous studies, two sources are available for the formation and expression
of attitudes. That is, the affective component associated with personal emotions and feel-
ings (e.g., pleasures), along with the cognitive component reflecting evaluations based on
previous knowledge, belief, and opinions about the advantages and disadvantages of an
object or behavior (Millar and Tesser 1989; Crites et al. 1994). Rosenberg (1968) found
that people whose attitudes are characterized by higher affective-cognitive congruence tend
to have greater attitudinal stability and resistance to persuasion from external environment,
following which congruence of affective and cognitive components of the attitudes as a
moderator on particular intention/behavior have been extensively examined and acknowl-
edged. With aim to enhance the explanatory power of TPB on green travel behavior, this
study extends the classical TPB by accommodating the moderating role of the affective-
cognitive congruence of attitudes. Based on a combination of cluster analysis and multi-
group SEM comparison, the main findings are summarized as follows.
First of all, four groups are obtained by accounting for both affective and cognitive
attitudes, where the former refers to individual’s feeling and emotions about private
driving, and the latter is defined as personal knowledge and evaluations about the con-
sequences on environment due to (reducing) private driving behavior. The four groups
are identified to be with significantly different profiles as negative/incongruent, posi-
tive/incongruent, positive/congruent, and negative/incongruent, where the first word
describes the environmental concern (i.e., positive) or lack (i.e., negative), and the sec-
ond word is the affective-cognitive (in)congruence of attitudes. For example, the group
with positive/incongruent feature possesses an affective-cognitive incongruence of atti-
tudes towards private driving. Specifically, these people appreciate the satisfaction and
pleasure provided by private driving, but at the same time, they have developed some
kinds of consciousness and belief about the negative consequences on environment due
to private driving, which leads to the incongruence and environmental concern.
Second, the multi-group comparison explicitly reflects the moderating role of affec-
tive-cognitive congruence in relations between TPB factors and green travel behavior.
Only in the situation with positive environmental concern, overall attitude to reduce pri-
vate driving has a significant effect on intention for green travel. Effect of subjective
norm on behavioral intention also greatly varies. When affective and cognitive attitudes
towards private driving are incongruent, subjective norm tends to play a great role in
promoting the intention for green travel. The active PBC-intention relation is the weak-
est when individuals have sufficient knowledge and belief about the negative conse-
quence on environment due to private driving (i.e., a strong environmental concern), but
at the same time appreciate the pleasure and fulfillment with private driving. In other
words, improvement of public transport service and infrastructures plays a minor role
in promoting their green travel intention. In contrast, results about the passive PBC are
intriguing. When individuals have negative and congruent attitudes towards green travel
(i.e., great affection for private driving and a lack of environmental concern), the effect
of passive PBC on green travel intention but not behavior is observed to be the largest
among the four groups. This finding might be a little bit unexpected at a first glance.
Nonetheless, it could be understood from the other view. When people have a strong
inclination and appreciation for private driving, and are indifferent to the negative

13
Transportation (2021) 48:3013–3035 3031

consequences on environment due to private driving, they are likely to believe that eve-
rything related to private driving will be achieved with limited difficulties. At this point,
they would easily get frustrated when being faced with obstacles resulted from regula-
tions and restrictions on private driving.
Next, the direct influences from intention, active and passive PBC vary from group
to group. With incongruence between cognitive and affective attitudes, the intention-
behavior relation is observed to be stronger. The conflicting attitudes towards private
driving promote the realization of intention for green travel behavior once the green
travel intention is necessarily acquired. Only when people have both negative affective
and cognitive attitudes towards private driving, active and passive PBC would signifi-
cantly contributes to the green travel behavior. However, none of the above three predic-
tors is significant in the case with both positive affective and cognitive attitudes towards
private driving.
In summary, the (in)congruency between affective and cognitive attitudes is translated
into significantly diverse structural effects of TPB-related predictors (i.e., overall attitude,
subjective norm, passive and active PBC) on green travel intention and behavior, support-
ing the hypothesis of moderating role played by the affective-cognitive attitude congruence
in the decision-making process on green travel behavior defined by the TPB.

