On the Cultural Guises of Cognitive Dissonance: The Case of Easterners and Westerners

Abstract

Cognitive dissonance and effects of self-affirmation on dissonance arousal were examined cross-culturally. In Studies 1 and 2, European Canadians justified their choices more when they made them for themselves, whereas Asian Canadians (Study 1) or Japanese (Study 2) justified their choices more when they made them for a friend. In Study 3, an interdependent self-affirmation reduced dissonance for Asian Canadians but not for European Canadians. In Study 4, when Asian Canadians made choices for a friend, an independent self-affirmation reduced dissonance for bicultural Asian Canadians but not for monocultural Asian Canadians. These studies demonstrate that both Easterners and Westerners can experience dissonance, but culture shapes the situations in which dissonance is aroused and reduced. Implications of these cultural differences for theories of cognitive dissonance and self-affirmation are discussed.

Full text

On the Cultural Guises of Cognitive Dissonance: The Case of Easterners
and Westerners
Etsuko Hoshino-Browne
Swarthmore College
Adam S. Zanna
McGill University
Steven J. Spencer and Mark P. Zanna
University of Waterloo
Shinobu Kitayama
University of Michigan
Sandra Lackenbauer
University of Western Ontario
Cognitive dissonance and effects of self-affirmation on dissonance arousal were examined cross-
culturally. In Studies 1 and 2, European Canadians justified their choices more when they made them for
themselves, whereas Asian Canadians (Study 1) or Japanese (Study 2) justified their choices more when
they made them for a friend. In Study 3, an interdependent self-affirmation reduced dissonance for Asian
Canadians but not for European Canadians. In Study 4, when Asian Canadians made choices for a friend,
an independent self-affirmation reduced dissonance for bicultural Asian Canadians but not for monocul-
tural Asian Canadians. These studies demonstrate that both Easterners and Westerners can experience
dissonance, but culture shapes the situations in which dissonance is aroused and reduced. Implications of
these cultural differences for theories of cognitive dissonance and self-affirmation are discussed.
Keywords: cognitive dissonance, culture, self-affirmation, self-concepts
Individuals encounter a myriad of choices every day ranging
from very simple decisions such as which cereal to eat for break-
fast to more difficult ones such as which of two job offers to
accept. Sometimes people have to make choices not only for
themselves but also for their family members or close friends. For
instance, parents often make choices for their children when they
are young (e.g., a mother might choose a pair of jeans for her son).
In some cultures, some decisions continue to be made by parents
even for adult children (e.g., a father might make a choice of
groom for his daughter). As originally proposed by Festinger
(1957) and subsequently demonstrated by many others (e.g.,
Brehm, 1956; Heine & Lehman, 1997; Steele, Spencer, & Lynch,
1993), regardless of the significance of the decisions, people faced
with equally attractive alternatives tend to experience cognitive
dissonance and justify their decisions.
Does cultural variation exist in the experience of cognitive
dissonance and the subsequent tendency to justify or rationalize
individual decision making? We argue that this tendency to justify
Etsuko Hoshino-Browne, Department of Psychology, Swarthmore Col-
lege; Adam S. Zanna, Faculty of Law, McGill University, Montreal,
Quebec, Canada; Steven J. Spencer and Mark P. Zanna, Department of
Psychology, University of Waterloo, Waterloo, Ontario, Canada; Shinobu
Kitayama, Department of Psychology, University of Michigan; Sandra
Lackenbauer, Department of Psychology, University of Western Ontario,
London, Ontario, Canada.
This article draws on a doctoral dissertation completed by Etsuko
Hoshino-Browne under the guidance of Steven J. Spencer and Mark P.
Zanna at the University of Waterloo. The research in this article was
supported by a postgraduate scholarship from the National Science and
Engineering Research Council of Canada, the University of Waterloo
Faculty of Arts Graduate Scholarship, and the collaborative research fund
from the Graduate School of Human Environmental Studies, Kyoto Uni-
versity, under the 21st-Century Center of Excellence Program, Ministry of
Education, Culture, Sports, Science, and Technology, Japan, to Etsuko
Hoshino-Browne; an Ontario Graduate Scholarship to Adam S. Zanna; and
a research grant from the Social Science and Humanities Research Council
of Canada to Steven J. Spencer and Mark P. Zanna.
Earlier versions of the research were reported at the February 2001 (San
Antonio, Texas) and the February 2002 (Savannah, Georgia) Annual
Meetings of the Society of Personality and Social Psychology, at the May
2001 New Perspectives on Dissonance and Culture Symposium in Kyoto,
Japan, at the June 2001 (Toronto, Ontario, Canada) Annual Convention of
the American Psychological Society, and at the December 2002 Interna-
tional Symposium on Socio-Cultural Foundations of Cognition, Kyoto,
Japan.
We thank Dov Cohen, Paul Davies, John Holmes, and Christie Lomore
for their helpful comments and suggestions on earlier versions of this
article. We also thank Andrea Miller and Greg Gunn at the University of
Waterloo and Kei Sokura at Kyoto University for their assistance with data
collection. We also thank people in the Kitayama Laboratory at Kyoto
University for helping in the development of the Japanese version of the
Chinese menu used in Study 2.
Correspondence concerning this article should be addressed to Et-
suko Hoshino-Browne, Department of Psychology, Swarthmore Col-
lege, 500 College Avenue, Swarthmore, PA 19081-1397. E-mail:
ehoshin1@swarthmore.edu
Journal of Personality and Social Psychology Copyright 2005 by the American Psychological Association
2005, Vol. 89, No. 3, 294 –310 0022-3514/05/$12.00 DOI: 10.1037/0022-3514.89.3.294
294

or rationalize decisions is a part of human nature. However, we
also argue that it emerges in culture-specific ways because we
believe that culture shapes how and when such rationalization
occurs. We base our argument on the conceptualization of cogni-
tive dissonance as part of a self-image maintenance process (Spen-
cer, Josephs, & Steele, 1993; Steele et al., 1993), and we partic-
ularly focus on cross-cultural variations in people’s self-concepts
(Heine, Lehman, Markus, & Kitayama, 1999; Markus &
Kitayama, 1991; Triandis, 1989, 1996).
As delineated in self-affirmation theory (Steele, 1988; see also
Spencer et al., 1993, for its extension, the self-image maintenance
process), people tend to experience threats to their self-concepts
and the concomitant arousal of cognitive dissonance to the extent
that they sense the possibility of having made a less-than-optimal
choice. However, because different cultures espouse different self-
construals or self-views as their cultural ideals (Heine et al., 1999;
Markus & Kitayama, 1991; Triandis, 1989, 1996), there should be
cross-cultural variability in the processes of cognitive dissonance
and self-image maintenance. In particular, we believe that cogni-
tive dissonance is experienced whenever people’s important self-
concepts are threatened, but dissonance reduction depends on the
particular nature of important self-concepts espoused in a given
culture. The research presented in this article demonstrates that
cross-cultural research helps the field understand relatively basic
social psychological processes better, in this case, both cognitive
dissonance and self-image maintenance processes.
Culturally Ideal Self-Concepts, Cognitive Dissonance,
and Self-Affirmation
Cross-cultural variations in the structure of the self among
individualistic Western culture and collectivistic East Asian cul-
ture have been delineated in recent decades (e.g., Heine et al.,
1999; Markus & Kitayama, 1991; Triandis, 1989, 1996). In rela-
tion to cognitive dissonance, the characteristics of ideal Western
self-concepts most relevant to decision making seem to include the
need to be rational and to make choices independently of or free
from other entities and contexts. For instance, in choosing between
business and medicine careers, decision makers’ greatest concerns
surface when they consider whether the profession is something
they wish to pursue or whether they are choosing a profession to
satisfy the desire of their parents. Optimal Western decision mak-
ing should appear to be rational, unique, and independent, express-
ing individuals’ own preferences and desires. Sensing the possi-
bility of having made an irrational choice or having made a
decision influenced by others, for instance, could induce Western-
ers to feel that their culturally ideal self-concepts are threatened.
Such threatened feelings could, in turn, lead them to sense that
their self-integrity is damaged and, consequently, to justify their
choices as a means of reducing cognitive dissonance.
In contrast, East Asians, who hold an interdependent self-view,
tend to attach greater importance to smooth and harmonious in-
terpersonal relationships with their in-group members (e.g., Heine
et al., 1999; Markus & Kitayama, 1991; Triandis, 1989, 1996). As
a result, they tend to have strong interpersonal concerns such as
appropriately fitting in with their in-groups and knowing the
preferences of their close others and correctly anticipating those
preferences in a timely manner. For instance, when getting coffee,
East Asians may be more likely to know how their close friends
take their coffee from previous observations of the friends’ behav-
ior. Optimal East Asian decision making should incorporate and
reflect the preferences and desires of close in-group members and
can be construed as a sign of good membership in that society.
Individuals espousing interdependent self-concepts as their cul-
tural ideal find making a suboptimal decision for themselves much
less threatening than making an inconsiderate decision for their
in-group members. Such an undesirable decision can have impli-
cations for maintaining harmonious relationships and may increase
interpersonal concerns. It could also become a source of cognitive
dissonance to the extent that the decision threatens the mainte-
nance of a culturally adaptive self-image as an individual sensitive
and sensible to the needs of in-group members. Thus, interdepen-
dent East Asians are expected to justify their decisions when they
fear they might have made interpersonally inconsiderate decisions,
which can have various implications for their relationships with
in-group members, and not when they have made suboptimal
decisions for themselves, which affect only the decision makers
alone.
Cross-culturally variable self-concepts also have implications
for self-image maintenance. Steele’s (1988; see also Spencer et al.,
1993) self-affirmation theory suggests that people are motivated to
maintain an image of self-integrity. When a negative event threat-
ens their beliefs that they are morally adequate and adaptive,
individuals try to restore their positive self-images by affirming
some positive, valuable aspects of their self-concepts. For exam-
ple, when people fail a driving test, they can affirm themselves by
recalling that they have very close, warm relationships with a
number of friends. When people feel threatened by a nagging
feeling that they have made a foolish decision, they can be af-
firmed by reminding themselves of some personally important
values, such as religious beliefs. Although these theories were
developed in North America, we believe that the theory and
process are cross-culturally viable.
As described above, culturally ideal self-concepts vary across
cultures. Thus, effective self-affirmation opportunities should also
be culturally configured. Specifically, the characteristics of peo-
ple’s self-concepts that are likely to be self-affirming should
depend on people’s cultural background. For example, the char-
acteristics of independent Western self-concepts that are likely to
contribute to effective self-affirmation tend to reinforce people’s
beliefs in their individual uniqueness or distinctiveness from other
people. In contrast, the self-affirming characteristics of interdepen-
dent East Asian self-concepts tend to reinforce a strong sense of
connectedness or belongingness with important in-group mem-
bers. When people’s culturally valued sense of self is threatened
and people experience cognitive dissonance, an effective self-
image maintenance method (other than justifying the decision they
have made) is to affirm the self in a culturally adaptive and
appropriate way (i.e., affirming the independent self for Western-
ers and the interdependent self for East Asians).
Past Research on Cognitive Dissonance From
Cross-Cultural Perspectives
Heine and Lehman (1997) demonstrated in cross-cultural dif-
ferences in dissonance reduction between North Americans and
East Asians. Using a conventional free-choice paradigm, under the
guise of conducting market research of music compact discs
295
CULTURAL GUISES OF COGNITIVE DISSONANCE

