Content uploaded by Jan H Schumann
Author content
All content in this area was uploaded by Jan H Schumann on Jan 13, 2014
Content may be subject to copyright.
Executive Summaries v
B
ecause of suggested beneficial effects of word-of-
mouth (WOM) referral, service companies have
invested large amounts of money in customer
referral programs as well as programs aimed at foster-
ing positive communication among their existing cus-
tomers. However, prior research has predominantly
focused on the effects of receiving WOM referrals in
the prepurchase phase and has neglected the effect of
WOM on existing customers. Although initial evidence
suggests that receiving WOM influences customers’
performance judgments, this effect has not been studied
among customers in existing service relationships.
Moreover, the question of cross-cultural differences in
the effectiveness of received WOM has recently gained
increased prominence. Missing completely thus far,
however, is research on the moderating effect of culture
on the link between received WOM and customer
evaluations among existing customers.
Schumann and colleagues contribute to research on this
topic by proposing a positive effect of received WOM
on service quality perceptions among existing
customers. Furthermore, they predict that the four
Hofstedian cultural dimensions—power distance,
uncertainty avoidance, individualism/collectivism, and
masculinity/femininity—moderate this effect. They sug-
gest stronger effects of received WOM on service
quality perceptions in high-power-distance, high-
uncertainty-avoidance, more collectivist, and more femi-
nine cultures. The authors test hierarchical linear mod-
els on 1910 bank customers in 11 countries. The results
show that received WOM has a strong positive effect
on customer service quality perceptions. Furthermore,
received WOM has a stronger effect on the evaluations
of customers in high-uncertainty-avoidance cultures
than in low-uncertainty-avoidance cultures. No other
cultural value is a significant moderator. Uncertainty
avoidance is expressed as the tolerance for unstruc-
tured, ambiguous, or unpredictable future events.
All other cultural values involve aspects that are
less likely to be related to the relevance of external
information sources. The results imply that received
WOM is also important to existing customers and that
managers should adjust their strategy of referral
marketing to match their target group’s uncertainty-
avoidance level.
Cross-Cultural Differences in the
Effect of Received Word-of-Mouth
Referral in Relational Service
Exchange
Jan H. Schumann, Florian v. Wangenheim, Anne Stringfellow, Zhilin Yang, Vera Blazevic,
Sandra Praxmarer, G. Shainesh, Marcin Komor, Randall M. Shannon, and
Fernando R. Jiménez
Journal of International Marketing
©2010, American Marketing Association
Vol. 18, No. 3, 2010, p. v
ISSN 1069-0031X (print) 1547-7215 (electronic)
EXECUTIVE SUMMARIES
S
ervice companies make substantial investments in
programs that foster customer referral and com-
munication among customers. These programs
are primarily directed at the acquisition of new cus-
tomers. However, service firms are also fostering com-
munication among existing customers by establishing
customer communities and customer clubs, particularly
on the Web (Srinivasan, Anderson, and Ponnavlou
2002). Prior research has predominantly focused on the
effects of receiving word-of-mouth (WOM) referrals in
the prepurchase phase (e.g., Gremler 1994; Wangen-
heim and Bayón 2007). However, there is initial evi-
dence for a positive effect of received WOM referrals on
loyalty among existing customers, showing that receipt
of WOM referrals reduces switching behavior (Money
2004; Wangenheim and Bayón 2004). Still, little is
understood how these referral sources lead to increased
customer loyalty. Although initial evidence suggests that
received WOM influences customers’ performance judg-
62
Journal of International Marketing
Journal of International Marketing
©2010, American Marketing Association
Vol. 18, No. 2, 2010, pp. 62–80
ISSN 1069-0031X (print) 1547-7215 (electronic)
Cross-Cultural Differences in the
Effect of Received Word-of-Mouth
Referral in Relational Service
Exchange
Jan H. Schumann, Florian v. Wangenheim, Anne Stringfellow, Zhilin Yang, Vera Blazevic,
Sandra Praxmarer, G. Shainesh, Marcin Komor, Randall M. Shannon, and
Fernando R. Jiménez
ABSTRACT
Because of suggested beneficial effects of word-of-mouth (WOM) referral, service companies have invested large
amounts of money in customer referral programs as well as programs aimed at fostering positive communication
among their existing customers. The question of cross-cultural differences in the effectiveness of WOM has recently
gained increased prominence. The authors contribute to research on this topic by proposing a positive effect of received
WOM on service quality perceptions among existing customers. Moreover, they predict that cultural values moderate
this effect. They test the model on 1910 bank customers in 11 countries. The results show that received WOM has a
positive effect on customer service quality perceptions. Furthermore, received WOM has a stronger effect on the evalua-
tion of customers in high-uncertainty-avoidance than in low-uncertainty-avoidance cultures. No other cultural value is
a significant moderator. The results imply that received WOM is also important to existing customers and that man-
agers should adjust their strategy of referral marketing to match their target group’s uncertainty-avoidance level.
Keywords: culture, word of mouth, service, service quality, consumer behavior
Effect of Received Word-of-Mouth Referral 63
ments (Bone 1995), this effect has not been studied
among customers in existing service relationships.
Such an effect may also differ significantly across
regions or countries. Thus, global service providers need
to consider potential differences in the effect of strength
of received WOM on customers’ service quality percep-
tions across countries. Although cross-cultural differ-
ences in the evaluation of service providers have recently
gained increased research interest (Zhang, Beatty, and
Walsh 2008), cross-cultural service research cannot suf-
ficiently answer this question. Prior research on cross-
cultural differences in the effect of WOM has addressed
only customer choice processes (e.g., Money, Gilly, and
Graham 1998), product diffusion (e.g., Dwyer, Mesak,
and Hsu 2005), and the general willingness of customers
to engage in WOM (Lam, Lee, and Mizerski 2009).
Missing completely thus far is research on the moderat-
ing effect of culture on the link between received WOM
and customer evaluations among existing customers.
Our aim is to address these issues and to contribute to
existing research in two ways. First, we examine the
effect of received WOM on customers’ service quality
perceptions. Second, we explore the moderating effects
of cultural values on the effect of received WOM on cus-
tomers’ service quality perceptions. The results of this
study should help international service providers adjust
their WOM strategy to fit their specific target groups in
different countries, thereby optimizing their allocation
of marketing resources. We conduct the analysis in the
context of professional service using survey data from
customers in 11 countries on four continents.
EFFECT OF RECEIVED WOM REFERRAL ON
SERVICE QUALITY PERCEPTIONS IN
SERVICE RELATIONSHIPS
It is widely accepted that services are more difficult to
evaluate and expose customers to greater risks than
products (Murray and Schlacter 1990; Zeithaml 1981).
This trait applies particularly to professional services,
such as medical, legal, and banking services. In profes-
sional services, customers perceive greater risk and vul-
nerability because they lack the experience and knowl-
edge to fully understand and confidently evaluate the
service results (Ostrom and Iacobucci 1995; Sharma and
Patterson 1999). To reduce the greater risk associated
with services, service customers have a decreased prefer-
ence for outright purchase and depend less on observa-
tion or trial (Murray 1991). Instead, they engage to a
larger extent in information acquisition activities when
evaluating service providers. When doing so, they prefer
personal sources, such as referrals by friends, to imper-
sonal sources, such as commercials, because they have
more confidence in personal sources and find them to be
more effective than impersonal sources, such as mass
media. Therefore, WOM, as “informal communications
directed at other consumers about the ... usage, or char-
acteristics of particular ... services and/or their sellers”
(Westbrook 1987, p. 261), is a powerful information
source in services (Zeithaml and Bitner 1996).
Prior research on social influence indicates that people
are susceptible to interpersonal influence (Asch 1951;
Sherif 1935). Using this research, Venkatesan (1966)
studies the effect of interpersonal influence in the context
Jan H. Schumann is an assistant professor (e-mail: jan.
schumann@wi.tum.de), and Florian v. Wangenheim is
Professor of Services and Technology (e-mail: florian.
wangenheim@wi.tum.de), Department of Service and Tech-
nology Marketing, Business School, Technische Universitaet
Muenchen, Germany.
Anne Stringfellow is Associate Professor of Global Marketing
and Academic Director of Executive MBA Programs,
Thunderbird School of Global Management (e-mail:
anne.stringfellow@thunderbird.edu).
Zhilin Yang is an associate professor, Department of Marketing,
City University of Hong Kong, and is the founding director of
the Masters’ of Science in Marketing Program (e-mail:
mkzyang@cityu.edu.hk).
Vera Blazevic is a visiting professor, Technology and Innovation
Management Group, RWTH Aachen University, Germany, and
is also affiliated with Maastricht University, The Netherlands
(e-mail: blazevic@tim.rwth-aachen.de).
Sandra Praxmarer is Assistant Professor of Marketing, Univer-
sity of Bamberg, Germany (e-mail: Sandra.Praxmarer@
uni-bamberg.de).
G. Shainesh is Associate Professor of Marketing, Indian Institute
of Management Bangalore (e-mail: shaineshg@iimb.ernet.in).
Marcin Komor is an assistant professor, University of Econom-
ics, Katowice, Poland (e-mail: komor@ae.katowice.pl).
Randall M. Shannon is an assistant professor, College of
Management, Mahidol University (CMMU), Bangkok, and is
the program chair of the Masters in Marketing program (e-mail:
A.Randall@gmail.com).
Fernando R. Jiménez is Assistant Professor of Marketing, Uni-
versity of Texas at El Paso (e-mail: frjimenezarevalo@utep.edu).