Practical implications

The congruence between affective and cognitive attitude is confirmed to have a mod-
erating effect on the TPB factors-green travel behavior relationships, based on which
the following implications are provided. Given the diverse features across groups, it is
strongly suggested that transport administrators dedicate the limited funds and resources
to fields and population groups where the most intensive effect would be achieved.
The experiment in Chaiken and Baldwin (1981) showed that subjects who were high
in affective-cognitive congruence were not significantly influenced by manipulation. In
this regard, for the group with positive and congruent affective-cognitive attitudes, i.e.,
possessing negative feelings and emotions towards private driving, as well as sufficient
knowledge and beliefs about the negative consequences on environment due to private
driving, it is reasonable to believe that they are the least likely to be attracted by the
anti-ecology behavioral cues. Hence, considerable efforts and resources could be saved
from these people. On the other hand, for those with strong affection for private cars and
a lack of consciousness about the negative consequences on environment due to private
driving, a great deal of effort would be consumed on their behavior initiation (Sheeran
and Abraham 2003). These implications also emphasize the importance for the develop-
ment of affective-cognitive congruence of attitudes (both negative to private cars) since
the attitude-behavior relationship is strengthened in the case ‘when the affective and
cognitive components of an attitude are mutually consistent, the attitude is in a stable
state’ (Rosenberg 1968).
A comparatively considerable effect of subjective norm on green travel intention is
obtained for the two group with incongruent attitudes. Hence, the practical behavior,
opinions and encouragement from important others on green travel are highly likely
to enhance people’s green travel intention, especially for those with shaky attitudes
towards private driving. For example, creation of pro-green-travel social environment,
as well as the role model, is expected to be effective.

13
3032 Transportation (2021) 48:3013–3035

As mentioned above, PBC is refined into active and passive components, which helps
to clarify which kind of transport management strategies would be effective towards
particular population group. The limited effect of active PBC on intention suggests that
improvement in services and infrastructures of public transport, walking and cycling
plays a minor role in promoting people’s green travel intention in the positive/incongru-
ent group. The considerable effect of passive PBC on green travel intention indicates
that regulations and restriction on cars should be implemented by administrations to
significantly push green travel intention of the negative/incongruent group. For people
firmly opposed to private driving, neither the improvement of green travel environment
nor administrative measures over private driving would effectively make them to carry
out green travel behavior. In this case, the resources and funds should not be allocated.
Though most studies have emphasized the differences between regions/cities/coun-
tries given their diverse culture, geography, economy, and etc., such that findings
obtained from one place cannot be simply applied to others. In this situation, it refers to
that empirical results and interventions utilizing the TPB are inconclusive and inconsist-
ent across regions. To provide further implications on this issue, the moderating effect
of affective-cognitive congruence in the TPB predicting green travel behavior is thus
accounted for, which has only been tested in this particular study. Empirical studies
applied on sample data collected from cities with highly-developed economy in China
find that people in Chinese cities are motivated by similar influences on their travel
behavior to people in Western cities, depending on the degree of affective-cognitive (in)
congruence of attitudes, where most of the published studies have been conducted in
the past. This is likely the flip side of the affective-cognitive congruence findings, ulti-
mately justifying the value to include its moderating role.
In spite of the value in predicting green travel, there are some limitations to be improved.
Restricted by the questionnaire, some items and factors, in particular affective and cogni-
tive attitudes, are directed towards private driving, while the TPB is utilized to explain
green travel behavior. Reasons have been given about this problem from the conceptual
point of view, and affective-cognitive attitudes is empirically designated as a moderator,
instead of directly included to explain the target behavior. Nevertheless, it is still hoped that
all items would focus on specific object or behavior to gain greater explanatory power. In
addition, the measurement of cognitive attitudes is limited, primarily towards the evalua-
tions about consequences on environment due to private driving. Though being focused,
general behavioral consequences should be also considered in further research. Last but
not the least, according to the TPB, correct measurements of PBC are beliefs about the
presence of factors that may facilitate or impede performance of the behavior. However,
only factors increasing the likelihood of green travel are considered in the PBC evaluation,
while the impeding factors are missing. Though passive and active PBC are utilized for
effective policy implications, a comprehensive measurement of PBC should be adopted.

Acknowledgments This research is supported by the National Natural Science Foundation of China under
Grant Nos. 72071121 and 71701125, the Social Science Planning Research Project of Shandong Province
under Grant No. 20DGLJO12, and the Fundamental Research Funds of Shandong University under Grant
No. 2019GN008. The authors would like to express their appreciation to anyone who has provided sugges-
tions and comments on this paper. In addition, the authors would like to thank Pr. Jing at Jiangsu University
for sharing the sample data with us.