(CDs), these researchers examined the relation between cognitive
dissonance and self-affirmation among Canadian and Japanese
participants and found cross-cultural differences in psychological
functioning between these two groups. In particular, Canadians
showed the usual justification of their choices of CDs as a means
of reducing dissonance. In keeping with the tenets of self-
affirmation theory (Steele, 1988; Steele et al., 1993), Canadian
participants did not justify their decisions when provided with an
opportunity to affirm themselves through positive feedback on a
personality test. Their Japanese counterparts, by contrast, did not
show a tendency to justify their choices of CDs, and thus, for these
participants, the self-affirmation manipulation turned out to be
meaningless. Specifically, if Japanese participants did not experi-
ence cognitive dissonance in choosing between comparably rated
CDs, they should also not have felt any threat to their self-integrity
and, thus, had no need to restore their positive self-images by
affirming themselves.
On the basis of these findings, Heine and Lehman (1997) argued
that their Japanese participants did not rationalize their decisions
because East Asians do not experience cognitive dissonance in the
conventional free-choice paradigm. Appealing to core differences
between the North American independent self-view and the East
Asian interdependent self-view, they suggested that cognitive dis-
sonance was a culturally constructed phenomenon specific to
North American culture. In other words, making a nonoptimal
choice for the self is threatening if one has an independent, but not
an interdependent, self-view.
However, Sakai’s (1981; Sakai & Andow, 1980) earlier research
with an induced compliance paradigm demonstrated contrasting
results with East Asian samples. Sakai (1981) asked Japanese high
school students to make a counterattitudinal speech on the aboli-
tion of coeducation either publicly (i.e., their names, affiliated
classes, and grades were included in an audiotaped speech) or
anonymously. He found that those who made the speech publicly
showed significant attitude change (i.e., higher endorsement of the
abolition of coeducation) than those who made the speech anon-
ymously. Sakai attributed this attitude change in the public speech
condition to dissonance reduction. Anticipating a counterargument
from the audience, the students making a public speech became
aware of the inconsistency between their private opinion in support
of coeducation and the public speech endorsing its abolition, which
led them to experience cognitive dissonance.
Unlike Heine and Lehman’s (1997) study, which used a free-
choice paradigm, Sakai (1981) found dissonance reduction among
Japanese individuals in an induced compliance paradigm. Was the
discrepancy in the results of these two lines of research the mere
consequence of differences in the research paradigms used? We
believe that the inconsistent results are due not to the different
paradigms used, but rather to differences in the nature of the self
being threatened in those two situations. In Sakai’s study, we
believe that participants’ interpersonal concerns accounted for the
attitude change in the public speech condition. As mentioned
earlier, in collectivistic cultures emphasizing interdependence
among in-group members, people are greatly concerned with pro-
moting smooth, harmonious relationships and avoiding unneces-
sary interpersonal friction with immediate in-group members. Any
incident that draws people’s interpersonal concerns might suggest
that they are not living up to their cultural ideals and could thus
become a source of cognitive dissonance. Although the speech
itself was audiotaped in Sakai’s study, the participants in the public
speech condition had to personally identify themselves by includ-
ing, for example, their names and classes. Revealing such personal
identification raised concerns about potential disagreement, ridi-
cule, personal attack, or ostracism from the would-be audience,
which might include friends or classmates. It is not difficult,
therefore, to imagine that such interpersonal concerns led those in
the public speech condition to change their attitudes to coincide
with their counterattitudinal speeches in anticipation of having to
defend themselves against their audience.
In the four studies described below, we investigated situations in
which East Asians and North Americans experienced cognitive
dissonance, engaged in postdecisional justification, and relied on
self-affirmation opportunities to restore threatened self-integrity.
These studies advance the understanding of when these two
groups’ culturally ideal self-concepts are threatened and subject to
cognitive dissonance and self-affirmation processes.
Unconventional Free-Choice Paradigm
Although the free-choice paradigm is known to present threats
to Westerners who hold independent self-concepts, we needed to
consider a realistic situation in which East Asians, who hold
interdependent self-concepts, were likely to experience cognitive
dissonance. In the traditional free-choice paradigm, participants
evaluate a set of objects on the basis of their own preferences and
choose between two equally evaluated alternatives for themselves.
In our studies, we called this the “self” condition. To pit against
this self condition, we needed a situation in which the experimental
task creates interpersonal concerns and has implications for inter-
dependent self-concepts, such as connectedness with close others.
For East Asians, knowing the preferences and anticipating the
desires of close in-group members are important aspects of inter-
personal relationships. Failing to meet these cultural standards
might not only hurt the others’ feelings and thereby harm close
interpersonal relationships, but also make the decision maker feel
incompetent. Such a situation could create interpersonal concerns
and thereby threaten the self-integrity of and lead to dissonance
arousal for East Asians. We therefore created the “friend” condi-
tion, in which participants were asked to evaluate a set of objects
on the basis of their close friends’ preferences and choose between
two object alternatives as a gift for their close friends.
Although music records or CDs were popular objects used in the
free-choice paradigm in past research (e.g., Heine & Lehman,
1997; Steele et al., 1993) to examine the degree of rationalization
or justification of choices people made, we chose Chinese food
entre´es as our free-choice materials. We reasoned that a food
domain would provide a good context to test our hypotheses
because a free lunch coupon as a chosen product would be enticing
to university students and would make the experimental task
realistic. Because of a relatively large population of Asian Cana-
dians at the University of Waterloo and in the surrounding area,
European Canadian undergraduates were likely to be relatively
familiar with Chinese food, and both Asian Canadian and Euro-
pean Canadian students were thought to consume it relatively
frequently, owing to its variety and inexpensiveness and to the
large number of local Chinese restaurants. We also noted the
importance of knowing and anticipating the food preferences of
important in-group members when serving them, a feature of East
296
HOSHINO-BROWNE ET AL.

Asian culture described by both Markus and Kitayama (1991) and
Heine and Lehman (1997).
Thus, for the purposes of our experiments, we created a free-
choice paradigm in which participants rank and rate a list of
Chinese food entre´es based on either their own preferences (the
self condition) or their close friends’ preferences (the friend con-
dition). The participants were then asked to choose a free lunch gift
certificate either for themselves (the self condition) or for their
close friends (the friend condition).
Study 1: Postdecisional Justification Among European
Canadians and Asian Canadians
In our first study, using the free-choice paradigm described
above, we examined how culturally ideal self-concepts (i.e., inde-
pendent self-concepts for European Canadians vs. interdependent
self-concepts for Asian Canadians) interact with situations in
which both cultural groups experience dissonance arousal and
rationalize their decisions as a means of dissonance reduction.
For European Canadians in the self condition who made their
choices for themselves, we expected to replicate the results ob-
tained by other researchers (e.g., Heine & Lehman, 1997; Steele et
al., 1993). Thus, we predicted that European Canadians would
justify their decisions when they made choices for themselves
because the prospect of making a suboptimal decision for them-
selves would be threatening to the independent self. We did not
make a clear prediction for European Canadians in the friend
condition who made their choices for their close friends. On one
hand, because previous research (Nel, Helmreich, & Aronson,
1969; Norton, Monin, Cooper, & Hogg, 2003) has shown that
Westerners can experience cognitive dissonance arising from in-
terpersonal concerns, European Canadians might justify their
choices for their friends. For example, Norton et al. (2003) found
that Westerners experience dissonance vicariously, justifying the
counterattitudinal behavior of in-group members. On the other
hand, because European Canadians have a more independent than
interdependent self-concept (Markus & Kitayama, 1991), they
may not justify choices for their friends to the same extent as their
choices for themselves or as much as Asian Canadians who made
their choices for their friends. Therefore, we refrained from mak-
ing a specific prediction for this group.
We hypothesized that Asian Canadians would justify their de-
cisions more when they made choices for their close friends than
when they made choices for themselves because the prospect of
making inconsiderate decisions for important in-group members
should evoke interpersonal concerns and threaten the interdepen-
dent self. When formulating this hypothesis, we considered that
the strength of their identification with Asian culture might be a
possible moderating factor for Asian Canadians’ interdependent
behavior. Although our Asian Canadian participants were born in
Asia, the environments in which these participants have lived in
Canada could be markedly different. Specifically, whereas some
Asian Canadians have lived in an environment similar to that of
their home country and have been surrounded by people with the
same ethnic background and cultural assumptions, others have
lived in an environment in which they have been more immersed
in North American culture. For these and likely for other reasons,
the level of identification of the Asian Canadian participants with
their Asian heritage varies greatly.
Therefore, we presumed that to the extent that Asian Canadians
identified strongly with Asian culture, they were likely to hold
stronger interdependent cultural ideals and thus to show stronger
postdecisional justification when they made the choices for their
close friends. Alternatively, the more weakly Asian Canadians
identified with their Asian background, the more likely they would
be acculturated to individualistic North American culture; thus,
they would hold interdependent cultural ideals less strongly than
those with strong identification with Asian culture. Accordingly, in
our research we focused on postdecisional justification among
Asian Canadians who indicated they strongly identified with Asian
culture.
Method
Participants
We surveyed the birth countries of potential participants during a mass
testing session and recruited European Canadians who indicated Canada as
their birth country and Asian Canadians who indicated one of the East
Asian countries (e.g., Hong Kong, China, Vietnam, Taiwan) as their birth
country. A total of 126 undergraduate students (73 women and 53 men) at
the University of Waterloo participated in the study. They received either
a partial course credit or $5 for their participation. Eight people who did not
properly follow the experimenter’s instructions for experimental tasks or
did not believe the cover story were excluded from the following data
analyses.
1
Two Asian Canadians who did not indicate that they identified
with their Asian background were also excluded from the data analyses. Of
the remaining 116 participants (69 women and 47 men), 64 were European
Canadians born in Canada (37 women and 27 men) and 52 were Asian
Canadians born in East Asia (32 women and 20 men). In all four studies
reported in this article, participants’ sex had neither a significant main
effect nor an interaction with other independent variables on the dependent
variable; thus, it is not discussed further. Among the Asian Canadian
participants, the mean length of stay in Canada was 7.5 years (SD ⫽ 4.6
years).
Procedure
Strength of identification with Asian and Canadian cultures. During
the mass testing session, Asian Canadians were asked (a) with which ethnic
group they most identify and (b) how much they identify with that ethnic
group on an 11-point scale with anchors ranging from 0 (not at all)to10
(very much). The levels of the identification strength with Asian culture
among Asian Canadian participants we originally recruited (N ⫽ 93) varied
greatly (the range was 2–10, M ⫽ 8.41). Using the median point of 9.0 on
the 11-point scale, we included in the main analyses only those who
indicated that they “most” identified with Asian culture and that they
“strongly” identified with Asian culture. The identification strength with
Asian culture and the length of stay in Canada significantly correlated,
r(91) ⫽⫺.31, p ⬍ .01. Thus, the longer Asian Canadians stayed in
Canada, the less they identified with Asian culture.
In the mass testing session, we also asked potential participants to rate
the extent to which they identify with Canadian culture on an 11-point
scale with anchors ranging from 0 (not at all)to10(very much). The
identification strength with Canadian culture significantly correlated with
1
Some participants were excluded from analyses in each study because
they did not complete the Time 1 ranking and rating task as instructed, they
did not write an essay under the self-affirmation manipulation as instructed,
or they did not believe the cover story of the Chinese restaurant and gift
certificates.
297
CULTURAL GUISES OF COGNITIVE DISSONANCE