64 Journal of International Marketing
of a consumer decision-making situation. He reveals that
customers asked to pick the best suit from among three
identical suits made decisions in accordance with an
instructed reference group. Subsequently, marketing
research has repeatedly shown that received WOM influ-
ences customer evaluations of products (Bone 1995;
Burnkrant and Cousineau 1975; Herr, Kardes, and Kim
1991; Pincus and Waters 1977) and services (Burzynski
and Bayer 1977). Burzynski and Bayer (1977) find that
moviegoers exposed to positive WOM before watching a
movie expressed more positive evaluations of the film
than moviegoers exposed to negative WOM.
Although it could be argued that the effect of received
WOM referral becomes less important if customers
already have their own prior consumption experiences,
initial evidence exists that this is not the case. Herr,
Kardes, and Kim (1991) show in a product context that
the positive impact of received WOM on the evaluation
of computers is also valid in situations in which cus-
tomers have their own prior consumption experiences.
They explain the phenomenon with the accessibility–
diagnosticity model (Feldman and Lynch 1988; Lynch,
Marmorstein, and Weingold 1988), according to which
the impact of specific pieces of information depends
partly on their accessibility from memory. This accessi-
bility is increased by the vividness of the information,
and a particularly vivid way of receiving information is
WOM. This positive effect of received WOM should not
be restricted to the evaluation of products but also
applies to service quality perceptions of customers in
service relationships who have a history of prior con-
sumption experiences with a service. Therefore, we pro-
pose the following:
H
1
: Among customers in existing service relation-
ships, higher levels of received WOM referral
result in greater service quality perceptions.
MODERATING EFFECTS OF CULTURAL
VALUES ON THE IMPACT OF RECEIVED
WOM REFERRAL ON CUSTOMERS’ QUALITY
PERCEPTIONS IN SERVICE RELATIONSHIPS
Culture can be defined as “the collective programming
of the mind, which distinguishes the members of one
group from another” (Hofstede 2001, p. 9), and has
repeatedly been shown to influence consumer cognitions
and behavior (McCort and Malhotra 1993; Triandis
1972). In line with this thought, we propose that cul-
tural values moderate the effect of received WOM on
customers’ service quality perceptions. Figure 1 depicts
the proposed research model.
Figure 1. Research Model
Cultural Values
(Target group level)
Service Quality Perceptions
(Individual level)
Control Variables
Target-group level: Gross national income based on purchasing power parity per capita
Individual level: Age, gender, length of relationship, and fixed contact service employee
Received WOM Referrals
(Individual level)
Drivers of Brand Commitment 65
Research findings on differences in the impact of WOM
across countries indicate that cultural values moderate
the cognitive processing of received WOM and, thus,
the relevance that customers attribute to received
WOM. Money, Gilly, and Graham (1998) find that
Japanese business customers consult more referral
sources in choosing a service provider than American
business customers. In the context of electronic discus-
sion boards, Chinese participants engage to a greater
extent in information-seeking behavior than their
American counterparts (Fong and Burton 2008).
Among existing customers, received WOM has a
stronger effect on customers’ switching behavior in
Japan than in the United States (Money 2004). In the
following sections, we derive hypotheses on the
moderating effects of cultural values on the impact of
WOM on customers’ quality perceptions in service
relationships. We focus on Hofstede’s (2001) four
original cultural values—power distance, uncertainty
avoidance, individualism/collectivism, and masculinity/
femininity—which have been used in prior cross-
cultural WOM research (e.g., Dawar, Parker, and Price
1996; Lam, Lee, and Mizerski 2009).
Power Distance
Power distance reflects the way a culture handles
inequality (Hofstede 2001). As a measure of interper-
sonal power and influence, it reflects the view of the less
powerful member of a hierarchy. People in high-power-
distance cultures share norms for differential prestige,
power, and wealth, as well as the belief that talents and
capabilities are unequally distributed across society
(Hofstede 2001). This inequality may pertain to physi-
cal and mental characteristics, social status and prestige,
wealth, power, or rules. The norms for differential
prestige, power, and wealth in high-power-distance cul-
tures often are expressed by authoritarian values and
support for conformity (Hofstede 2001). Research find-
ings have repeatedly shown that people in high-power-
distance cultures engage in more information acquisi-
tion behavior than people in low-power-distance
cultures (Dawar, Parker, and Price 1996; Lam, Lee, and
Mizerski 2009). Lam, Lee, and Mizerski (2009) argue
that this behavior can be explained by the higher exter-
nal locus of control among people in high-power-
distance cultures, which makes them prone to informa-
tion search behavior in their decision making. In con-
trast, Dawar, Parker, and Price (1996) argue that people
in high-power-distance cultures have a distrust of others
because they are aware that power that rests in people is
often coercive and not legitimately based. Therefore, in
high-power-distance cultures, authorities are accepted
but not necessarily trusted. Empirical findings show that
people in high-power-distance cultures rely less on
impersonal information sources, such as salespeople,
than on personal information sources (Dawar, Parker,
and Price 1996).
In contrast, people in low-power-distance cultures
adopt a norm for more equal distribution of prestige,
power, and wealth. In their opinion, inequality in a soci-
ety should be minimized. Thus, people in low-power-
distance cultures—especially national elites—hold
relatively antiauthoritarian values. People in low-power-
distance cultures are more prone to impersonal informa-
tion acquisition (Dawar, Parker, and Price 1996).
Moreover, they should be more comfortable with mak-
ing evaluations based on their own experiences because
of their higher internal locus of control (Lam, Lee, and
Mizerski 2009). Therefore, we suggest the following:
H
2
: The effect of received WOM referral on
customer service quality perceptions is greater
for customers in high-power-distance cultures
than for customers in low-power-distance
cultures.
Uncertainty Avoidance
Uncertainty avoidance is expressed as the tolerance for
unstructured, ambiguous, or unpredictable future events
(Hofstede 2001). Hofstede (2001) states that the unpre-
dictability of the future is a given fact of human exis-
tence, of which all people are conscious. However,
people in different cultures deal with this fact in differ-
ent ways.
High-uncertainty-avoidance cultures are characterized
by a need to reduce ambiguity and risk (Kale and Barns
1992), which is manifested in a high prevalence of strict
rules and regulations. Compared with people in low-
uncertainty-avoidance cultures, members of high-
uncertainty-avoidance cultures perceive life more as a
threat and experience higher levels of anxiety. To lower
this anxiety, they should be more motivated to reduce
the perceived ambiguity and uncertainty of life (Doney,
Cannon, and Mullen 1998). A way to reduce ambiguity
and uncertainty in the context of services is to seek
advice or assurance from trusted others. Consistent with
this notion, high uncertainty avoidance is associated
with a higher level of opinion-sharing (Lam, Lee, and
Mizerski 2009; Liu, Furrer, and Sudharshan 2001), as
well as opinion-seeking (Dawar, Parker, and Price 1996;
Money, Gilly, and Graham 1998), behavior. In the con-
text of service relationships, this level would suggest
more reliance on received WOM referral from reliable
others who already have experience with or knowledge
of the service when evaluating a provider’s service
quality. People in high-uncertainty-avoidance cultures
will also seek these referrals in already existing service
relationships because they want to reassure themselves
of their opinion about the service provider.
In low-uncertainty-avoidance cultures, anxiety can be
reduced through passive relaxation; thus, people are
expected to control their emotions (Hofstede 2001).
Moreover, low-uncertainty-avoidance cultures are char-
acterized by a greater openness to change and new
ideas, making people more comfortable with ambiguity
and chaos. People in low-uncertainty-avoidance cultures
often embrace diversity and are more curious about
things that do not conform to the norm. Finally, people
in low-uncertainty-avoidance cultures have a greater
belief that they can influence their own lives and the
world in general (Hofstede 2001). Therefore, they are
less engaged in information acquisition behavior
(Dawar, Parker, and Price 1996; Money, Gilly, and Gra-
ham 1998) and should be less susceptible to external
influences on their consumer behavior and cognitions.
Therefore, we predict the following:
H
3
: The effect of received WOM referral on cus-
tomer service quality perceptions is greater for
customers in high-uncertainty-avoidance cul-
tures than for customers in low-uncertainty-
avoidance cultures.
Individualism/Collectivism
According to Hofstede’s (2001) definition, individualism/
collectivism reflects a culture’s relation to individual
goals and accomplishments. The core origins and char-
acteristics of this dimension are differences in family
units and the extent to which they influence people’s
lives and everyday behavior. Whereas in individualist
cultures, the most important distinction is between self
and others, in collectivist cultures, the self is always
defined in the context of social networks, and the
important distinction is the line between in-group and
out-group. Collectivists are characterized by a “we”
consciousness, which means that their identity is based
on the social system in which they are embedded. More-
over, Hofstede argues that collectivist cultures are char-
acterized by high-context communication (Hall 1976)
because the tightly knit social system encompasses many
rules that regulate people’s behavior.
In contrast, individualists live in a society in which
everyone is supposed to take care of him- or herself and
his or her immediate family only. These cultures are
characterized by a strong “I” consciousness and the
emotional independence of individuals from institutions
and organizations. Parsons and Shils (1951) character-
ize individualist cultures by their strong self-orientation.
Furthermore, individualist cultures are characterized by
low-context communication (Hall 1976). Without
strong group norms and regulations, the social context
in which communication takes place indicates little
information. Therefore, individualists need to communi-
cate more openly and directly about what they have to
say.