Authors’ contribution XF Literature Search and Review, Data analysis and content planning, Manuscript
Writing and Editing.

13
Transportation (2021) 48:3013–3035 3033

Compliance with ethical standards

Conflict of interest The authors declare that they have no conflict of interest.

References
Ajzen, I.: From Intentions to Actions: A Theory of Planned Behavior. Springer, Berlin (1985)
Annual report on the environmental management of motor vehicles in China (2016). Beijing, China
Baron, R.M., Kenny, D.A.: The moderator–mediator variable distinction in social psychological research:
conceptual, strategic, and statistical considerations. J. Pers. Soc. Psychol. 51(6), 1173–1182 (1986)
Beige, S., Axhausen, K.W.: The dynamics of commuting over the life course: swiss experiences. Transp.
Res. Part A Policy Pract. 104, 179–194 (2017). https​://doi.org/10.1016/j.tra.2017.01.015
Bentler, P.M.: Multivariate analysis with latent variables: causal modeling. Annu. Rev. Psychol. 31(1), 419–
456 (1980)
Bentler, P.M.: Theory and Implementation of EQS: A Structural Equations Program. BMDP Statistical Soft-
ware Inc, Los Angeles (1988)
Bissing-Olson, M.J., Iyer, A., Fielding, K.S., Zacher, H.: Relationships between daily affect and pro-envi-
ronmental behavior at work: the moderating role of pro-environmental attitude. J. Organ. Behav. 34(2),
156–175 (2013)
Black, C., Collins, A., Snell, M.: Encouraging walking: the case of journey-to-school trips in compact urban
areas. Urban Stud. 38(7), 1121–1141 (2001). https​://doi.org/10.1080/00420​98012​4102
Bradley, M.: Workshop Synthesis: Measuring the Influence of Attitudes and Perceptions. In: Johanna,
Z., Martin, L.-G., Marcela, M., Juan, A.C. (eds.) Transport Survey Methods, pp. 643–646. Emerald
Group Publishing Limited, New York (2013). https​://doi.org/10.1108/97817​81902​882-035
Castanier, C., Deroche, T., Woodman, T.: Theory of planned behaviour and road violations: the mod-
erating influence of perceived behavioural control. Transp. Res. Part F Traffic Psychol. Behav. 18,
148–158 (2013)
Chaiken, S., Baldwin, M.W.: Affective-cognitive consistency and the effect of salient behavioral infor-
mation on the self-perception of attitudes. J. Pers. Soc. Psychol. 41(1), 1–12 (1981)
Chan, D.K.-S., Fishbein, M.: Determinants of College Women’s Intentions to Tell Their Part-
ners to Use Condoms1. J. Appl. Soc. Psychol. 23(18), 1455–1470 (1993). https​://doi.
org/10.1111/j.1559-1816.1993.tb010​43.x
Change IC (2014) Mitigation of climate change. Contribution of Working Group III to the Fifth Assess-
ment Report of the Intergovernmental Panel on Climate Change 1454
Conner, M., Povey, R., Sparks, P., James, R., Shepherd, R.: Moderating role of attitudinal ambivalence
within the theory of planned behaviour. Br. J. Soc. Psychol. 42(1), 75–94 (2003)
Conner, M., Warren, R., Close, S., Sparks, P.: Alcohol consumption and the theory of planned behav-
ior: an examination of the cognitive mediation of past behaviorid. J. Appl. Soc. Psychol. 29(8),
1676–1704 (1999)
Cooke, R., Sheeran, P.: Moderation of cognition-intention and cognition-behaviour relations: a meta-
analysis of properties of variables from the theory of planned behaviour. Br. J. Soc. Psychol. 43(2),
159–186 (2004)
Crites, S.L., Fabrigar, L.R., Petty, R.E.: Measuring the affective and cognitive properties of attitudes:
conceptual and methodological issues. Pers. Soc. Psychol. Bull. 20(6), 619–634 (1994)
Deng, X., Doll, W.J., Hendrickson, A.R., Scazzero, J.A.: A multi-group analysis of structural invariance:
an illustration using the technology acceptance model. Inf. Manag. 42(5), 745–759 (2005)
Didi media research institute.: Report on urban traffic in China in the first half year of 2016. Beijing,
China (2016)
Donald, I.J., Cooper, S.R., Conchie, S.M.: An extended theory of planned behaviour model of the psy-
chological factors affecting commuters’ transport mode use. J. Environ. Psychol. 40, 39–48 (2014)
Dudow, A.: Eco-Logistics: Incorporating business transport issues into environmental management sys-
tems. Int. Inst. Ind. Environ. Econ. 5, 12 (1998)
Fishbein, M., Ajzen, I.: Predicting and Changing Behavior: The Reasoned Action Approach. Psychology
Press, New York (2010)
Fu, X., Juan, Z.: Understanding public transit use behavior: integration of the theory of planned behavior
and the customer satisfaction theory. Transportation 44(5), 1021–1042 (2017)
Hair, J.F., Black, B., Babin, B., Anderson, R.E., Tatham, R.L.: Multivariate Data Analysis, 6th edn.
Pearson Education Inc., NJ (2006)