the identification strength with Asian culture, r(91) ⫽⫺.21, p ⬍ .05, and
it correlated significantly with the length of stay in Canada, r(91) ⫽ .40,
p ⬍ .001. In all studies, the experimenter was blind to participants’ strength
of identification with Asian and Canadian cultures.
Materials and free-choice paradigm. Participants reported to the lab-
oratory individually and were greeted by a female or male experimenter.
They were then randomly assigned to either the self or friend condition.
They were provided with a cover story stating that a group of researchers
were investigating the psychology of decision making in real-life situa-
tions, in collaboration with the university administrative office, and that the
researchers wanted to evaluate the popularity of a menu in a soon-to-be-
opened Chinese restaurant among University of Waterloo students. Fur-
thermore, participants were told that the university administration would
use the student feedback to determine whether the restaurant should be
included in a student prepaid card system that can be used in restaurants on
and around campus. Also, the restaurant would use this feedback to create
a special lunch menu that would be attractive to university students.
In the friend condition, participants were further told that “Past research
has shown that survey responses are more meaningful and more accurate
when respondents picture themselves making decisions for another person.
For the purpose of this survey, we would like you to picture a close friend,
someone whose food preferences you feel you know fairly well, and
respond as you make the decision for your friend.” Then, these participants
were given a minute to think about their close friends and select one. Once
they indicated that they had chosen their friends, the experimenter con-
firmed that the friend was a close friend and not merely an acquaintance
and that they knew the friend’s food preferences well. These questions
were asked to prevent participants in the friend condition from substituting
their own preferences for their friends’ preferences. Those in the self
condition, on the other hand, were not given these additional explanations
and questions. To ensure that participants remembered whose preferences
they should use, especially for those in the friend condition, each ques-
tionnaire had explicit instructions on whose preferences they should base
their ratings and decisions.
All participants were then given a list of Chinese entre´es that contained
25 dishes and asked to choose 10 items on the basis of either their own
preferences (the self condition) or their close friends’ preferences (the
friend condition). Once they had chosen the 10 most preferred entre´es, they
were asked to rank order them on the basis of either their own or their
friends’ preferences and then rate each item in terms of how much they or
their friends would like to order each entre´e on a 7-point scale with anchors
ranging from 1 (not at all)to7(very much); the Time 1 rating measure.
After filling out a demographic questionnaire, the participants were pre-
sented with two gift certificates that corresponded to their 5th and 6th
ranked entre´es. Depending on the condition they were assigned to, they
were asked to choose one for themselves or one for their friends that they
or their friends could redeem for one free lunch when the restaurant opened
the following month. On indicating their choice of a free coupon, partic-
ipants were asked whether they were certain that they or their friends
would really like the choice. Then, those in the self condition were asked
to write down their own names beside “Client’s Name” on the coupon,
whereas those in the friend condition were asked to write down their
friends’ names beside “Client’s Name” and their own names beside “Com-
pliments of.” After participants personalized their choices of coupons in
this manner, the experimenter left the laboratory, ostensibly to fetch
another participant, and left them alone for 10 min. During this period, the
chosen coupons were left on the desk in front of the participants. When the
experimenter returned to the lab, participants were asked to look at a more
detailed version of the same menu of 25 entre´es, which contained an
elaborated description and the price of each entre´e. The 10 items initially
chosen by each participant were highlighted. Participants were asked to
transcribe these items in the order they appeared in the menu and then to
rerate each of the 10 items in terms of how much they or their friends
would like it on a 9-point scale with anchors ranging from 1 (not at all)to
9(extremely); the Time 2 rating measure. We changed the rating scales
slightly between Time 1 and Time 2 to prevent our participants from
simply remembering their ratings at Time 1 and using them at Time 2 out
of their motivation to be consistent between the two measures.
Because our experimental materials involved food preferences, we
avoided scheduling experimental sessions at mealtime. Still, we included a
question probing how hungry participants were on a 4-point scale with
anchors ranging from 1 (not at all hungry)to4(very hungry).
2
We also
included three questions in the friend condition. One question probed what
type of relationship participants had with their friends (e.g., romantic
partner, best friend, relative). The second question probed how close
participants felt toward their friends for whom they chose the gift certifi-
cates on a 5-point scale with anchors ranging from 1 (not at all close)to5
(very close).
3
The third question probed how participants thought their
chosen gift certificates would be delivered to their friends. Participants
normally assumed that they themselves would give the gift certificates to
their friends in person. After a few questions assessing whether participants
were suspicious about the deception, participants were thoroughly de-
briefed and received an unexpected $5 in addition to their participation
credit or remuneration. We made this additional payment to compensate for
the invalid gift certificates.
In summary, the study had a 2 (cultural group: European Canadian vs.
Asian Canadian) ⫻ 2 (target of coupon choice: self vs. friend) between-
subjects factorial design.
Results
The dependent variable in Study 1 was the postdecisional jus-
tification expressed by the spread of alternatives. The spread of
alternatives was calculated from the rating measures of the 5th and
6th ranked entre´es at Time 1 and Time 2. First, the 9-point scale of
the Time 2 rating was converted to a 7-point scale to match with
the Time 1 rating. Then, the sum of an increase in the rating of the
2
The grand mean across the four studies of how hungry participants felt
during the experimental session was M ⫽ 1.70 (SD ⫽ 0.85). The mean
difference among the four studies was not significant, F(3, 439) ⫽ 1.87,
p ⬍ .14. A post hoc analysis indicated that the only significant mean
difference was between Study 2 (M ⫽ 1.81, SD ⫽ 0.68) and Study 4 (M ⫽
1.59, SD ⫽ 0.94). In Study 2, there was a significant main effect for
cultural group on the felt hunger, F(1, 177) ⫽ 4.23, p ⬍ .05. European
Canadians felt hungrier than Japanese. There were no other significant
effects in Study 2. In Study 4, there was a significant main effect for the
self-affirmation condition on felt hunger, t(99) ⫽ 2.05, p ⬍ .05. Partici-
pants in the no self-affirmation condition felt hungrier than those in the
independent self-affirmation condition. There were no other significant
effects in Study 4. In all four studies, the hunger variable did not have any
significant main effects or interactions with other independent variables on
the main dependent variable. Thus, it is not discussed further.
3
The grand mean across the four studies of how close participants in the
friend condition felt toward their friends was M ⫽ 4.04 (SD ⫽ 0.94). There
was a significant mean difference among four studies, F(3, 293) ⫽ 12.43,
p ⬍ .001. A post hoc analysis indicated that Study 1 (M ⫽ 3.47, SD ⫽
1.38) was significantly different from Studies 2, 3, and 4 (M ⫽ 4.33, SD ⫽
0.80, M ⫽ 3.95, SD ⫽ 0.58, and M ⫽ 4.20, SD ⫽ 0.71, respectively), and
Study 2 was significantly different from Study 3. Within each study, the
means of the felt closeness variable did not differ significantly among
different cultural groups or conditions. In Studies 1, 3, and 4, this variable
did not have any significant main effects or interactions with other inde-
pendent variables on the dependent variable. However, there was a signif-
icant interaction between the felt closeness variable and cultural groups on
the spread of alternatives in Study 2; thus, it is discussed in the Study 2
discussion section in relation to in-group identification.
298
HOSHINO-BROWNE ET AL.

entre´e of the chosen coupon and a decrease in the rating of the
entre´e of the nonchosen coupon between Time 1 and Time 2
measures was calculated. It did not matter whether we analyzed the
data by adjusting the Time 2 scale to the Time 1 scale or vice versa
or by standardizing both scales. Thus, we used the simplest method
in this and subsequent studies.
We conducted a 2 (cultural group: European Canadian vs. Asian
Canadian ) ⫻ 2 (target of choice: self vs. friend) analysis of
variance (ANOVA) to test our prediction for an interaction be-
tween the two cultural groups and the target person (i.e., self or
friend) for whom participants chose gift certificates. The results
indicated that there were no significant main effects for cultural
group or for target of coupon choice, F(1, 112) ⫽ 0.25, ns, and
F(1, 112) ⫽ 0.38, ns, respectively. However, as predicted, the
two-way interaction was significant, F(1, 112) ⫽ 4.38, p ⬍ .05
(see Figure 1). European Canadians in the self condition tended to
show a greater spread of alternatives than Asian Canadians in the
self condition, F(1, 112) ⫽ 3.22, p ⫽ .07. Among Asian Canadi-
ans, those in the friend condition tended to show a greater spread
of alternatives than those in the self condition, F(1, 112) ⫽ 3.32,
p ⬍ .07. European Canadians in the friend condition showed less
spread of alternatives compared with European Canadians in the
self condition and Asian Canadians in the friend condition, al-
though not significantly so. Among European Canadians, the dif-
ference between the self and friend conditions was F(1, 112) ⫽
1.22, ns. In the friend condition, the difference between European
Canadians and Asian Canadians was F(1, 112) ⫽ 1.33, ns.In
addition, we conducted one-sample t tests to examine whether the
spread of alternatives was significantly different from zero in each
group. In all studies, we used two-tailed tests. As predicted, the
means of the spread of alternatives of European Canadians in the
self condition (M ⫽ 0.74, SD ⫽ 1.62), t(31) ⫽ 2.57, and of
strongly identified Asian Canadians in the friend condition (M ⫽
0.77, SD ⫽ 1.53), t(27) ⫽ 2.68, were significantly different from
zero ( ps ⬍ .05). Also, as expected, the means of the spread of
alternatives of European Canadians in the friend condition (M ⫽
0.27, SD ⫽ 2.13), t(31) ⫽ 0.73, and of strongly identified Asian
Canadians in the self condition (M ⫽⫺0.07, SD ⫽ 1.12), t(23) ⫽
0.32, were not significantly different from zero (see Figure 1).
Although Asian Canadians who weakly identified with Asian
culture (i.e., those who indicated less than 9 on the 11-point scale
of Asian cultural identification strength) were not the main focus
of the first study, the inclusion of this group was important, as they
provided a nice comparison with the strongly identified Asian
Canadian group because many variables that might be relevant to
the Asian Canadian participants were equivalent in these two
groups.
4
Thus, we conducted a regression analysis, including 41
weakly identified Asian Canadians and using the strength of iden-
tification with Asian culture as a continuous variable. In particular,
we tested the spread of alternatives for an interaction between the
target person for whom Asian Canadian participants chose gift
coupons (i.e., the self or friend condition) and their identification
strength with Asian culture. Reflecting our speculation that the
stronger the Asian cultural identification the more interdependent
tendency there would be in postdecisional justification, we ob-
tained a significant crossover interaction between the target con-
dition and the identification strength with Asian culture, t(89) ⫽
2.74, p ⬍ .01 (
␤
⫽ .35). In the friend condition, the more strongly
Asian Canadians identified with Asian culture, the more they
justified their choices for their friends; the simple slope was
significant,
␤
⫽ .25, t(89) ⫽ 2.41, p ⬍ .05. In contrast, in the self
condition, the more strongly Asian Canadians identified with
Asian culture, the less they showed postdecisional justification for
the choices made for themselves; the simple slope was also sig-
nificant,
␤
⫽⫺.33, t(89) ⫽ 2.09, p ⬍ .05.
Discussion
The results of the first study mainly supported our predictions.
Replicating previous findings, we found that European Canadians
showed significant postdecisional justification when they made
4
Although the mean number of years residing in Canada was signifi
-
cantly different between weakly identified Asian Canadians and strongly
identified Asian Canadians (M ⫽ 10.9 years vs. 7.5 years), t(91) ⫽ 3.39,
p ⫽ .001, the mean strength of identification with Canadian culture was not
significantly different between the two groups (M ⫽ 6.32 vs. 5.62 on an
11-point scale), t(91) ⫽ 1.45, ns.
Figure 1. Study 1 postdecisional justification in 2 Cultural Groups ⫻ 2 Choice Targets (dependent variable ⫽
mean spread of alternatives). An asterisk indicates that the mean spread of alternatives of the condition is
significantly different from zero.
299
CULTURAL GUISES OF COGNITIVE DISSONANCE