There are contrasting opinions and findings on the effect
of individualism on customer information acquisition
behavior. Dawar, Parker, and Price (1996) argue that
individualists engage in more information acquisition
behavior and explain this by the higher level of initiative
among people in individualist cultures. However, they
do not find empirical support for their claim. Other
findings on higher information acquisition behavior per-
tain to the use of external information sources (Lam,
Lee, and Mizerski 2009). Most authors argue and find
evidence for more pronounced information sharing and
acquisition behavior among people in collectivist cul-
tures (Dwyer, Mesak, and Hsu 2005; Fong and Burton
2008; Liu, Furrer, and Sudharshan 2001; Money, Gilly,
and Graham 1998). The rationale behind this is that
collectivist cultures have stronger social ties and are
more cohesive than individualist cultures. Therefore,
people in collectivist cultures are more likely to share
information and allow reference groups to influence
their decision making (Doran 2002). We argue that
received WOM referral should also influence customer
evaluations of their service provider. Therefore, we pro-
pose the following:
H
4
: The effect of received WOM referral on cus-
tomer service quality perceptions is greater for
customers in collectivist cultures than for cus-
tomers in individualist cultures.
Masculinity/Femininity
The masculinity/femininity dimension reflects how a cul-
ture defines and deals with the gender roles of men and
women (Hofstede 2001). Whereas in more masculine
66
Journal of International Marketing
Effect of Received Word-of-Mouth Referral 67
cultures, men adhere to traditionally masculine gender
roles and only women adopt the feminine roles, in more
feminine cultures, both men and women adhere to tradi-
tionally feminine gender roles. Therefore, masculinity/
femininity refers to the way “tough” values (e.g.,
assertiveness, success, competition) dominate “tender”
values (e.g., solidarity, nurturance, service) (Hofstede
2001).
Feminine cultures are characterized by a stronger rela-
tionship orientation. For them, the quality of life and
people are more important. They stress who a person is,
and they work to live rather than the other way around.
Moreover, Hofstede (2001) characterizes feminine cul-
tures as having sympathy for the weak and perceiving
small and slow things as beautiful.
Cultures high in masculinity are characterized by a
stronger ego orientation, such that people define them-
selves and their reason for being according to their work
and money or belongings. Hofstede (2001) also charac-
terizes masculine cultures as sympathetic to the strong
and perceiving big and fast things as beautiful. Because
of the materialistic and possession-oriented nature of
masculine cultures, researchers have proposed and
found evidence for higher levels of information sharing
as well as information acquisition activities (Dwyer,
Mesak, and Hsu 2005; Lam, Lee, and Mizerski 2009;
Liu, Furrer, and Sudharshan 2001). Belongings are
highly valued in masculine cultures because they reflect
success and status. People define themselves much more
through their possessions and therefore place a greater
emphasis on information associated with these posses-
sions (Dwyer, Mesak, and Hsu 2005). This should also
be valid for service relationships, which are part of the
materialistic mind-set of people in masculine cultures.
Therefore, we suggest the following:
H
5
: The effect of received WOM referral on cus-
tomer service quality perceptions is greater for
customers in more masculine cultures than for
customers in more feminine cultures.
METHOD
Study Context
We chose retail banking as our research setting for sev-
eral reasons. First, it is one of the most internationalized
service industries (Zeithaml and Bitner 1996). Second,
retail banking services are relatively comparable across
different countries (Malhotra et al. 2005), which ensures
functional equivalence (Sekaran 1983). We collected data
in 11 countries on four continents and chose countries
that vary considerably according to Hofstede’s (2001)
cultural framework and their gross national income
based on purchasing power parity per capita (GNI/PPP;
World Bank 2009). The sample consists of business stu-
dents from leading universities and business schools in
the respective countries. A sample of business students is
appropriate in this context because business students
constitute a well-defined target group that is homoge-
neous and, therefore, highly comparable across countries
(Erdem, Swait, and Valenzuela 2006). With this sample,
we also ensure subject pool equivalence (Alden,
Steenkamp, and Batra 1999), which minimizes the affect
of other potentially influential factors, such as education,
social status, family status, wealth, and age (Bearden,
Money, and Nevins 2006). Moreover, students are
appropriate because our study is based on theory with
hypotheses at a fundamental cognitive level (Bello et al.
2009). If the results support the hypotheses, it is likely
that they also generalize to other populations. We col-
lected the data between May 2006 and February 2007,
and therefore they should be unaffected by the subse-
quent major financial crisis.
Data Collection
Survey Instrument. We conducted a paper-and-pencil
survey on customers’ relationship with their primary
bank in different countries. When possible, we applied
the English version of the survey to reduce potential
translation biases. For the translated surveys, we
ensured equivalence by asking several researchers to
back-translate the scales (Brislin 1970).
The survey consisted of three parts. The first part
involved the received WOM referrals and the customers’
service quality perceptions, which consisted of items
that were either adapted from the literature (Gefen and
Straub 2004; McKnight, Choudhury, and Kacmar 2002)
or self-developed. Second, we used the CVSCALE (Don-
thu and Yoo 1998; Yoo, Donthu, and Lenartowicz
2001) to assess the cultural values of power distance,
uncertainty avoidance, individualism/collectivism, and
masculinity/femininity. We chose this scale because
recent research has pointed out the lack of reliability
and validity of the Hofstede VSM 94 (Spector, Cooper,
and Sparks 2001; Van de Vijver and Poortinga 2002).
The CVSCALE possesses a good reliability and validity
and has proved to be cross-cultural invariant (Patterson,
Cowley, and Prasongsukarn 2006; Yoo and Donthu
2002; Yoo, Donthu, and Lenartowicz 2001). We meas-
68 Journal of International Marketing
ured all items on seven-point Likert scales (1 = “strongly
disagree,” and 7 = “strongly agree”; for the scales, see
Appendix A). Third, we assessed customer characteris-
tics and demographics, including length of relationship,
gender, age, nationality, and time spent in the country.
We also controlled for the potential effect of a fixed
contact service employee (i.e., a service employee with
whom respondents had regular contact) on service
quality perceptions by means of a yes/no question.
Secondary Data. We included the GNI/PPP of all coun-
tries in our analysis, obtained from the World Bank Key
Development Data and Statistics (World Bank 2009).
We did this to control for differences in the standard of
living and level of development across countries.
RESULTS
Demographic Profile of the Sample
A total of 2284 business students from major universi-
ties in the United States, Mexico, Australia, China,
Hong Kong, Thailand, India, Germany, the Nether-
lands, Poland, and Russia participated in the study. To
exclude other major cultural influences, we analyzed
only cases that had no missing demographic informa-
tion and were identifiable citizens of the respective
countries who had lived there since birth. The final sam-
ple consisted of 1910 respondents and displayed an
equal distribution of male and female respondents (see
Appendix B). More than two-thirds of the participants
were between 20 and 25 years of age. On average, the
length of the customer relationship with the bank was
more than eight years, and 17% of the respondents had
a fixed contact service employee. There were consider-
able differences among countries. The chi-square test
shows that gender was unequally distributed between
the countries (χ
2
= 279.53, d.f. = 10, p < .001), as were
the distribution of the respondents’ ages (χ
2
= 1066.10,
d.f. = 30, p < .001) and the average length of the cus-
tomer relationship (F = 76.21, d.f. = 10, p < .001). Fur-
thermore, the countries differed in prevalence of a fixed
contact service employee among the respondents (χ
2
=
223.72, d.f. = 10, p < .001). We controlled for these dif-
ferences during the analysis.
Reliability Tests
First- and Second-Generation Reliability Tests. We
applied first- and second-generation reliability tests to
the scales. We conducted the tests first on the largest
sample (Germany) and then extended them to the other
countries. As a result of these analyses, we reduced the
power distance scales by two items. Subsequent
exploratory factor analysis confirmed the factor struc-
ture of our survey. We conducted the exploratory factor
analysis with principal component analysis and Varimax
rotation; this resulted in six factors that represented the
proposed constructs. The total variance explained is
67.49%. Cronbach’s alpha is acceptable for most scales.
Only the WOM and power distance scales are slightly
below the recommended level of .70 (Nunnally 1978) in
single countries. Following Bollen (1989), we further
built a measurement model with the factorial structure
confirmed in the exploratory factor analysis using
AMOS 17.0. The model displays a good overall fit
(χ
2
(259) = 1398.47, p < .001; χ
2
/d.f. = 5.40; goodness-
of-fit index = .94; adjusted goodness-of-fit index = .93;
comparative fit index = .95; root mean square error of
approximation = .05), and intercorrelations between the
constructs are acceptable (see Table 1). Furthermore,
because the data meet Fornell and Larcker’s (1981) cri-
terion that the average variance explained of a factor
must be greater than any squared correlation of that fac-
tor with another, we confirmed discriminant validity for
all scales. In addition, all factor reliability scores are
well above the recommended level of .60 (Bagozzi and
Yi 1988).
Common Method Variance. The cross-sectional survey
design of this study indicates the potential for common
method bias (Podsakoff, MacKenzie, and Lee 2003).
Therefore, in this research, we applied several proce-
dural remedies that have been proposed to reduce this
potential (Podsakoff, MacKenzie, and Lee 2003; Rind-
fleisch et al. 2008). We instructed the participants to
answer spontaneously and honestly and reassured them
of a confidential and anonymous treatment of the data.
By including the GNI/PPP, retrieved from a secondary
source, we further reduced potential biases.