13
3034 Transportation (2021) 48:3013–3035

Heath, Y., Gifford, R.: Extending the theory of planned behavior: predicting the use of public transporta-
tion 1. J. Appl. Soc. Psychol. 32(10), 2154–2189 (2002)
Hoffmann, C., Abraham, C., White, M.P., Ball, S., Skippon, S.: What cognitive mechanisms predict
travel mode choice? a systematic review with meta-analysis. Transp. Rev. 37(5), 631–652 (2017)
Jeon, C.M., Amekudzi, A.A., Guensler, R.L.: Evaluating plan alternatives for transportation system sus-
tainability: atlanta metropolitan region. Int. J. Sustain. Transp. 4(4), 227–247 (2010)
Jia, N., Li, L., Ling, S., Ma, S., Yao, W.: Influence of attitudinal and low-carbon factors on behavioral
intention of commuting mode choice–A cross-city study in China. Transp. Res. Part A Policy Pract.
111, 108–118 (2018)
Khoo, H.L., Ong, G.P.: Understanding sustainable transport acceptance behavior: a case study of Klang
Valley, Malaysia. Int. J. Sustain. Transp. 9(3), 227–239 (2015)
Klöckner, C.A., Matthies, E.: How habits interfere with norm-directed behaviour: a normative decision-
making model for travel mode choice. J. Environ. Psychol. 24(3), 319–327 (2004)
Kuhnimhof, T., Chlond, B., von der Ruhren, S.: Users of transport modes and multimodal travel behav-
ior steps toward understanding travelers’ options and choices. Transp. Res. Rec. J. Transp. Res.
Board 1985(1), 40–48 (2006)
Kuiper, N.A., Rogers, T.B.: Encoding of personal information: self-other differences. J. Pers. Soc. Psy-
chol. 37(4), 499–514 (1979)
Lavine, H.G., Thomsen, C.J., Zanna, M.P., Borgida, E.: On the primacy of affect in the determination of
attitudes and behavior: the moderating role of affective-cognitive ambivalence. J. Exp. Soc. Psy-
chol. 34(4), 398–421 (1998)
Markus, H.R.: Self-schemata and processing information about the self. J. Pers. Soc. Psychol. 35(2),
63–78 (1977)
Mcmillan, T.: The relative influence of urban form on a child’s travel mode to school. Transp. Res. Part A
Policy Pract. 41(1), 69–79 (2007)
Millar, M.G., Tesser, A.: The effects of affective-cognitive consistency and thought on the attitude-behavior
relation. J. Exp. Soc. Psychol. 25(2), 189–202 (1989)
Nilsson, M., Kuller, R.: Travel behaviour and environmental concern. Transp. Res. Part D Transp. Environ.
5(3), 211–234 (2000)
Norman, R.: Affective-cognitive consistency, attitudes, conformity, and behavior. J. Pers. Soc. Psychol.
32(1), 83 (1975)
Petty, R.E., Krosnick, J.A.: Attitude strength: antecedents and consequences. Erlbaum, Hillsdale (1995)
Povey, R., Conner, M., Sparks, P., James, R., Shepherd, R.: Application of the theory of planned behaviour
to two dietary behaviours: roles of perceived control and self-efficacy. Br. J. Health. Psychol. 5(2),
121–139 (2000)
Prillwitz, J., Barr, S.: Moving towards sustainability? Mobility styles, attitudes and individual travel behav-
iour. J. Transp. Geogr. 19(6), 1590–1600 (2011). https​://doi.org/10.1016/j.jtran​geo.2011.06.011
Pronello, C., Camusso, C.: Travellers’ profiles definition using statistical multivariate analysis of attitudinal
variables. J. Transp. Geogr. 19(6), 1294–1308 (2011)
Ramezani, S., Pizzo, B., Deakin, E.: An integrated assessment of factors affecting modal choice: towards
a better understanding of the causal effects of built environment. Transportation 45(5), 1351–1387
(2018)
Rosenberg, M.J.: A structural theory of attitude dynamics. Public Opin. Q. 24(2), 319–340 (1960)
Rosenberg, M.J.: Hedonism, inauthenticity, and other goads toward expansion of a consistency theory. The-
ories Cognit. Consist. A Sourcebook 9, 73–111 (1968)
Ru, X., Wang, S., Chen, Q., Yan, S.: Exploring the interaction effects of norms and attitudes on green travel
intention: an empirical study in eastern China. J. Clean. Prod. 197, 1317–1327 (2018)
Sanchezgarcia, M., Batistafoguet, J.M.: Congruency of the cognitive and affective components of the atti-
tude as a moderator on intention of condom use predictors. Soc. Indic. Res. 87(1), 139–155 (2008)
Schleicher, D.J., Watt, J.D., Greguras, G.J.: Reexamining the job satisfaction-performance relationship: the
complexity of attitudes. J. Appl. Psychol. 89(1), 165–177 (2004)
Sheeran, P., Abraham, C.: Mediator of moderators: temporal stability of intention and the intention-behavior
relation. Pers. Soc. Psychol. Bull. 29(2), 205–215 (2003). https​://doi.org/10.1177/01461​67202​23904​6
Sigurdardottir, S.B., Kaplan, S., Møller, M., Teasdale, T.W.: Understanding adolescents’ intentions to com-
mute by car or bicycle as adults. Transp. Res. Part D Transp. Environ. 24, 1–9 (2013)
Skår, S., Sniehotta, F.F., Araújo-Soares, V., Molloy, G.J.: Prediction of behaviour versus prediction of
behaviour change: the role of motivational moderators in the theory of planned behaviour. Appl. Psy-
chol. 57(4), 609–627 (2008)
Tagler, M.J., Cozzarelli, C.: Feelings toward the poor and beliefs about the causes of poverty: the role of
affective-cognitive consistency in help-giving. J. Psychol. 147(6), 517–539 (2013)