choices for themselves. We also replicated past findings and found
that Asian Canadians did not engage in postdecisional justification
when they made choices for themselves. As hypothesized, Asian
Canadians tended to engage in postdecisional justification when
they made choices for their close friends. On the other hand,
European Canadians’ postdecisional justification of their choices
made for their friends was less than that of choices made for
themselves and less than Asian Canadians’ justification of the
choices made for their friends.
As mentioned earlier, the weakly identified Asian Canadian
group was a nice comparison group for the strongly identified
Asian Canadian group, and the inclusion of this group was partic-
ularly important in terms of the current free-choice paradigm that
used Chinese food. One might argue that the use of a Chinese
menu could prime Asian Canadians with Asian-ness or Asian
cultural norms and could encourage them to think that they should
be expert in Chinese food. Such a priming effect or encouragement
could in turn motivate them to make the right decisions and
consequently justify their decisions, particularly in the friend con-
dition. If so, then, all Asian Canadian participants, many of whom
were Chinese, should have justified their choices for their friends
to demonstrate their expertise in Chinese food, regardless of their
identification level with Asian culture. The fact that weakly iden-
tified Asian Canadians did not justify their choices for their friends
counteracts such concerns about the use of Chinese food in this
research.
The variables that might have influenced the results in the friend
condition were the types of relationships participants had with
their friends and the degree of closeness they felt toward them.
Participants categorized their friends in multiple ways, such as a
best friend and romantic partner, and there were no striking dif-
ferences in the friend categorization. Between the two cultural
groups, 70% to 78% of participants classified their friends as
“friend” or “best friend” and about 28% indicated a “romantic
partner.” Participants also rated how close they felt toward their
friends on a 5-point scale. A t test indicated that there was no
significant difference between European Canadians (M ⫽ 4.28,
SD ⫽ 0.58) and Asian Canadians (M ⫽ 4.21, SD ⫽ 0.96), t(58) ⫽
0.33, ns. Thus, possible differences in types of friends or felt
closeness toward friends can be safely eliminated as alternative
explanations for the mean differences of the spread of alternatives
in the friend condition between the two cultural groups.
Although the pattern of the means among the four conditions
was as predicted and the interaction was significant, the contrasts
of the means were not as statistically significant as we expected.
Thus, we conducted a second study in Canada and in Japan in an
effort to replicate the pattern of postdecisional justification for the
choice made for the self and a friend.
Study 2: Replication of Study 1 Among European
Canadians and Japanese
Study 2 was meant to be a conceptual replication of Study 1,
with Japanese participants replacing Asian Canadians. We main-
tained the same hypotheses outlined in Study 1. It was also
important to see whether we obtained results consistent with Study
1 among European Canadians in the friend condition.
Thus, our specific hypotheses were that European Canadians
would engage in postdecisional justification when they made
choices for themselves but that they might or might not do so when
they chose the gift certificates for their friends. We expected that
Japanese would justify their choices when they made the choices
for their friends but that they would not when they chose the gift
certificates for themselves.
Method
Participants
A total of 197 students participated in this cross-national study. In
Canada, 104 European Canadian students (61 women and 43 men) from the
University of Waterloo participated. They received either partial course
credit or $5 for their participation. As in Study 1, we measured partici-
pants’ strength of identification with Canadian culture on an 11-point scale
during the mass testing session. The East Asian group was composed of 93
Japanese students (55 women and 38 men) from Kyoto University, Kyoto,
Japan. They were compensated for their participation with a bookstore gift
coupon that was worth 500 Japanese yen, which was comparable to the $5
at the University of Waterloo. Fifteen people who did not properly follow
the instructions given by the experimenters were excluded from the data
analyses. Also, 1 participant was excluded because his response on the
main dependent variable was over 3.5 standard deviations from the mean,
and thus regarded as an outlier. Thus, 181 participants (99 European
Canadians and 82 Japanese; 109 women and 72 men) were included in the
analyses.
Procedure
Participants were randomly assigned to either the self or friend condi-
tion. Two research assistants, one female one male, ran the study at the
University of Waterloo, which was conducted in English; Etsuko Hoshino-
Browne and a female research assistant ran the study at Kyoto University,
which was conducted in Japanese.
The free-choice materials and procedures used were basically the same
as those used in Study 1, with a few modifications. The materials used in
Japan were translated into Japanese by Etsuko Hoshino-Browne and pre-
tested among the members of the research laboratory at Kyoto University.
To eliminate any potential social aspects that go along with dining out, the
self condition in Study 2 was slightly modified, so that participants chose
a coupon for a take-out rather than for eating in a restaurant. This was
especially a concern for Japanese participants because it is common for
people in Japan to go to a Chinese restaurant with a group of friends or
family members and share various dishes. Consequently, it was possible
that people might take the other diners’ preferences into account in choos-
ing a coupon even when they chose the coupon for themselves. By making
the menu and the coupons in the self condition for take-out, this possibility
was reduced. The dishes listed on the Japanese version of the menu were
also modified according to popular Chinese dishes in Japan.
At the end of the experiment, both Japanese and European Canadian
participants were thoroughly debriefed. Only European Canadian partici-
pants received an unexpected $5 in addition to their participation credit or
remuneration to replace the invalid gift certificates, because the additional
payment was unconventional at Kyoto University.
In summary, the study had a 2 (cultural group: European Canadians vs.
Japanese) ⫻ 2 (target of coupon choice: self vs. friend) between-subjects
factorial design.
Results
The dependent variable was the postdecisional justification ex-
pressed by the spread of alternatives, and we calculated it in the
same manner as in Study 1.
300
HOSHINO-BROWNE ET AL.

The purpose of Study 2 was to replicate the pattern of the means
of postdecisional justification obtained in Study 1; thus, we pre-
dicted a significant interaction between the two cultural groups and
the target of coupon choice.
We conducted a 2 (cultural group: European Canadians vs.
Japanese) ⫻ 2 (target of coupon choice: self vs. friend) ANOVA,
which indicated that the predicted interaction was significant, F(1,
177) ⫽ 4.65, p ⫽ .03 (see Figure 2). As expected, European
Canadians in the self condition tended to show a greater spread of
alternatives than Japanese in the self condition, F(1, 177) ⫽ 3.60,
p ⬍ .06. Also as predicted, Japanese in the friend condition tended
to show a greater spread of alternatives than Japanese in the self
condition, F(1, 177) ⫽ 2.57, p ⬍ .11.
5
Thus, the pattern of the
means among European Canadians and Japanese was similar to
that of Study 1 among European Canadians and Asian Canadians,
although the simple contrasts comparing the means according to
the prediction were not statistically significant. We also conducted
one-sample t tests to examine whether the spread of alternatives
was significantly different from zero in each group. As predicted,
the means of the spread of alternatives of European Canadians in
the self condition (M ⫽ 0.63, SD ⫽ 1.40), t(60) ⫽ 3.51, and of
Japanese in the friend condition (M ⫽ 0.58, SD ⫽ 1.09), t(43) ⫽
3.50, were significantly different from zero, ps ⫽ .001. Also, as
expected, the means of the spread of alternatives of European
Canadians in the friend condition (M ⫽ 0.24, SD ⫽ 1.43), t(37) ⫽
1.03, and of Japanese in the self condition (M ⫽ 0.12, SD ⫽ 1.21),
t(37) ⫽ 0.60, were not significantly different from zero (see Figure
2).
Meta-Analyses Across Studies 1 and 2
The pattern of the means of postdecisional justification obtained
in Study 2 replicated the results in Study 1, and the interaction
between the two cultural groups and the target of decision was
significant. However, the simple contrasts of the means were not
statistically significant. Therefore, we conducted a series of meta-
analyses (see Rosenthal, 1991) across these two studies to test the
interaction, simple contrasts of the means, and one-sample t tests
on the spread of alternatives to ascertain the extent European
Canadians and East Asians engaged in postdecisional justification.
The meta-analysis for the interaction between the two cultural
groups and the target of coupon choice was significant, z ⫽ 2.98,
p ⬍ .005. Both European Canadians and East Asians showed the
most justification of their choices when their culturally important
self-concepts were threatened by making choices. That is, Euro-
pean Canadians justified their choices made for themselves,
whereas East Asians justified their choices made for their friends.
A series of meta-analyses for simple contrasts indicated that the
difference between European Canadians and East Asians in the
self condition was significant, z ⫽ 2.62, p ⬍ .01. Among East
Asians, the difference between the self and friend condition was
also significant, z ⫽ 2.43, p ⬍ .02. Among European Canadians,
however, the contrast between the self and friend condition was
only marginally significant, z ⫽ 1.80, p ⫽ .07. In the friend
condition, the contrast between the two cultural groups was also
only marginally significant, z ⫽ 1.64, p ⫽ .10.
Meta-analyses for the one-sample t tests indicated that European
Canadians’ spread of alternatives in the self condition was signif-
icantly different from zero, z ⫽ 4.02, p ⬍ .001, whereas their
spread of alternatives in the friend condition was not, z ⫽ 1.22, ns.
East Asians’ spread of alternatives in the self-condition was not
different from zero, z ⫽ 0.64, ns, whereas their spread of alterna-
tives in the friend condition was significantly different from zero,
z ⫽ 4.07, p ⬍ .001.
Discussion
The results of the meta-analyses of first two studies present a
clearer picture of the nature of the interaction between the two
5
Among European Canadians, the difference between the self and friend
conditions (M ⫽ 0.63 vs. M ⫽ 0.24, respectively) was F(1, 177) ⫽ 2.08,
p ⬍ .15. In the friend condition, the difference between European Cana-
dians and Japanese (M ⫽ 0.24 vs. M ⫽ 0.58, respectively) was F(1, 177) ⫽
1.38, ns.
Figure 2. Study 2 postdecisional justification in 2 Cultural Groups ⫻ 2 Choice Targets (dependent variable ⫽
mean spread of alternatives). An asterisk indicates that the mean spread of alternatives of the condition is
significantly different from zero.
301
CULTURAL GUISES OF COGNITIVE DISSONANCE