Another characteristic that diminishes the potential
impact of common method bias is the multilevel design
applied in this study. We use a nested, hierarchical struc-
ture with the cultural values being aggregated to the
level of country groups. Therefore, the analysis is based
on two separate data sets: one with individual-level data
and one containing aggregated group-level data. Such
an aggregation can cancel out much of the random error
and sources of bias that occur at an individual level
(Glick 1985; Kark, Boas, and Chen 2003). Finally,
H
2
–H
5
are between-group hypotheses, and we have no
reason to believe that the groups differ systematically in
common method variance (Hofman, Morgeson, and
Effect of Received Word-of-Mouth Referral 69
Gerras 2003). In summary, we argue that common
method bias should not substantially affect the results of
this research.
Test for Measurement Invariance. We tested for meas-
urement invariance across cultures using Steenkamp and
Baumgartner’s (1998) procedure. First, we assessed the
configural invariance of the scales across countries.
Configural invariance can be confirmed if a model fits
the data well in all countries. Overall, the configural
invariance models possessed an excellent model fit.
Only the Tucker–Lewis index of uncertainty avoidance
is somewhat lower than the recommended level of .90
(Hu and Bentler 1999); however, this is not a significant
concern, because all the other fit indexes of this scale are
in an acceptable range. In addition, all scales must pos-
sess equal scale intervals across countries to allow for a
meaningful comparison of the covariation across cul-
tures (Steenkamp and Baumgartner 1998).
Therefore, in the next step, we constrained the factor
loadings to be equal across the country groups to con-
trol for metric invariance. Metric invariance requires
equal scale intervals across countries, which enables us
to compare difference scores in a meaningful way. Most
scales fulfill this criterion and are at least partially met-
ric invariant (see Appendix C). Only masculinity/
femininity differs significantly from the unconstrained
model. In this case, Steenkamp and Baumgartner (1998)
suggest consulting fit indexes, which are less sensitive to
sample size (Van Birgelen et al. 2002). An assessment of
the other fit indexes showed that the partial metric
invariance model of the masculinity/femininity scale had
a very good model fit and that the fit indexes only dif-
fered marginally from the configural invariance model.
Therefore, we conclude that partial invariance is also
supported for masculinity/femininity.
For the cultural values, we further tested for scalar
invariance because we expected differences in the
absolute levels of these variables. Scalar invariance also
requires that the intercepts of the items are equal, which
enables us to compare the latent means across countries.
All full and partial scalar invariance models have a sig-
nificantly lower model fit than the configural invariance
model. Yet, again, we assessed the change in the other fit
indexes and found only smaller decreases or even
increases in model fit. Therefore, we considered the
scales partially scalar invariant.
Hypothesis Testing
Analysis Procedure. Among cross-cultural researchers,
there is an extensive debate and diverse practice on how
to account for customers’ cultural values, ranging from
the use of secondary data at the country level, to pri-
mary data at the target-group level, to primary data at
the individual level (Bearden, Money, and Nevins 2006;
Steenkamp 2001). Although the other approaches have
Table 1. Intercorrelations, Squared Intercorrelations, Average Variance Explained, and Factor Reliabilities
Intercorrelations
Scale 1 2 3 4 5 6
1. WOM .22 .01 .03 .01 .00
2. Service quality perceptions .47** .01 .04 .01 .00
3. Power distance .08** –.11** .01 .00 .20
4. Uncertainty avoidance .18** .21** .08** .12 .01
5. Individualism/collectivism –.11** –.11** –.06* –.34** .04
6. Masculinity/femininity –.04 –.05 .45** .08** .19**
Average variance extracted .58 .69 .53 .55 .50 .56
Factor reliability .80 .90 .77 .86 .86 .83
*p < .05.
**p < .01.
Notes: Intercorrelations are depicted in the lower-left part; squared intercorrelations are depicted in the upper-right part.
70 Journal of International Marketing
merit in certain contexts, our aim was to investigate the
effect of shared cultural values of a specific target group
in different parts of the world on their consumer behav-
ior. Therefore, we based our operationalization of cul-
tural values on the definition of culture as a group-level
phenomenon (Hill 1997; Hofstede 2001). Following
Lenartowicz and Roth (1999), we analyzed the effect of
culture on consumer behavior by first grouping the
respondents according to the unit of analysis, which in
our case is country. As Lenartowicz and Roth suggest,
we included only respondents who possess the pertinent
nationality and have always lived in their country to
ensure that the respondents belong to the unit of analy-
sis. Furthermore, Lenartowicz and Roth suggest assess-
ing the cultural values of the groups with primary data.
Table 2 displays the group means and standard errors of
all cultural values. Finally, Lenartowicz and Roth rec-
ommend verifying the homogeneity of the groups and,
thus, the aggregated use of the cultural values.
To justify this aggregation, we calculated the intraclass
correlation coefficients, ICC(1) and ICC(2) (Bliese
2000). The ICC(1) values range from .06 to .16 and can
be considered good (see Table 2). The ICC(2) values,
which should be .60 or higher (Ostroff and Schmitt
1993), are all above .90. This shows that the group
means are highly reliable. The significant difference
between the countries confirmed our assumption of cul-
tural distance between them.
Although we acknowledge that this measure can be only
an approximation of the target groups’ actual cultural
values, we argue that it is the best possible indicator of
the shared cultural values of each group. We rule out the
possibility of applying secondary data on cultural values
because research findings show considerable differences
in cultural values within countries (Huo and Randall
1991; Koch and Koch 2007).
Multilevel Analysis. In the next step, we used three mul-
tilevel models to test the direct effects of WOM and the
moderating effect of uncertainty avoidance. We applied
multilevel analysis because of the nested data set, with
1910 customers nested in 11 countries. These two levels
of aggregation are reflected in the model that entails
country-level as well as individual-level data.
Table 2. Group Means for Cultural Values by Country: Results of an Analysis of Variance and ICC(1) and ICC(2)
Power Distance Uncertainty Avoidance Individualism/Collectivism
a
Masculinity/Femininity
b
Country M SE M SE M SE M
c
SE
Australia 2.35 .10 5.01 .09 2.99 .08 2.49 .11
China 2.62 .12 4.84 .11 2.49 .09 3.76 .12
Germany 2.37 .06 4.24 .05 2.40 .06 3.70 .07
Hong Kong 3.12 .09 4.76 .06 2.47 .06 4.13 .10
India 2.87 .11 5.16 .10 2.16 .10 3.42 .11
Mexico 2.55 .11 4.84 .09 2.18 .10 2.79 .11
Netherlands 2.37 .08 4.44 .08 2.76 .07 3.08 .10
Poland 2.54 .08 4.80 .09 2.86 .08 3.64 .10
Russia 3.73 .12 4.97 .12 2.70 .11 4.66 .13
Thailand 3.12 .08 4.89 .06 2.24 .06 3.94 .09
United States 2.19 .08 5.11 .08 2.69 .09 2.88 .10
F (d.f. 10) 22.73* 14.97* 12.25* 34.08*
ICC(1) .11 .07 .06 .16
ICC(2) .96 .93 .92 .97
*p < .01.
a
Reversed coding of the CVSCALE to display level of Individualism.
b
According to Hofstede (2001), the responses of men and women usually differ on the masculinity/femininity dimension. Because the samples from the different
countries entail significant differences in gender distribution, we controlled for these differences when calculating the country means.
c
Estimated marginal means.
Effect of Received Word-of-Mouth Referral 71
The data’s hierarchical structure requires a hierarchical
or multilevel analysis (Bryk and Raudenbush 1992;
Steenkamp, Ter Hofstede, and Wedel 1999). When
applied to hierarchically nested data, ordinary linear
models underestimate the standard error and therefore
lead to erroneous results (Bryk and Raudenbush 1992).
Multilevel modeling avoids this error by simultaneously
analyzing effects at two or more levels of aggregation.
Multilevel modeling also provides the opportunity to ana-
lyze cross-level interaction effects between individual-
and country-level variables. We conducted the analyses
with the HLM 6.0 software. However, multilevel mod-
els might be susceptible to multicollinearity effects (De
Jong and De Ruyter 2004). Perhaps because of the small
number of groups, a full model results in a solution with
a poor data fit. Therefore, we tested our hypotheses on
the moderating effects of the cultural values with sepa-
rate models. At the individual level, the models encom-
pass customers’ received WOM, as well as customer
demographics and the information on the fixed contact
service employee as control variables. At the group level,
the models comprise GNI/PPP as a control variable and
the cultural values as predictors. The dependent variable
is perceived service quality. Following Bryk and Rauden-
bush (1992), we group-centered the individual-level
variables and grand-mean-centered the group-level
variables. In addition, we specified the beta coefficients
of WOM as random. The results of the intercept-only
models (Bryk and Raudenbush 1992) indicate that the
ICC for perceived service quality is .13, which shows a
considerable amount of between-group variance.
Finally, we tested the hypotheses.
Table 3 shows the final models. We show only the test
results for H
1
(power distance) to increase readability;
the results for the other models do not differ substan-
tially. In support of H
1
, received WOM referral had a
significant effect on customers’ service quality percep-
tions (β = .329, p < .001). Furthermore, the model
showed a significant, positive effect of a fixed contact
service employee (β = .211, p < .01). At the individual
level, the model explained 17% of the variance in the
customers’ service quality perceptions.