13
Transportation (2021) 48:3013–3035 3035

Trafimow, D., Sheeran, P., Conner, M., Finlay, K.A.: Evidence that perceived behavioural control is a multi-
dimensional construct: perceived control and perceived difficulty. Br. J. Soc. Psychol. 41(1), 101–121
(2002)
Trafimow, D., Trafimow, J.H.: Predicting back pain sufferers’ intentions to exercise. J. Psychol. 132(6), 581–
592 (1998). https​://doi.org/10.1080/00223​98980​95992​91
Yang H, Yoo Y It’s all about attitude: revisiting the technology acceptance model. In: decision support sys-
tems, 2004. vol 1, pp. 19–31
Yang, R., Li, L., Zhu, J.: Impact of the consciousness factor on the green travel behavior of urban residents:
an analysis based on interaction and regulating effects in Chinese cultural context. J. Clean. Prod. 274,
122894 (2020). https​://doi.org/10.1016/j.jclep​ro.2020.12289​4
Zhang, D., Schmöcker, J.-D., Fujii, S., Yang, X.: Social norms and public transport usage: empirical study
from Shanghai. Transportation 43(5), 869–888 (2016)
Zhang, J., Yu, B., Chikaraishi, M.: Interdependences between household residential and car ownership
behavior: a life history analysis. J. Transp. Geogr. 34, 165–174 (2014)

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.

Xuemei Fu obtained her Ph.D in Management at Shanghai Jiao Tong University, China. Now she is an asso-
ciate professor at school of Management, Shandong University. Her research focuses on travel behavior and
transport demand analysis. In recent years, she has made a great amount of efforts on the integration of
latent variables into activity-travel behavior modeling framework.

13