cultural groups and the target of decision. The meta-analyses also
indicated that the extent to which European Canadians engaged in
justification for the choices they made for their friends was not
different from zero. The meta-analyses also indicated that their
justification in the friend condition was less than their justification
in the self condition or Asians’ justification in the friend condition,
but neither contrast was statistically different. Thus, it is difficult
to claim that European Canadians justified the choices they made
for their friends or that they did not justify the choices they made
for their friends.
As mentioned earlier, there is evidence that North Americans do
experience dissonance interpersonally or vicariously when they
observe their in-group members engage in attitude-inconsistent
behavior (Norton et al., 2003). On the basis of these findings, we
speculated earlier the possibilities of North Americans engaging in
postdecisional justification of their choices for their friends. On
one hand, if European Canadians felt personal responsibility for
the choices they made for their friends or if they held strong
in-group identification with their friends (i.e., in-group identifica-
tion at a personal level), they would be likely to justify their
choices for their friends. On the other hand, if they held strong
in-group identification with their individualistic Canadian culture
(i.e., in-group identification at a cultural level), they would es-
pouse more individualistic orientations and thus would be unlikely
to justify their choices for their friends.
We examined two proxy variables of European Canadian in-
group identification, that is, the strength of felt closeness as a
personal level identification and the strength of identification with
Canadian culture as a cultural level identification. However, the
additional analyses did not shed light on the nonsignificant differ-
ence between the self and friend conditions among European
Canadians, and we did not find clear, consistent relations between
either type of identification and postdecisional justification.
6
The
most plausible interpretation at this point seems to be that although
they care about their choices made for their close friends to a
certain extent, they are more concerned about their personal
choices than their interpersonal choices.
Study 3: Postdecisional Justification and Self-Affirmation
Among Asian Canadians
Past cognitive dissonance research has demonstrated that self-
affirmation buffers threats and reduces defensiveness in the expe-
rience of cognitive dissonance and that it can be used as an
alternative means to postdecisional justification to alleviate disso-
nance arousal and maintain one’s self-image (Heine & Lehman,
1997; Steele & Liu, 1983; Steele et al., 1993). In other words,
postdecisional justification is not the only route to restoring self-
integrity. Thus, in a third study, we investigated the effect of
self-affirmation as a means of countering cognitive dissonance in
the friend condition among Asian-born Asian Canadians who
strongly identified with Asian culture. We examined this particular
group because they were the group who engaged in postdecisional
justification because of a threat to their interdependent cultural
ideals in Study 1, and they are, therefore, the group that should
show clear effects of an interdependent self-affirmation.
We used both independent and interdependent self-affirmation
manipulations and a no self-affirmation control group. We ex-
pected that only the interdependent self-affirmation would be
effective as an alternative to justifying their choices for Asian
Canadians because this self-affirmation manipulation would pro-
vide them with an opportunity to maintain their cultural ideals
even when confronting the feelings of threatened self-integrity tied
to the prospect of choosing the wrong gift coupons for important
in-group members. Thus, we expected that when choosing cou-
pons for their close friends, Asian Canadians presented with an
opportunity to affirm their interdependent selves would not show
postdecisional justification, whereas those who did not have such
an opportunity would continue to engage in postdecisional
justification.
Method
Newly Devised Self-Affirmation Manipulations
Given the fact that a particular aspect of the Asian self is threatened in
the experience of cognitive dissonance, we wished to create a new self-
affirmation manipulation tailored to affirm culturally ideal interdependent
self-concepts.
Conventional self-affirmation manipulations typically use a form of
value survey. In one common manipulation used by Fein and Spencer
(1997), participants select the most personally important value from a list
6
As a proxy variable for in-group identification at a personal level, we
examined the variable that measured how close participants felt toward
their friends (i.e., the felt closeness variable) on a 5-point scale with
anchors ranging from 1 (not at all close)to5(very close). A t test indicated
that there was no significant difference between two cultural groups (M ⫽
4.45, SD ⫽ 0.69 for European Canadians and M ⫽ 4.23, SD ⫽ 0.89 for
Japanese), t(80) ⫽ 1.27, ns. European Canadians indicated that they based
their food preference judgments on and chose as their free lunch coupon
recipients their best friends or romantic partners, whereas Japanese indi-
cated that they chose either friends or best friends. Using the felt closeness
variable as a continuous variable, we performed separate regression anal-
yses for Studies 1 and 2 to test whether feeling closer toward friends (i.e.,
stronger personal-level in-group identification) makes justification of the
choices made for the friends stronger. In Study 1 there were no significant
effects; however, in Study 2 there was a marginally significant main effect
for felt closeness, t(78) ⫽ 1.92, (
␤
⫽ .25), p ⬍ .06, and a significant
interaction between felt closeness and cultural groups on the spread of
alternatives, t(78) ⫽ 2.52, (
␤
⫽⫺.34), p ⬍ .02. The pattern of the
interaction suggests that the closer Japanese participants felt toward their
friends, the more they justified their choices for their friends, whereas the
closer European Canadians felt toward their friends, the less they justified
their choices for their friends. Considering this negative relation between
felt closeness and the degree of justification in Study 2 and our finding of
no relation between the two variables in Study 1 among European Cana-
dians, the speculation that stronger in-group identification at a personal
level leads to more postdecisional justification of the decisions for friends
seems unlikely, at least in the current free-choice paradigm. As a proxy
variable for in-group identification at a cultural level, we examined
strength of identification with Canadian culture, which was measured on an
11-point scale with anchors ranging from 0 (not at all)to10(very much).
Using this as a continuous variable, we performed a regression analysis for
Studies 1 and 2 separately. There were no significant main effects for the
identification variable or significant interactions between the identification
variable and the target of coupon choice on the spread of alternatives in
both studies, t’s ⬍ 1.0, ns. Thus, the other speculation that stronger
in-group identification at a cultural level (i.e., stronger individualistic
orientations) leads to less postdecisional justification of the decisions for
friends also seems unlikely, at least in our research.
302
HOSHINO-BROWNE ET AL.

of various values, such as business– economics, social life–relationships,
and religion–spirituality, and then explain, in written form, why the value
is important to them. Considering the culturally ideal interdependent self-
concept for East Asians, we created an interdependent self-affirmation
manipulation, noting that East Asians’ culturally important self-concepts
are their feelings of connectedness and belongingness. To affirm these
cultural ideals, we asked participants under an interdependent self-
affirmation manipulation to select the most important value for themselves
and their families and explain the reasons why they and their families share
that particular value. Parallel to this new interdependent self-affirmation
manipulation, we revised the conventional self-affirmation manipulation to
create an independent self-affirmation manipulation that provides people
with a more salient opportunity to affirm their uniqueness and indepen-
dence. Specifically, under the new independent self-affirmation manipula-
tion, participants explained why their selected values uniquely describe
who they are.
7
Participants
As in Study 1, we surveyed potential Asian Canadian participants during
a mass testing session at the University of Waterloo about their birth
countries and recruited only those who indicated an East Asian country as
their birth country. The mean length of stay in Canada among the partic-
ipants was 8.99 years. To measure the strength of these participants’
identification with Asian culture, we included the same two questions in
the mass testing session as we did in Study 1 (i.e., which ethnic group did
they most identify with and how much did they identify with that ethnic
group on an 11-point scale with anchors ranging from 0 [not at all]to10
[very much]). Although the median point on the scale in Study 1 was 9.0,
for this study, we selected Asian-born Asian Canadians who indicated that
they identified with Asian culture at 8.0 or above on the 11-point scale in
order to obtain more Asian Canadian participants. We also included the
same question as in Studies 1 and 2 to measure how much people identify
with Canadian culture on an 11-point scale with anchors ranging from 0
(not at all)to10(very much).
A total of 61 undergraduate students (36 women and 25 men) at the
University of Waterloo participated in the study. They received partial
course credit for their participation. Seven people who did not properly
follow the instructions given by the experimenter were excluded from the
data analyses. Thus, 54 participants (32 women and 22 men) were included
in the analyses.
Procedure
Participants were randomly assigned to one of three self-affirmation
conditions: no self-affirmation, independent self-affirmation, or interde-
pendent self-affirmation. A female experimenter conducted the experimen-
tal sessions, and all participants took part in the study individually. We
used the same free-choice materials and procedures that we used in Study
1. However, unlike Study 1, we used only the friend condition, in which
Asian Canadians chose coupons for a free lunch entre´e for their close
friends. In both the independent and the interdependent self-affirmation
conditions, the self-affirmation manipulation was given after participants
finished their Time 1 ranking and rating and completed a demographic
questionnaire. Thus, immediately before participants were presented with
two coupons for their 5th and 6th ranked entre´es to make their choices for
their friends, they were given an opportunity to affirm their independent or
interdependent selves. The experimenter presented the participants with the
self-affirmation manipulation in the form of a questionnaire purportedly
unrelated to our study, explaining that she was helping another researcher
to collect data. As described earlier, participants in the independent self-
affirmation condition were asked to choose one value that was the most
personally important to them from a list of six different values and to write
a paragraph about how the selected value uniquely describes who they
are. Those in the interdependent self-affirmation condition were asked to
select the most important value for both themselves and their family
members and to explain in a paragraph why they share those particular
values with their family members. Participants in the no self-affirmation
condition did not receive these manipulations. At the end of the experi-
ment, participants were thoroughly debriefed and received unexpected $5
in addition to their participation credit to replace the invalid gift
certificates.
Results
The dependent variable was the postdecisional justification ex-
pressed by the spread of alternatives, and we used the same
calculation as we did in Studies 1 and 2.
We hypothesized that because the interdependent self-
affirmation should help Asian Canadians to buffer or reduce dis-
sonance arousal, these participants would not show much postde-
cisional justification of the coupons they chose for their close
friends. On the other hand, we expected that those who did not
have an opportunity to affirm their interdependent self, that is,
those in the no self-affirmation condition or in the independent
self-affirmation condition, would engage in postdecisional
justification.
We conducted a one-way ANOVA, which indicated that the
mean differences among the three self-affirmation groups were
marginally significant, F(2, 51) ⫽ 2.69, p ⬍ .08 (see Figure 3). To
test our specific hypothesis, however, we conducted a planned
contrast to compare the mean of the two conditions for which we
expected a significant spread of alternatives (i.e., the no self-
affirmation and independent self-affirmation conditions) against
the mean of the interdependent self-affirmation condition for
which we did not expect such a spread of alternatives. As pre-
dicted, we obtained a significant mean difference between these
groups, F(1, 51) ⫽ 4.80, p ⫽ .03. That is, the mean of the spread
of alternatives of the no self-affirmation and independent self-
affirmation conditions was significantly different from the mean of
7
To ascertain the effectiveness of these new self-affirmation manipula
-
tions, we conducted a study in which we gave Asian-born Asian Canadians
either the independent or the interdependent self-affirmation manipulation
and then measured their state self-esteem (cf. McFarland & Ross, 1982)
immediately after they completed their essays. As expected, those Asian-
born Asian Canadians who presumably held strong interdependent self-
concepts, and thus could affirm their interdependent selves through an
interdependent self-affirmation opportunity, showed a significantly higher
level of state self-esteem than did those who were given an independent
self-affirmation opportunity. We also examined the values the Asian Ca-
nadians selected and the content of the essays they wrote in either of the
two conditions. The most popular value that the participants chose was
“social life–relationships” in both conditions. Also, the participants in both
conditions provided similar reasons in explaining why the selected value
uniquely described who they are (the independent self-affirmation) or why
they share the selected value with their family (see Hoshino-Browne,
Zanna, Spencer, & Zanna, 2004, for the details). The only major difference
between the two self-affirmation conditions was that these reasons were
framed in terms of self versus family in the respective conditions. It is
interesting that, despite the fact the majority of the Asian Canadian par-
ticipants selected the same values and listed similar reasons in both
conditions, the interdependent self-affirmation made such a positive dif-
ference in the participants’ state self-esteem.
303
CULTURAL GUISES OF COGNITIVE DISSONANCE