The analysis of the group-level effects showed that H
2
,
H
4
, and H
5
are not supported. The effect of WOM on
service quality perceptions did not differ between high-
and low-power-distance cultures (β = .068, not signifi-
cant), between individualist and collectivist cultures (β =
.080, not significant), and between masculine and femi-
nine cultures (β = .008, not significant). However, we find
support for H
3
. The effect of received WOM on service
quality perceptions was signifiantly stronger in high- than
in low-uncertainty-avoidance cultures (β = .115,
p < .01). The models accounted for 9% (individualism/
collectivism), 10% (power distance), 13% (uncertainty
avoidance), and 19% (masculinity/femininity) of the
group-level variance in customers’ service quality
perceptions.
DISCUSSION
Theoretical Implications
Our study of cross-cultural differences in the effect of
received WOM makes four important contributions to
marketing theory and practice. First, it shows that
received WOM has a strong positive effect on cus-
tomers’ service quality perceptions in a professional
service setting. This finding contributes to research on
interpersonal influences on customer evaluations (Bone
1995; Burzynski and Bayer 1977).
Second, the results show that the effect of received
WOM is also valid in existing service relationships. Even
when customers have their own extensive experience,
received WOM has a significant impact on their evalua-
tions of the service provider. However, the explained
variance at the individual level is relatively low. We
expected this given respondents’ personal experience
with the service provider. Nevertheless, the model
explains 17% of the variance in the service quality per-
ceptions. This finding extends prior WOM research,
which has been primarily directed at the purchase deci-
sion process (Murray 1991) and behavioral effects of
received WOM among existing customers (Wangenheim
and Bayón 2004). Although the results of this study
emerge from the context of banking services, they should
also transfer to existing service relationships in other
contexts because the effects are on a fundamental cogni-
tive level.
Third, we show differences in the effect of received
WOM across cultures. In line with our hypotheses, these
differences can be explained by the level of uncertainty
avoidance among the given target groups. Received
WOM has a significantly stronger effect on customers’
service quality perceptions in high-uncertainty-
avoidance than in low-uncertainty-avoidance cultures.
The explained group-level variance is rather low. The
model explains only between 9% and 19% of the vari-
ance in service quality perceptions. Yet we expected this
because interaction effects usually do not increase
explained variance; rather, their focus is to help under-
72 Journal of International Marketing
stand relationships, not to predict the dependent
variable more effectively (Aiken and West 1991; Jones
and Reynolds 2006). This finding extends prior research
findings on the effect of uncertainty avoidance on infor-
mation search behavior, which show that people in high-
uncertainty-avoidance cultures consult more external
sources before choosing a product or seeking a service
(Dawar, Parker, and Price 1996; Lam, Lee, and Mizerski
2009; Money, Gilly, and Graham 1998). The current
findings show that uncertainty avoidance not only is a
powerful moderator in the prepurchase phase but also
affects the cognitions of existing consumers.
Fourth, we did not find support for the proposed mod-
erating effects of power distance, individualism/
collectivism, and masculinity/femininity. Although prior
research has repeatedly shown that these cultural values
are determinants of information acquisition behavior in
Table 3. Results of Multilevel Analyses: Moderating Effects of Cultural Values
Model 1 Model 2 Model 3 Model 4
DV: Service Quality DV: Service Quality DV: Service Quality DV: Service Quality
Perceptions Moderator: Perceptions Moderator: Perceptions Moderator: Perseptions Moderator:
Power Distance Uncertainty Avoidance Individualism/Collectivism Masculinity/Femininity
Coefficient T Coefficient T Coefficient T Coefficient T
Intercept 5.041 46.76**
Individual-Level
Antecedents
Age .020 .54
Gender .048 1.11
Length of relationship .419E+3 1.03
Fixed contact service .211 2.72*
employee
Received WOM .329 17.63**
Group-Level
Antecedents
GNI/PPP .130E+4 1.48 .015(E+3) 2.29 .120E+4 1.45 1.10E+5 1.66
Cultural value –.042 –.15 .325 .17 .129 .37 –.207 –1.20
Cross-Level
Interactions
Received WOM × .068 1.83 .115 3.40* .080 1.52 .008 .32
cultural value
Model Fit
Deviance 5513.26 5509.62 5512.25 5515.75
d.f. 4 4 4 4
Explained Variance
Individual level .17
Group level .10 .13 .09 .19
*p < .01.
**p < .001.
Notes: DV = dependent variable. N = 1910.
Effect of Received Word-of-Mouth Referral 73
the prepurchase phase (e.g., Fong and Burton 2008;
Lam, Lee, and Mizerski 2009; Money, Gilly, and Gra-
ham 1998), we are unable to support these findings in
existing service relationships. The reason for these dif-
ferences might be found in the different level of analysis
we applied. Prior research has either used correlation
analysis (Dawar, Parker, and Price 1996; Liu, Furrer,
and Sudharshan 2001) or analyzed direct effects of cul-
tural values on customers’ information exchange or
acquisition behavior (Lam, Lee, and Mizerski 2009;
Money, Gilly, and Graham 1998). These analyses are
directed at the level of behavior but do not investigate
which cultural values affect the strength of the effect of
acquired information on customer information process-
ing. The effects of cultural values on customer informa-
tion acquisition behavior and customer information
processing need to be differentiated. Customer informa-
tion processing seems to be affected only by the cultural
value that is conceptually most closely linked with infor-
mation processing. Uncertainty avoidance is expressed
as the tolerance for unstructured, ambiguous, or unpre-
dictable future events (Hofstede 2001). All other cul-
tural values involve aspects that are less likely to be
related to the relevance of external information sources.
However, further research is needed to test these
assumptions and to provide a better understanding of
cross-cultural differences in customer cognitions in serv-
ice relationships.
Managerial Implications
Customer referrals are an established tool for customer
acquisition. The findings show that referrals also have
strong effects on customer evaluations in existing service
relationships. Received WOM influences customers’
service quality perceptions and therefore is an important
tool that marketing managers can use to increase cus-
tomer retention (Money 2004; Wangenheim and Bayón
2004). This benefit should be of particular importance in
noncontractual settings, in which service firms depend
largely on relationship building for their customer
retention.
The findings also show substantial differences in the
effectiveness of received WOM across countries. Service
marketing managers should take this difference into
account when planning their marketing strategy across
different cultures in an attempt to allocate their
resources most effectively. Marketing activities should
be targeted at fostering WOM communication among
existing customers in high-uncertainty-avoidance cul-
tures, in which received WOM is especially influential.
In such cultures, received WOM also should be a par-
ticularly effective tool for customer acquisition, and
marketing managers should install appropriate reward
programs for existing customers, whose recommenda-
tions are more likely to result in the acquisition of new
customers. In contrast, in low-uncertainty-avoidance
cultures, received WOM is a less effective tool, and serv-
ice managers should instead invest in service quality and
direct communication with their customers. Programs
directed at new customer acquisition might focus on giv-
ing the potential customers the opportunity to experi-
ence the service, such as through free trials.
The results suggest that the level of uncertainty avoid-
ance is the sole cultural metric that is significantly cor-
related with the impact of received WOM on customers’
service quality perceptions. Although WOM activity
and acquisition behavior might be affected by different
cultural values, our study gives managers a simple single
metric that informs them when fostering WOM is par-
ticularly effective in increasing service quality percep-
tions among existing customers.
We obtained our findings in a cross-cultural setting, but
they should also apply to differences in the cultural val-
ues of different target groups within a single country.
Customers from different societal milieus may differ
strongly in their cultural values, so service managers
should analyze the value system of their specific target
group to determine the extent to which received WOM
may affect their service evaluation.
LIMITATIONS AND FUTURE RESEARCH
DIRECTIONS
The results of this study again highlight the importance
of considering cross-cultural differences in customer
decision making (McCort and Malhotra 1993). How-
ever, the limitations need to be mentioned, and we pres-
ent them as avenues for further research. First, we used
a cross-sectional design to analyze service relationships,
which does not allow for an investigation of the devel-
opment of customers’ evaluations of the service
provider. Furthermore, longitudinal analyses are needed
to understand the dynamics of information acquisition
and evaluation processes of service customers over time.
Second, this study does not include information about
exactly when, how, and by whom the customers
received their WOM. This also pertains to the question
of information search behavior (Moorthy, Ratchford,
and Talukdar 1997)—that is, whether the received
WOM referral was actively searched for or passively
received. Additional research should assess these pro-
cesses in more detail to help clarify the most common
and most effective methods for WOM referrals in serv-
ice relationships. Third, we assessed only positive WOM
and analyzed its beneficial effects on customers’ evalua-
tions of the service provider. Although we would argue
that negative WOM should lead to less favorable cus-
tomer evaluations of the service provider and also that
this effect should be stronger in high-uncertainty-
avoidance cultures, further research is necessary to
investigate this proposition. Fourth, our study focuses
on a target group of business students in the banking
service industry. The hypotheses pertain to a fundamen-
tal level and should generalize to other target groups
and marketing contexts, but further research is needed
to test this claim. Fifth, we study the effects of shared
cultural values on individual behavior in 11 countries,
which puts restrictions on the effects that can be tested
with multilevel models. Further research should extend
this scope, which would allow including more variables
in the models and test several moderating effects against
each other. Sixth, we focus on Hofstede’s (2001) four
original cultural dimensions, which have been the focus
of prior research on WOM. Further research should also
investigate potential effects of Confucian dynamism
(short-term versus long-term orientation), which has
been added as a fifth cultural dimension to the Hofste-
dian framework.