the spread of alternatives of the interdependent self-affirmation
condition.
We also conducted a series of follow-up contrast tests to com-
pare differences in the mean spread of alternatives among different
pairs of self-affirmation conditions. As expected, the difference
between the no-self-affirmation condition and the independent
self-affirmation condition was not significant, F(1, 51) ⫽ 0.66, ns.
Also, as expected, the difference between the no self-affirmation
condition and the interdependent self-affirmation condition was
significant, F(1, 51) ⫽ 5.19, p ⬍ .03. However, the difference
between the independent self-affirmation condition and the inter-
dependent self-affirmation condition was not significant, F(1,
51) ⫽ 2.18, p ⬍ .14.
8
We return to this unexpected result in the
Discussion section. In addition, we conducted one-sample t tests to
examine whether the spread of alternatives was significantly dif-
ferent from zero in each group. As predicted, the means of the
spread of alternatives in the no self-affirmation condition (M ⫽
0.69, SD ⫽ 1.32), t(16) ⫽ 2.16, was significantly different from
zero, p ⬍ .05, but those in the independent self-affirmation con-
dition (M ⫽ 0.36, SD ⫽ 1.03), t(17) ⫽ 1.50, and in the interde-
pendent self-affirmation condition (M ⫽⫺0.22, SD ⫽ 1.22),
t(18) ⫽ 0.78, were not (see Figure 3).
Discussion
The results of the third study partially supported our predictions.
Asian Canadians who did not have a chance to affirm themselves
engaged in postdecisional justification when they made choices for
their close friends. This result nicely replicated the pattern of
results in the friend condition among Asian Canadians in Study 1
and among Japanese participants in Study 2. Moreover, the inter-
dependent self-affirmation was an effective means for Asian Ca-
nadians to buffer or reduce their threatened feelings. Through an
interdependent self-affirmation opportunity, the Asian Canadian
participants seemed to be able to maintain their cultural ideals,
even after making choices for their close friends, and to signifi-
cantly reduce both their levels of dissonance arousal and their
concomitant need to justify their choices. By contrast, although
Asian Canadians in the independent self-affirmation condition had
an opportunity to affirm themselves, affirming an independent
self-concept was not an effective means of neutralizing the threat
from the possibility of making nonoptimal decisions for valued
in-group members. As a result, they tended to engage in postde-
cisional justification of the choices that they made for their close
friends.
Note, however, that the degree to which those in the indepen-
dent self-affirmation condition engaged in the justification fell
between the mean spread of alternatives obtained in the no self-
affirmation and interdependent self-affirmation conditions. Recall
that the mean difference between the independent self-affirmation
and interdependent self-affirmation conditions was not statistically
significant. If the independent self-affirmation opportunity was not
helpful at all to Asian Canadians, the degree to which they engaged
in postdecisional justification should have been equivalent to the
no self-affirmation condition. We speculated as to possible expla-
nations of this somewhat unexpected result.
One possible explanation for the more pronounced effect of the
interdependent self-affirmation on postdecisional justification could
be that the interdependent self-affirmation is somehow stronger or has
better efficacy than the independent self-affirmation, apart from the
fact that it affirms Asian Canadians’ culturally important interdepen-
dent self-concepts. If this is the case, then regardless of the types of
important self-concepts people possess, the interdependent self-
8
In this study we also asked participants whether they ate with the friend
to whom they were giving the coupon. On the basis of responses to this
question, we excluded 3 people who indicated that they never or rarely ate
with the friend. We reasoned that these people might not experience
dissonance because they would be less likely to face a threat to their
interdependent self-concept if the friend did not like their choice. In this
analysis, the results were slightly stronger. The main effect for the self-
affirmation manipulation was significant, F(2, 48) ⫽ 3.06, p ⫽ .05. The
condition means were M ⫽ 0.82 in the no self-affirmation condition, M ⫽
0.36 in the independent self-affirmation condition, and M ⫽⫺0.22 in the
interdependent self-affirmation condition. As in the analysis reported in the
text, the only significant difference between means was between the no
affirmation condition and the interdependent affirmation condition, F(1,
48) ⫽ 6.03, p ⬍ .01.
Figure 3. Study 3 postdecisional justification among Asian Canadians in three self-affirmation conditions
(dependent variable ⫽ mean spread of alternatives). An asterisk indicates that the mean spread of alternatives
of the condition is significantly different from zero.
304
HOSHINO-BROWNE ET AL.

affirmation should always have a stronger effect in reducing psycho-
logical discomfort or buffering people from self-threats. Then, exam-
ining European Canadian men in the self condition would be ideal to
test this possibility. The self condition has been demonstrated in
Studies 1 and 2 to threaten the independent self most, and European
Canadian men are thought to hold stronger independent self-concepts
than their female counterparts (see Cross & Madson, 1997, for a
review) or Asian counterparts (e.g., Markus & Kitayama, 1991;
Triandis, 1989, 1996). Thus, we conducted a small study by having 37
European Canadian men make choices for themselves after receiving
either an independent or an interdependent self-affirmation. The
experimental procedure was exactly the same as that used for the
self condition in Studies 1 and 2 and the independent and inter-
dependent affirmation conditions were exactly the same as those
used in Study 3.
The results demonstrated that an independent self-affirmation
eliminated the spread of alternatives (M ⫽⫺0.14), whereas the
interdependent self-affirmation did not (M ⫽ 0.80), t(35) ⫽ 2.75,
p ⬍ .01. This additional study suggests that the interdependent
self-affirmation manipulation had greater efficacy than the inde-
pendent manipulation seen in Study 3 that was not due simply to
the fact that the interdependent self-affirmation is a stronger affir-
mation for all people. In contrast, these results suggest that the
strength of the form of self-affirmation depends on the nature of
the self-concept that is being affirmed.
A more compelling theoretical account for the pattern of results
seen in Study 3 is that the independent self-affirmation was effective
for some but not all Asian Canadians. Some of the Asian Canadians
who strongly identified with Asian culture also strongly identified
with individualistic Canadian culture and thus may have embraced
independent self-concepts as well as more culturally inherent inter-
dependent self-concepts. If so, these biculturally identified Asian
Canadians may have been able to use the independent self-affirmation
to help them buffer or reduce threatened feelings when they experi-
enced dissonance, just as they seemed able to use the interdependent
self-affirmation effectively. By contrast, those who strongly identified
with Asian culture and not much, if at all, with Canadian culture (i.e.,
“monocultural” Asian Canadians) may not have embraced indepen-
dent self-concepts, and therefore, an opportunity to affirm indepen-
dent self-concepts would be functionally irrelevant to any effort to
maintain their interdependent cultural ideals. We examine this hy-
pothesis in Study 4.
Study 4: Postdecisional Justification and Independent
Self-Affirmation Among Bicultural and Monocultural
Asian Canadians
In Study 3, although Asian Canadians who were given a chance
to affirm their independent selves tended to show postdecisional
justification, the degree to which they engaged in the justification
was somewhat attenuated compared with the results obtained in
the no self-affirmation condition. We explain this finding by
hypothesizing that the independent self-affirmation would be ef-
fective for Asian Canadians if they were bicultural and thus
embraced both interdependent and independent self-concepts. We
also hypothesize that affirming the independent self would not be
of much help to those who were monocultural and thus held only
interdependent self-concepts.
It is important to note that whether Asian Canadians also iden-
tify with Canadian culture is only relevant to the independent
self-affirmation. In Study 3 after the interdependent self-
affirmation, we found no evidence of dissonance reduction. Given
that the participants were all Asian Canadians who held interde-
pendent self-concepts as their cultural ideals, this finding was
expected. Therefore, in the fourth study we excluded the interde-
pendent self-affirmation condition and examined the effects of the
independent self-affirmation only among Asian Canadians who
either identified only with their Asian culture (monoculturals) or
identified with both Canadian and Asian cultures (biculturals).
We expected that after making choices for their close friends,
Asian Canadians in the no self-affirmation condition would show
postdecisional justification, regardless of the strength of identifi-
cation with Canadian culture. In the independent self-affirmation
condition, in contrast, we predicted that bicultural Asian Canadi-
ans would show attenuated postdecisional justification because
these bicultural people are thought to have integrated Canadian
culture and its individualistic cultural ideals in their self-view, and
thus, they have independent self-concepts that could be affirmed.
We predicted that monocultural Asian Canadians, however, would
continue to engage in the justification. We did not expect the
independent self-affirmation to affect the monocultural Asian Ca-
nadians because we expected that they would not have strong
independent self-concepts if they did not identify with individual-
istic Canadian culture.
Method
Participants
As in Studies 1 and 3, we surveyed the birth countries of potential
participants, measured the strength of identification with Asian culture
during a mass testing session, and recruited Asian Canadians born in East
Asian countries. As in Study 3, we recruited Asian-born Asian Canadians
who indicated that they identified strongly with Asian culture (at 8.0 or
above on the 11-point scale). As in the previous three studies, to measure
the strength of identification with Canadian culture, we also asked potential
participants to rate the extent to which they identify with Canadian culture
on an 11-point scale with anchors ranging from 0 (not at all)to10(very
much). In addition, we asked participants the length of time that they had
spent in Canada, reasoning that the longer they had been in Canada the
stronger their identification would be. The mean length of stay in Canada
among our participants was 9.19 years.
A total of 120 undergraduate students (70 women and 50 men) at the
University of Waterloo participated. They received either a partial course
credit or $5 for their participation. Seventeen people who did not follow the
instructions given by the experimenter were excluded from the data anal-
yses. Also, 1 participant was excluded because his response on the main
dependent variable was over 3.5 standard deviations from the mean, and
thus regarded as an outlier, and 1 participant was excluded because
although she was born in Asia, she moved to Canada within days of her
birth. Thus, 101 participants (60 women and 41 men) were included in the
analyses. The identification strength with Canadian culture correlated
significantly with the length of stay in Canada, r(97) ⫽ .54, p ⬍ .001
(based on 99 participants, as 2 participants did not indicate the length of
stay in Canada in Study 4). Thus, the longer Asian Canadians stayed in
Canada, the more strongly they identified with Canadian culture. Given the
high correlation of these two variables, we standardized both measures and
combined them as an index of identification with Canadian culture.
305
CULTURAL GUISES OF COGNITIVE DISSONANCE