Research findings suggest that the willingness to engage
in positive WOM differs across cultures (Lam, Lee, and
Mizerski 2009; Liu, Furrer, and Sudharshan 2001). Fur-
ther research should attempt to analyze the various driv-
ers of referral behavior across cultures. Research should
also investigate cross-cultural differences in the willing-
ness to engage in negative WOM. In an increasingly
global service industry, these results may provide service
managers with greater knowledge that will enable them
to optimize their relationship marketing tools to appeal
to the cultural values of their respective target groups
when exporting their services as well as to counterbal-
ance unwanted effects of negative WOM.
REFERENCES
Aiken, Leona S. and Stephen G. West (1991), Multiple Regres-
sion: Testing and Interpreting Interactions. Newbury Park,
CA: Sage Publications.
Alden, Dana L., Jan-Benedict E.M. Steenkamp, and Rajeev
Batra (1999), “Brand Positioning Through Advertising in
Asia, North America, and Europe,” Journal of Marketing, 63
(January), 75–87.
Asch, Solomon E. (1951), “Effects of Group Pressure on the
Modification and Distortion of Judgements,” in Groups,
Leadership and Men, Harold Guetzkow, ed. Pittsburgh:
Carnegie Press, 177–90.
Bagozzi, Richard P. and Youjae Yi (1988), “On the Evaluation
of Structural Equation Models,” Journal of the Academy of
Marketing Science, 16 (March), 74–94.
Bearden, William O., R. Bruce Money, and Jennifer L. Nevins
(2006), “Multidimensional Versus Unidimensional Measures
in Assessing National Cultural Values,” Journal of Business
Research, 59 (February), 195–203.
Bello, Daniel, Kwok Leung, Lee Radebaugh, Rosalie L. Tung,
and Arien van Witteloostuijn (2009), “Student Samples in
International Business Research,” Journal of International
Business Studies, 40 (April), 361–64.
Bliese, Paul D. (2000), “Within-Group Agreement, Non-
Independence, and Reliability: Implications for Data Aggrega-
tion and Analysis,” in Multi-Level Theory: Research and
Methods in Organizations, Katharine J. Klein and Steve W.J.
Kozlowski, eds. San Francisco: Jossey-Bass, 349–81.
Bollen, Kenneth A. (1989), Structural Equations with Latent
Variables. New York: John Wiley & Sons.
Bone, Paula F. (1995), “Word-of-Mouth Effects on Short-Term
and Long-Term Product Judgements,” Journal of Business
Research, 32 (March), 213–23.
Brislin, Richard W. (1970), “Back-Translation for Cross-
Cultural Research,” Journal of Cross-Cultural Psychology, 1
(September), 185–216.
Bryk, Anthony S. and Stephen W. Raudenbush (1992), Hierar-
chical Linear Models: Applications and Data Analysis Meth-
ods. Newbury Park, CA: Sage Publications.
Burnkrant, Robert E. and Alain Cousineau (1975), “Informa-
tional and Normative Social Influence in Buyer Behavior,”
Journal of Consumer Research, 2 (December), 206-215.
Burzynski, Michael H. and Dewey J. Bayer (1977), “The Effect
of Positive and Negative Prior Information on Motion Picture
Appreciation,” Journal of Social Psychology, 101, 215–18.
Dawar, Niraj, Philip M. Parker, and Lydia J. Price (1996), “A
Cross-Cultural Study of Interpersonal Information
Exchange,” Journal of International Business Studies, 27
(September), 497–516.
De Jong, Ad and Ko de Ruyter (2004), “Adaptive Versus Proac-
tive Behavior in Service Recovery: The Role of Self-Managing
Teams,” Decision Sciences, 35 (Summer), 457–91.
Doney, Patricia M., Joseph P. Cannon, and Michel P. Mullen
(1998), “Understanding the Influence of National Culture on
74 Journal of International Marketing
Effect of Received Word-of-Mouth Referral 75
Cronbach’s
Factor Loadings M (SD) Alpha
WOM
.78
Friends of mine already have made good experiences with my bank. .62 4.76 (1.41)
Friends of mine have recommended my bank to me. .74 3.70 (1.90)
Friends of mine have told me positive things about my bank. .91 4.16 (1.67)
Service Quality Perceptions
.90
My bank knows how to provide excellent service. .82 4.84 (1.33)
My bank is competent and has a lot of expertise. .86 5.02 (1.29)
The quality of my bank’s services is very high. .89 4.95 (1.34)
Overall my bank is an experienced financial institute. .77 5.37 (1.33)
Power Distance
.79
People in higher positions should make most decisions
without consulting people in lower positions. excluded 3.14(1.61)
People in higher positions should not ask people in lower
positions too frequently. excluded 3.27 (1.57)
People in higher positions should avoid social interaction with people
in lower positions. .74 2.38 (1.49)
People in lower positions should not disagree with decisions
by people in higher positions. .75 2.66 (1.51)
People in lower positions should not delegate important tasks
to people in lower positions. .69 2.97 (1.52)
Uncertainty Avoidance
.86
It is important to have instructions spelled out in detail so that I always know
what I’m expected to do. .60 4.46 (1.58)
It is important to closely follow instructions and procedures. .80 4.67 (1.40)
Rules and regulations are important because they inform me
of what is expected of me. .86 4.84 (1.32)
Standardized work procedures are helpful. .68 4.84 (1.32)
Instructions for operations are important. .73 5.02 (1.25)
Individualism/Collectivism
.85
Individuals should sacrifice self-interest for the group
(either at school or the workplace). (r) .63 3.71 (1.44)
Individuals should stick with the group even through difficulties. (r) .58 2.11 (1.34)
Group welfare is more important than individual rewards. (r) .85 2.52 (1.38)
Group success is more important than individual success. (r) .83 2.46 (1.39)
Individuals should only pursue their goals after considering the welfare of the group. (r) .68 2.66 (1.40)
Group loyalty should be encouraged even if individual goals suffer. (r) .64 2.64 (1.40)
Masculinity/Femininity .83
It is more important for men to have a professional career than it is for women. .73 3.11 (1.96)
Men usually solve problems with logical analysis; women usually
solve problems with intuition. .73 3.63 (1.79)
Solving difficult problems usually requires an active, forcible approach,
which is typical of men. .87 3.32 (1.79)
There are some jobs that a man can always do better than a woman. .65 3.99 (2.05)
Notes: Items were measured on a seven-point “strongly disagree/strongly agree” scale; (r) = reversed items.
Appendix A. Scales Used in the Research
76 Journal of International Marketing
the Development of Trust,” Academy of Management
Review, 23 (July), 601–620.
Donthu, Naveen and Boonghee Yoo (1998), “Cultural Influ-
ences on Service Quality Expectations,” Journal of Service
Research, 1 (November), 178–86.
Doran, Kathleen B. (2002), “Lessons Learned in Cross-Cultural
Research of Chinese and North American Consumers,” Jour-
nal of Business Research, 55 (October), 823–29.
Dwyer, Sean, Hani Mesak, and Maxwell Hsu (2005), “An
Exploratory Examination of the Influence of National Culture
on Cross-National Product Diffusion,” Journal of Inter-
national Marketing, 13 (2), 1–27.
Erdem, Tülin, Joffre Swait, and Ana Valenzuela (2006), “Brands
as Signals: A Cross-Country Validation Study,” Journal of
Marketing, 70 (January), 34–49.
Feldman, Jack M. and John G. Lynch (1988), “Self-Generated
Validity and Other Effects of Measurement on Belief, Atti-
tude, Intention, and Behavior,” Journal of Applied Psy-
chology, 73 (August), 421–35.
Fong, John and Suzan Burton (2008), “A Cross-Cultural Com-
parison of Electronic Word-of-Mouth and Country-of-Origin
Effects,” Journal of Business Research, 61 (March), 233–42.
Fornell, Claes and David F. Larcker (1981), “Evaluating Struc-
tural Equation Models with Unobservable Variables and
Measurement Error” Journal of Marketing Research, 18
(February), 39–50.
Gefen, David and Detmar W. Straub (2004), “Consumer Trust
in B2C E-Commerce and the Importance of Social Presence:
Experiments in E-Products and E-Services,” Omega, 32
(December), 407–424.
Appendix B. Sample Size, Gender, Age, and Length of Relationship and Fixed Contact Person by Country
Age Length of
N Gender (Years) Relationship (Months)
Fixed Contact
Country Included Male Female ≤19 20–5 26–30 ≥31 M SD Person
Germany 330 214 116 13 289 28 0 138.72 86.08 139
(64.8%) (35.2%) (3.9%) (87.6%) (8.5%) (0%) (42.1%)
Mexico 155 82 73 36 116 3 0 36.41 26.31 26
(52.9%) (47.1%) (23.2%) (74.8%) (1.9%) (0%) (16.8%)
Poland 181 42 139 0 176 2 3 41.27 27.71 9
(23.2%) (76.8%) (0%) (97.2%) (1.1%) (1.7%) (5.0%)
Australia 136 48 88 43 80 6 0 111.51 71.90 9
(35.3%) (64.7%) (31.6%) (58.8%) (4.4%) (0%) (6.6%)
India 147 120 27 6 51 23 67 90.50 88.25 30
(81.6%) (18.4%) (4.1%) (34.7%) (15.6%) (45.6%) (20.4%)
Netherlands 165 75 90 20 121 7 17 178.18 101.96 16
(45.5%) (54.5%) (12.1%) (73.3%) (4.2%) (10.3%) (9.7%)
China 126 112 14 0 126 0 0 43.76 26.03 5
(88.9%) (11.1%) (0%) (100%) (0%) (0%) (4.0%)
Hong Kong 161 52 109 0 147 12 2 101.08 57.06 11
(32.3%) (67.7%) (0%) (91.3%) (7.5%) (1.2%) (6.8%)
Russia 112 46 66 43 48 14 7 36.32 29.70 9
(41.1%) (58.9%) (38.4%) (42.9%) (12.5%) (6.3%) (8.0%)
United States 163 82 81 1 147 9 6 74.00 48.90 20
(50.3%) (49.7%) (.6%) (90.2%) (5.5%) (3.7%) (12.3%)
Thailand 234 78 156 0 95 88 51 101.27 77.80 52
(33.3%) (66.7%) (0%) (40.6%) (37.6%) (21.8%) (22.2%)
Pooled Sample 1910 951 959 162 1369 192 160 97.79 80.02 326
(49.8%) (50.2%) (8.5%) (73.1%) (10.1%) (8.4%) (17.1%)
Effect of Received Word-of-Mouth Referral 77
Glick, William H. (1985), “Conceptualizing and Measuring
Organizational and Psychological Climate: Pitfalls in Multi-
level Research,” Academy of Management Review, 10 (July),
601–616.
Gremler, Dwayne D. (1994), “Word-of-Mouth About Service
Providers: An Illustration of Theory Development in Market-
ing,” in AMA Winter Educators’ Conference Proceedings,
Vol. 5, C. Whan Park and Daniel C. Smith, eds. Chicago:
American Marketing Association, 62–70.
Hall, Edward T. (1976), Beyond Culture. New York: Doubleday.
Herr, Paul M., Frank R. Kardes, and John Kim (1991), “Effects
of Word-of-Mouth and Product Attribute on Persuasion,”
Journal of Consumer Research, 17 (March), 454–62.
Hill, Charles W. (1997), International Business: Competing in
the Global Market Place. Chicago: Richard D. Irwin.
Hofman, David A., Frederick P. Morgeson, and Stephen J. Ger-
ras (2003), “Climate as a Moderator of the Relationship
Model Fit Indexes
Scale
χχ
2
d.f.
ΔΔχχ
2
ΔΔ
d.f. RMSEA TLI CFI
WOM
Configural invariance N.A. N.A. N.A. 1.00 1.00
Full metric invariance 44.92 20 44.92 20*** .03 .98 .99
Partial metric invariance 13.90 10 13.90 10 .02 .99 1.00
Service Quality Perceptions
Configural invariance 105.37 22 .05 .95 .98
Full metric invariance 186.13 52 80.76 30** .05 .96 .97
Partial metric invariance 116.10 32 10.73 10 .04 .96 .98
Power Distance
Configural invariance N.A. N.A. N.A. 1.00 1.00
Full metric invariance 35.74 20 35.74 20* .02 .98 .99
Partial metric invariance 15.60 10 15.60 10 .02 .99 1.00
Full scalar invariance 305.61 40 305.61 40** .06 .85 .81
Partial scalar invariance 139.78 20 139.78 20** .06 .86 .92
Uncertainty Avoidance
Configural invariance 396.25 55 .06 .85 .93
Full metric invariance 466.21 95 69.96 40* .05 .91 .92
Partial metric invariance 423.70 75 27.45 20 .05 .89 .93
Full scalar invariance 942.67 125 546.42 70** .06 .85 .83
Partial scalar invariance 564.77 95 168.52 40** .05 .88 .90
Individualism/Collectivism
Configural invariance 360.51 99 .04 .91 .94
Full metric invariance 419.67 149 59.16 50* .03 .94 .94
Partial metric invariance 392.77 129 32.26 30 .03 .93 .94
Full scalar invariance 828.59 179 468.08 80** .04 .87 .86
Partial scalar invariance 531.21 149 170.70 50** .04 .91 .92
Masculinity/Femininity
Configural invariance 32.63 22 .02 .99 1.00
Full metric invariance 113.16 52 80.53 30** .03 .97 .98
Partial metric invariance 51.79 32 19.16 10* .02 .98 .99
Full scalar invariance 550.79 72 518.16 50** .06 .83 .81
Partial scalar invariance 174.42 42 141.79 20** .04 .92 .95
*p < .05.
**p < .01.
***p < .001.
Notes: RMSEA = root mean square error of approximation, TLI = Tucker–Lewis index, and CFI = comparative fit index. N.A. = not applicable.
Appendix C. Test for Measurement Invariance
78 Journal of International Marketing
Between Leader-Member Exchange and Content Specific Cit-
izenship: Safety Climate as an Exemplar,” Journal of Applied
Psychology, 88 (February), 170–78.
Hofstede, Geert (2001), Culture’s Consequences: Comparing
Values, Behaviors, Institutions, and Organizations Across
Nations, 2d ed. Thousand Oaks, CA: Sage Publications.
Hu, Li-tze and Peter M. Bentler (1999), “Cutoff Criteria for Fit
Indexes in Covariance Structure Analysis,” Structural Equa-
tion Modeling, 6 (January), 1–55.
Huo, Y. Paul and Donna M. Randall (1991), “Exploring Sub-
cultural Differences in Hofstede’s Value Survey,” Asia Pacific
Journal of Management, 8 (October), 159–73.
Jones, Michael A. and Kristy E. Reynolds (2006), “The Role of
Retailer Interest on Shopping Behavior,” Journal of Retailing,
82 (2), 115–26.
Kale, Sudhir H. and John B. Barns (1992), “Understanding the
Domain of Cross-Cultural Buyer-Seller Interactions,” Journal
of International Business Studies, 23 (Winter), 101–132.
Kark, Ronit, Shamir Boas, and Gilad Chen (2003), “The Two
Faces of Transformational Leadership: Empowerment and
Dependency,” Journal of Applied Psychology, 88 (April),
246–55.
Koch, Bradley J. and Pamela T. Koch (2007), “Collectivism,
Individualism, and Outgroup Cooperation in a Segmented
China,” Asia Pacific Journal of Management, 24 (June),
207–225.
Lam, Desmond, Alvin Lee, and Richard Mizerski (2009), “The
Effects of Cultural Values in Word-of-Mouth Communica-
tion,” Journal of Intemational Marketing, 17 (3), 55–70.
Lenartowicz, Tomasz and Kendall Roth (1999), “A Framework
for Culture Assessment,” Journal of International Business
Studies, 30 (Fall), 781–98.
Liu, Ben Shaw-Ching, Olivier Furrer, and D. Sudharshan
(2001), “The Relationships Between Culture and Behavioral
Intentions Toward Services,” Journal of Service Research, 4
(November), 118–29.
Lynch, John G., Howard Marmorstein, and Michael F. Weingold
(1988), “Choices from Sets Including Remembered Brands:
Use of Recalled Attributes and Prior Overall Evaluations,”
Journal of Consumer Research, 15 (September), 169–84.
Malhotra, Naresh K., Francis M. Ulgado, James Agarwal, G.
Shainesh, and Lan Wu (2005), “Dimensions of Service
Quality in Developed and Developing Economies: Multi-
Country Cross-Cultural Comparisons,” International Mar-
keting Review, 22 (3), 256–78.
McCort, Daniel John and Naresh K. Malhotra (1993), “Culture
and Consumer Behavior: Toward an Understanding of Cross-
Cultural Consumer Behavior in International Marketing,”
Journal of International Consumer Marketing, 6 (2), 91–127.
McKnight, Harrison D., Vivek Choudhury, and Charles Kacmar
(2002), “Developing and Validating Trust Measures for E-
Commerce: An Integrative Typology,” Information Systems
Research, 13 (September), 334–59.
Money, R. Bruce (2004), “Word-of-Mouth Promotion and
Switching Behavior in Japanese and American Business-to-
Business Service Clients,” Journal of Business Research, 57
(March), 297–305.
———, Mary C. Gilly, and John L. Graham (1998), “Explo-
rations of National Culture and Word-of-Mouth Referral
Behavior in the Purchase of Industrial Services in the United
States and Japan,” Journal of Marketing, 62 (October),
76–87.
Moorthy, Sridhar, Brian T. Ratchford, and Debarata Talukdar
(1997), “Consumer Information Search Revisited: Theory and
Empirical Analysis,” Journal of Consumer Research, 23
(March), 263–77.
Murray, Keith B. (1991), “A Test of Services Marketing Theory:
Consumer Information Acqusition Activities,” Journal of
Marketing, 55 (January), 10–25.
——— and John L. Schlacter (1990), “The Impact of Services
Versus Goods on Consumers’ Assessment of Perceived Risk
and Variability,” Journal of the Academy of Marketing Sci-
ence, 18 (Winter), 51–65.
Nunnally, Jum C. (1978), Psychometric Theory. New York:
McGraw-Hill.
Ostroff, Cheri and Neal Schmitt (1993), “Configurations of
Organizational Effectiveness and Efficiency,” Academy of
Management Journal, 36 (December), 1345–61.
Ostrom, Amy L. and Dawn Iacobucci (1995), “Consumer
Trade-Offs and the Evaluation of Services,” Journal of
Marketing, 59 (January), 17–28.
Parsons, Talcott and Edward A. Shils (1951), Toward a General
Theory of Action. Cambridge, MA: Harvard University Press.
Patterson, Paul G., Elizabeth Cowley, and Kriengsin Prasong-
sukarn (2006), “Service Failure Recovery: The Moderating
Impact of Individual-Level Cultural Value Orientation on
Perceptions of Justice,” International Journal of Research in
Marketing, 23 (September), 263–77.
Pincus, Steven and L.K. Waters (1977), “Informational Social
Influence and Product Quality Judgments,” Journal of
Applied Psychology, 62 (October), 615–19.