Procedure
Participants were randomly assigned to either the no self-affirmation or
the independent self-affirmation condition. As in Study 3, a female exper-
imenter conducted individual sessions for the participants. Also as in Study
3, we used only the friend condition in which Asian Canadians chose
coupons for a free lunch entre´e for their close friends. The materials and
procedures were the same as those used in Study 3. In the independent
self-affirmation condition, the self-affirmation manipulation was given as
a questionnaire under the guise of helping another researcher just before
participants were presented with two coupons that corresponded to their
5th and 6th ranked entre´es. Participants in the independent self-affirmation
condition were asked to choose one value that was the most personally
important to them from a list of six different values and write a short
paragraph why their selected values uniquely described who they are.
Participants in the no self-affirmation condition did not receive this ma-
nipulation. As in the previous three studies, participants were thoroughly
debriefed and received an unexpected $5 in addition to their participation
credit or remuneration to replace the invalid gift certificates.
Results
The dependent variable was the postdecisional justification ex-
pressed by the spread of alternatives and was calculated according
to the same procedure used in the previous three studies.
Because our measure of identification with Canadian culture
was a continuous variable, we analyzed the results using multiple
regression. We regressed participants’ spread of alternatives onto
their self-affirmation condition, their identification with Canadian
culture, and the interaction of these two variables. This analysis
revealed a marginally significant effect for the interaction between
self-affirmation condition and identification with Canadian cul-
ture,
␤
⫽⫺.25, t(95) ⫽ 1.70, p ⫽ .09. The predicted means at one
standard deviation above and below the mean on identification
with Canadian culture by self-affirmation condition are depicted in
Figure 4.
The simple slope predicting participants’ spread of alternatives
from their identification with Canadian culture was significant in
the independent self-affirmation condition,
␤
⫽⫺.34, t(95) ⫽
2.10, p ⬍ .05, but was not significant in the no self-affirmation
condition,
␤
⫽ .02, ns. Thus, the more participants identified with
Canadian culture in the independent self-affirmation condition, the
less postdecisional justification they engaged in. In contrast, iden-
tification with Canadian culture was unrelated to postdecisional
justification in the no self-affirmation condition.
In addition, at one standard deviation above the mean in iden-
tification with Canadian culture (i.e., among people high in iden-
tification with Canadian culture and high in identification with
Asian culture, or biculturals) the predicted mean of those in the no
self-affirmation condition (M
pred
⫽ 0.85) was significantly greater
than the predicted mean for those in the independent affirmation
condition (M
pred
⫽⫺0.03), t(95) ⫽ 3.33, p ⬍ .01.
9
Thus, among
bicultural participants, those in the independent self-affirmation
condition were predicted to show less postdecisional justification
than those in the no self-affirmation condition.
Discussion
Replicating the results of the previous three studies, in Study 4
our results demonstrated that Asian Canadians who did not have an
opportunity to affirm themselves engaged in postdecisional justi-
fication when they made choices for their close friends. The
strength of their identification with Canadian culture did not matter
in this situation. However, the strength of identification with
9
We also analyzed the data after excluding 6 people who indicated that
they never or rarely ate with the friend to whom they were giving the
coupon. In this analysis, the results were somewhat stronger. The interac-
tion between the self-affirmation and identification with Canadian culture
was significant,
␤
⫽⫺.32, t(89) ⫽⫺2.14, p ⬍ .05. The simple slope
predicting spread of alternatives was significant in the independent self-
affirmation condition,
␤
⫽⫺.44, t(89) ⫽ 2.72, p ⬍ .01, but was not
significant in the no self-affirmation condition (
␤
⫽ .02, ns). Finally, at one
standard deviation above the mean on identification with Canadian culture,
the predicted mean for bicultural Asian Canadians in the no self-
affirmation condition (M
pred
⫽ 0.76) was significantly greater than the
predicted mean in the independent self-affirmation condition (M
pred
⫽
⫺0.27), t(89) ⫽ 3.96, p ⬍ .001.
Figure 4. Study 4 postdecisional justification among Asian Canadians as predicted by self-affirmation
condition and identification with Canadian culture (dependent variable ⫽ predicted mean spread of alternatives).
306
HOSHINO-BROWNE ET AL.

Canadian culture did matter when Asian Canadians could affirm
their independent selves. Bicultural Asian Canadians who strongly
identified with both Asian and Canadian cultures were able to use
the independent self-affirmation opportunity effectively to buffer
or reduce threatened feelings and maintain their cultural ideals
even after making choices for their close friends. They did not,
therefore, need to justify their choices to reduce their dissonance
arousal. On the other hand, the independent self-affirmation op-
portunity was not helpful to monocultural Asian Canadians, who
did not strongly identify with individualistic Canadian culture and
thus did not embrace independent self-concepts. These participants
continued to engage in postdecisional justification of the choices
that they made for their close friends.
The bicultural Asian Canadians’ results led to a consideration of
the notion of “frame switching” (Hong, Morris, Chiu, & Benet-
Martinez, 2000, p. 709) or the “alternation” model in the accul-
turation process (LaFramboise, Coleman, & Gerton, 1993, p. 399).
The versatility that bicultural Asian Canadians demonstrated in the
use of self-affirmation suggests that they can switch or alternate
the psychological frames of reference that originated in two dif-
ferent cultures to act appropriately according to the social envi-
ronment in which they are situated. They seem to be able to cross
with ease the boundary between an interdependent self, which is a
more indigenous Asian self-concept, and an independent self,
which is a more predominant North American self-concept.
10
General Discussion
We began this research by asking whether there is cultural
variation in the experience of cognitive dissonance and in the
subsequent tendency to justify or rationalize individual decision-
making. On the basis of the four studies delineated above, we
found the answer was both “yes” and “no.” It was “no” in the sense
that at least in the two different cultural groups we examined,
people regardless of their cultural backgrounds experienced cog-
nitive dissonance after making choices that were important to them
and subsequently engaged in efforts to justify their decisions.
However, the answer was “yes” in the sense that the situation in
which they experienced cognitive dissonance and justified their
decisions was different across these cultural groups.
Our research demonstrates that, just like Westerners, East
Asians do engage in rationalization of their decisions. Across the
four studies conducted, we consistently found that East Asians
justified their choices when those choices pertained to their cul-
turally important self-concepts. To ascertain the extent to which
they engaged in postdecisional justification, we conducted a meta-
analysis across the four studies on the spread of alternatives for the
conditions in which Asian Canadians or Japanese made their
choices for their close friends without having any self-affirmation
opportunities. The meta-analysis indicated that the evidence is not
only consistent but extraordinarily strong, z ⫽ 5.63, p ⬍ .0001.
East Asians indeed engaged in dissonance reduction through ra-
tionalization of their decisions when their culturally important
self-concepts were threatened by the need to make choices for their
in-group members.
Our finding that Asian Canadians and Japanese justified their
choices for their close friends is consistent with recent findings by
Kitayama, Conner Snibbe, Markus, and Suzuki (2004). In a series
of cross-cultural studies, Kitayama et al. found that European
Americans showed the usual justification of their choices of CDs
across conditions, whereas their Japanese counterparts engaged in
justification of their choices when they were reminded of self-
relevant others by interpersonal priming. These researchers argued
that the saliency of self-relevant others had an effect on the
Japanese participants because such social cues evoked “interper-
sonal worries” about possible criticism or rejection by the self-
relevant others, which led to cognitive dissonance. The results of
the current studies together with Kitayama et al.’s findings provide
strong support for the argument that interpersonal concerns lead
East Asians to experience cognitive dissonance and subsequently
engage in dissonance reduction.
In Kitayama et al.’s (2004) study, the Japanese participants who
did not receive the interpersonal prime did not justify their choices
of CDs, presumably because their interpersonal worries were not
evoked. This result is also consistent with our finding that Asian
Canadians and Japanese who chose the gift certificate for them-
selves did not justify their choices because they did not have any
interpersonal concerns. These particular groups in our research and
10
Consistent with this reasoning, one might propose that bicultural
Asian Canadians, who presumably hold both independent and interdepen-
dent self-concepts, justify their choices for themselves more than monocul-
tural Asian Canadians. Considering this possibility, we regressed the
spread of alternatives onto the condition that participants were in (self vs.
friend), their identification with Asian culture (using the same measure that
was reported in Study 1), and their identification with Canadian culture
(using the same measure that was reported in Study 4) for all 93 Asian
Canadians in Study 1. This analysis revealed a significant three-way
interaction between these variables,
␤
⫽⫺.35, t(85) ⫽ 2.09, p ⬍ .05.
Examination of the predicted means on the basis of this analysis is
consistent with our reasoning. In the friend condition both biculturals (i.e.,
at one standard deviation above the mean on both Asian identification and
Canadian identification), M
pred
⫽ 0.95, and monoculturals (i.e., at one
standard deviation above the mean on Asian identification and one stan-
dard deviation below the mean on Canadian identification), M
pred
⫽ 0.72,
were predicted to show a substantial spread of alternatives, and there was
no difference between these two predicted means, t(85) ⬍ 1. In the self
condition, however, both biculturals, M
pred
⫽ 0.31, and monoculturals,
M
pred
⫽⫺0.54, showed less dissonance reduction, but importantly, the
predicted mean for biculturals was significantly higher than the predicted
mean for monoculturals, t(85) ⫽ 2.59, p ⬍ .05. Thus, biculturals were
predicted to show more justification of their choices for themselves than
monoculturals, but both of these groups were predicted to show more
justification for their friends than for themselves.
In contrast, those who identified with Canadian culture only (i.e., at one
standard deviation below the mean on Asian identification and one stan-
dard deviation above the mean on Canadian identification) and those who
identified with neither culture (i.e., at one standard deviation below the
mean on both Asian identification and Canadian identification) were both
predicted to show more spread of alternatives in the self condition, M
pred
⫽
0.63 and M
pred
⫽ 1.60, respectively, than in the friend condition, M
pred
⫽
0.27 and M
pred
⫽⫺0.87, respectively. The pattern of predicted means for
those identified with Canadian culture was very similar to the means
obtained for European Canadians, whereas those identified with neither
culture were predicted to show an especially large spread of alternatives in
the self condition and a reversal of the spread of alternatives in the friend
condition. However, these latter findings should probably be interpreted
with caution because only 9 of the 93 participants in the study actually fell
below the median on identification with both Asian and Canadian cultures.
307
CULTURAL GUISES OF COGNITIVE DISSONANCE