Podsakoff, Philip M., Scott B. MacKenzie, and Jeong-Yeon Lee
(2003), “Common Method Biases in Behavioral Research: A
Critical Review of the Literature and Recommended Reme-
dies,” Journal of Applied Psychology, 88 (October), 879–903.
Rindfleisch, Aric, Alan J. Malter, Shankar Ganesan, and Chris-
tine Moorman (2008), “Cross-Sectional Versus Longitudinal
Survey Research: Concepts, Findings, and Guidelines,” Jour-
nal of Marketing Research, 45 (June), 261–79.
Effect of Received Word-of-Mouth Referral 79
Sekaran, Uma (1983), “Methodological and Theoretical Issues
and Advancements in Cross-Cultural Research,” Journal of
International Business Studies, 14 (Fall), 61–73.
Sharma, Neeru and Paul G. Patterson (1999), “The Impact of
Communication Effectiveness and Service Quality on Rela-
tionship Commitment in Consumer, Professional Services,”
Journal of Services Marketing, 13 (June), 151–70.
Sherif, Muzafer (1935), “A Study of Some Social Factors in Per-
ception,” Archives of Psychology, 27 (187), 1–60.
Spector, Paul E., Cary L. Cooper, and Kate Sparks (2001), “An
International Study of the Psychometric Properties of the Hof-
stede Values Survey Module 1994: A Comparison of Indi-
vidual and Country/Province Level Results,” Applied Psy-
chology: An International Review, 50 (April), 269–81.
Srinivasan, Srini S., Rolph Anderson, and Kishore Ponnavlou
(2002), “Customer Loyalty in E-Commerce: An Exploration
of Its Antecedents and Consequences,” Journal of Retailing,
78 (Spring), 41–50.
Steenkamp, Jan-Benedict E.M. (2001), “The Role of National
Culture in International Marketing Research,” International
Marketing Review, 18 (1), 30–44.
——— and Hans Baumgartner (1998), “Assessing Measure-
ment Invariance in Cross-National Consumer Research,”
Journal of Consumer Research, 25 (June), 78–90.
———, Frenkel ter Hofstede, and Michel Wedel (1999), “A
Cross-National Investigation and National Antecedents of
Consumer Innovativeness,” Journal of Marketing, 63 (April),
55–69.
Triandis, Harry C. (1972), The Analysis of Subjective Culture.
New York: John Wiley & Sons.
Van Birgelen, Marcel, Ko de Ruyter, Ad de Jong, and Martin
Wetzels (2002), “Customer Evaluations of After-Sales Service
Contact Modes: An Empirical Analysis of National Culture’s
Consequences,” International Journal of Research in Market-
ing, 19 (March), 43–64.
Van de Vijver, Fons J.R. and Ype H. Poortinga (2002), “Struc-
tural Equivalence in Multilevel Research,” Journal of Cross-
Cultural Psychology, 33 (March), 141–56.
Venkatesan, M. (1966), “Experimental Study of Consumer
Behavior Conformity and Independence,” Journal of Market-
ing Research, 3 (November), 384–87.
Wangenheim, Florian v. and Tomás Bayón (2004), “The Effect
of Word of Mouth on Services Switching,” European Journal
of Marketing, 38 (9–10), 1173–85.
——— and ——— (2007), “The Chain from Customer Satis-
faction via Word-of-Mouth Referrals to New Customer
Acquisition,” Journal of the Academy of Marketing Science,
35 (Summer), 233–49.
Westbrook, Robert A. (1987), “Product/Consumption-Based
Affective Responses and Postpurchase Processes,” Journal of
Marketing Research, 24 (August), 258–70.
World Bank (2009), “World Bank Key Development Data &
Statistics,” (accessed April 21, 2009), [available at
http://web.worldbank.org/WBSITE/EXTERNAL/DATASTA
TISTICS/0,,contentMDK:20535285~menuPK:1192694~page
PK:64133150~piPK:64133175~theSitePK:239419,00.html].
Yoo, Boonghee and Naveen Donthu (2002), “The Effects of
Marketing Education and Individual Cultural Values on Mar-
keting Ethics of Students,” Journal of Marketing Education,
24 (August), 92–103.
———, ———, and Tomasz Lenartowicz (2001), “Individual
Cultural Values: A Multi-Country Investigation,” working
paper, School of Business, Hofstra University.
Zeithaml, Valarie A. (1981), “How Consumer Evaluation Pro-
cesses Differ Between Goods and Services,” in Marketing of
Services, James H. Donnely and William R. George, eds.
Chicago: American Marketing Association, 186–90.
——— and Mary Jo Bitner (1996), Services Marketing. New
York: McGraw-Hill.
Zhang, Jingyun, Sharon E. Beatty, and Gianfranco Walsh
(2008), “Review and Future Directions of Cross-Cultural
Consumer Services Research,” Journal of Business Research,
61 (March), 211–24.
THE AUTHORS
Jan H. Schumann is an assistant professor at the Depart-
ment of Service and Technology Marketing at the Busi-
ness School of Technische Universitaet Muenchen, Ger-
many. His primary research interests include cross-
cultural service marketing, relationship marketing, and
technology-intensive services. His research has been
published in Journal of Consumer Behaviour and
diverse conference proceedings. He received a best con-
ference paper award at QUIS 11, and his dissertation
proposal was honored by the AMA SERVSIG and Rela-
tionship Marketing SIG, among others.
Florian v. Wangenheim is Professor of Service and Tech-
nology Marketing at the Business School of Technische
Universitaet Muenchen, Germany. His primary research
interests are in customer management, service technology
marketing, and functional interfaces in service manage-
ment. His research has been published in Journal of Mar-
keting, Journal of the Academy of Marketing Science,
Journal of Service Research, and other outlets. He has
received awards from the American Marketing Associa-
tion, the Academy of Marketing Science, and the German
Ministry for Education and Research, among others.
80 Journal of International Marketing
Anne Stringfellow is Associate Professor of Global Mar-
keting and Academic Director of Executive MBA Pro-
grams at the Thunderbird School of Global Manage-
ment. Her research focuses primarily on cross-cultural
issues in marketing and new product development and
has been published in several academic journals, includ-
ing Management Science, Journal of Product Innovation
Management, Journal of Operations Management, and
Journal of Retailing.
Zhilin Yang is an associate professor in the Department
of Marketing at City University of Hong Kong. He is the
founding director of MSc in Marketing Program. He has
published in Journal of Marketing Research, Journal of
International Business Studies, International Business
Review, and Journal of Business Research, among oth-
ers. His main research interests include governance
strategies in marketing channels, trust in business mar-
ket, and marketing in China. He served as coeditor of a
special issue for Journal of Business Ethics and a special
issue for Industrial Marketing Management.
Vera Blazevic is a visiting professor at the Technology
and Innovation Management Group at RWTH Aachen
University, Germany, and is also affiliated with Maas-
tricht University, the Netherlands. Her research interests
include marketing strategy, customer cocreation in serv-
ices and innovation, and knowledge interfaces. Her
work has been published in Journal of Service Research,
Journal of the Academy of Marketing Science, Journal
of Business Research, and International Journal of Ser-
vice Industry Management.
Sandra Praxmarer is Assistant Professor of Marketing at
the University of Bamberg, Germany. She received her
doctoral degree at the University of Augburg, taught at
the University of Wollongong, Australia, and was an
investment manager at Allianz SE, Germany. One of her
main research interests lies in the development of trust
in business relationships.
G. Shainesh is an associate professor at the Indian Insti-
tute of Management Bangalore. His research and teach-
ing interests are in customer relationship management
and services marketing. His research has appeared in
International Journal of Technology Management, Jour-
nal of Relationship Marketing, International Marketing
Review, and IIMB Management Review, among others.
His books include Customer Relationship Management:
A Strategic Perspective (Macmillan India) and Customer
Relationship Management: Emerging Concepts, Tools
and Applications (Tata McGraw-Hill).
Marcin Komor is an assistant professor at the University
of Economics in Katowice (Poland). His scientific inter-
ests include international marketing, euromarketing, and
cross-culture communication. He has stayed on scientific
assignment at the University of Cologne, University of
Dortmund, and University of Goettingen in the frame-
work of scholarships: German Academic Exchange Ser-
vice (DAAD), The Catholic Academic Exchange Service
(KAAD), and The Union of the German Academies of
Sciences and Humanities.
Randall M. Shannon is an assistant professor at the
College of Management, Mahidol University (CMMU),
Bangkok, and is the program chair of the Masters in
Marketing program. He specializes in cross-cultural
consumer behavior and retailing, including private-label
brands. He has published more than 30 articles and
papers in international journals or conferences.
Fernando R. Jiménez is Assistant Professor of Marketing
at the University of Texas at El Paso. His research
interests include customer cocreation, service encounters,
and international marketing. His work has been pub-
lished in Journal of Marketing Theory and Practice and in
conference proceedings of the American Marketing Asso-
ciation, Association for Consumer Research, and Acad-
emy of Marketing Science, among others.
ACKNOWLEDGMENTS
This work was supported by the German Federal
Ministry of Education and Research (FKZ: 01HQ0553).
A second source of funding was through the German
Academic Exchange Service. The authors thank Ruth
Bolton, Antony Peloso, and Lonnie Ostrom from the
W.P. Carey School of Business at Arizona State Univer-
sity for their support in data collection. They also thank
the three anonymous JIM reviewers for their guidance in
improving the article.