in Kitayama et al.’s work replicated Heine and Lehman’s (1997)
result that Japanese did not rationalize their choices of CDs.
Another interesting finding in the Studies 1 and 2 came from
European Canadians in the friend condition. The European Cana-
dians seemed to experience some dissonance arousal when they
made their choices for their close friends, but the dissonance they
experienced was not to the same degree as when they made their
choices for themselves, nor as much as that experienced by East
Asians who made choices for their friends. In contrast, Nel et al.
(1969) and Norton et al. (2003) demonstrated that individualistic
North Americans experience dissonance when others are involved
in the induced compliance paradigm. Is the current finding among
European Canadians in the friend condition specific to a free-
choice paradigm or the free-choice materials used? It is possible
that the two research paradigms threaten different aspects of an
interdependent self and that Westerners’ interdependent selves are
more sensitive to attitude-behavior consistency involved in in-
duced compliance than they are to making a decision for a friend.
Further investigation in relation to these considerations is
warranted.
Alternative Explanations for East Asians’ Justification in
the Friend Condition
Regarding East Asians’ postdecisional justification in the friend
condition, there may be at least three alternative explanations as to
why people justify their choices for their close friends. One ex-
planation is that participants used their own food preferences,
pretending that they were their friends’, and thus their choices of
the gift certificates and their subsequent postdecisional justifica-
tion were for themselves rather than for their friends.
To preempt this possibility, we took a number of precautions in
the friend condition to ensure that participants did not use their
own preferences in place of their friends’ preferences. After par-
ticipants indicated that they had selected a friend, the experimenter
confirmed with participants that the friend was a close friend and
that they knew the friend’s food preferences well. Each question-
naire explicitly stated whose preferences they needed to base their
judgment on, and the experimenter underscored it in the instruc-
tions. After they chose gift certificates for their friends, they were
asked whether they were sure that their friends would like the
choices, and then they had to write down their friends’ names as
well as their own on the gift certificates. They were also asked a
probing question at the end of the experimental session regarding
how they thought the gift certificates might be delivered to their
friends. Thus, we tried painstakingly to ensure that participants in
the friend condition considered their friends’ preferences and used
them for their decisions and ratings.
Another alternative explanation is that the dissonance experi-
enced by East Asians in the friend condition is due to vicarious
dissonance. Recall that Norton et al. (2003) found that witnessing
in-group members’ attitude-inconsistent behavior could lead to an
experience of vicarious dissonance. In the friend condition in the
current research, if participants had witnessed their friends making
decisions and seen the impact of the decisions on the friends, they
might have experienced vicarious dissonance. One might even
argue that they do not need to witness their friends making deci-
sions to experience vicarious dissonance; they only need to em-
pathize with the situation in which their friends make their deci-
sions. However, the participants in this research neither witnessed
their friends making their decisions nor were they asked to imagine
themselves in their friends’ place. As outlined above, they were
instructed to make their own decisions for their friends on the basis
of their beliefs about their friends’ preferences, not as their friends
might actually make decisions. Thus, we believe that the current
friend condition would make it difficult for people to role-play as
their friends. Consequently, it would be difficult for them to
experience dissonance vicariously. Therefore, we believe that the
postdecisional justification we obtained in the friend condition was
due to psychological discomfort elicited by actually making
choices for close friends rather than to vicarious dissonance
evoked by imagining their friends being in the decision-making
situation.
Still, we note that these precautions could not entirely eliminate
a third possibility of the unintended social projection. That is,
participants might have inadvertently mistaken their own prefer-
ences for their perceptions of their friends’ preferences, and there-
fore, the ratings of preferred items and the justification of the
chosen gift certificate were the result of social projection of the
participants’ own preferences. After participants indicated their
beliefs about their friends’ preferences, we did not ask them to rate
the same items in accordance with their own preferences. If we had
included this self-rating measure, we would have been able to
distinguish the friend preference rating and the self preference
rating and assess the extent to which social projection occurred.
However, what really matters in interpersonal dissonance is that
people believed that their ratings and decisions were based on their
perceptions of their friends’ preferences and justified their inter-
personal perceptions or beliefs about their friends’ preferences
regardless of whether it was social projection. With the procedures
and precautions described above, we believe that we successfully
measured justification of their beliefs or change of their beliefs
about their friends’ preferences.
Efficacy of Self-Affirmation Among Asian Canadians
In the third and fourth studies, we found that self-affirmation, as
an alternative to postdecision justification, can be an effective
means for Asian Canadians to counter threats to their self-
concepts. In keeping with past research with Westerners, we found
that providing Asian Canadians with an opportunity to affirm
themselves reduced their need to engage in dissonance reduction
as a strategy to restore threatened self-integrity. Our findings
provide a strong test for the utility of self-affirmation processes.
Although East Asians’ culturally valued self-concepts may be
quite different from those of Westerners, affirming the important
aspects of the self is equally effective for both cultural groups.
Equally important to note is that our research also demonstrates
that this self-affirmation finding depends substantially on both the
locus and strength of Asian Canadians’ cultural identification.
Specifically, we found that providing Asian Canadians who
strongly identified with Asian culture with an opportunity to affirm
their interdependent self-concepts significantly reduced their need
to counter threatened self-integrity by justifying the choices they
made for close friends. This finding on the effect of interdependent
self-affirmation is consistent with Heine and Lehman’s (1997)
speculation that Japanese individuals might experience cognitive
dissonance if their interdependent self was somehow threatened
308
HOSHINO-BROWNE ET AL.

and that they might alleviate such threats by affirming their
interdependence.
Our findings with respect to the independent self-affirmation
manipulation for Asian Canadians were more nuanced. In partic-
ular, providing bicultural Asian Canadians who identified strongly
with both Asian and Canadian cultures with an independent self-
affirmation opportunity reduced their need to counter threatened
self-integrity by engaging in postdecisional justification. In con-
trast, providing monocultural Asian Canadians, who only identi-
fied strongly with Asian culture, with the same opportunity did not
reduce their tendency to justify their choices. Given that bicultural
Asian Canadians embrace both interdependent and independent
self-concepts as their cultural ideals, the independent (as well as
the interdependent) self-affirmation manipulation served as an
effective means for countering threatened self-integrity.
The results of Studies 3 and 4 also nicely address another
alternative explanation for the Studies 1 and 2 friend condition
findings. One might argue that East Asians engage in postdeci-
sional justification for the choices they made for their close friends
because of impression management or their effort to appear that
they care about their choices for their friends. If East Asians’
justification is due merely to impression management, then self-
affirmation is not likely to have any effect on the degree to which
East Asians justify their choices made for their friends. Thus, we
believe that East Asians justify their choices for their friends not
under the pretense of being considerate to their friends but because
they experience interpersonal dissonance by making choices for
their friends, and they try to resolve the psychological discomfort
they experience by rationalizing their choices.
Asian Canadians’ Biculturalism and Fluidity of
Self-Affirmation
Two aspects of our findings are particularly noteworthy. First,
the strength of East Asians’ identification with Asian culture
seems to be of crucial importance. Although the Asian Canadian
participants in this research were all born in East Asian countries,
the degree to which they identified with Asian culture varied
greatly. Moreover, this variability in the strength of Asian cultural
identification was an important factor in their experience of cog-
nitive dissonance and their subsequent justification of their
choices. The results we obtained provide a clear link between the
cultural identification and culturally ideal self-concepts people
hold. Although East Asians who only weakly identified with Asian
culture are ethnically Asian, their cultural ideals seemed to be
similar to the independent self-concepts of Westerners, at least in
our research. Thus, they experienced cognitive dissonance and
reacted to it in the same situation as Westerners did. In contrast,
throughout the four studies described above, it is clear that East
Asians who strongly identified with Asian culture held interdepen-
dent self-concepts as their cultural ideals. They experienced cog-
nitive dissonance when their interdependent self-concepts were
threatened, and they engaged in dissonance reduction through
either justification of their choices for their close others or inter-
dependent self-affirmation.
Furthermore, equally important evidence for the link between
the cultural identification and the culturally ideal self-concepts
people hold has emerged from East Asians’ identification with
individualistic Canadian culture. Bicultural East Asians who
strongly identify with both Asian and Canadian cultures seem to
embrace both interdependent and independent self-concepts as
their cultural ideals. Unlike monocultural East Asians, who
strongly identify only with Asian culture, bicultural East Asians
are capable of using the independent self-affirmation as a means of
reducing dissonance, when their interdependent self is threatened.
They seem to be able to smoothly cross the boundary of two
cultures and readily switch between the two cultural mind-sets
(Hong et al., 2000).
Note that not only do these findings provide evidence for the
relation between cultural identification and cultural ideals, but they
also demonstrate the “fluidity of self-affirmation processes”
(Steele, 1988, p. 267). Steele and his colleagues (Spencer et al.,
1993; Steele, 1988; Steele et al., 1993) argued that individuals
have a pool of positive self-concepts within a large self-system,
and they can affirm themselves and maintain an overall self-
integrity by using some of these positive attributes that are not
necessarily under threat. Our findings provide support for this
argument.
The first piece of evidence in support of the fluidity of self-
affirmation processes is that both monocultural and bicultural East
Asians who experience a threat to a particular interdependent
self-concept (i.e., a threat to their friendships) can use another
interdependent self-concept (i.e., family relationships) to affirm
themselves. Namely, when they experience a threat to their rela-
tionships with their close friends by the possibility of making an
inconsiderate choice or the wrong choice of Chinese food, they can
affirm themselves by using an important value that they share with
their family.
The second piece of evidence for the fluidity is that bicultural
East Asians, whose interdependent self-concepts are threatened by
interpersonal concerns that arise from the possibility of making
nonoptimal choices for their close others, can affirm themselves
using either their independent or their interdependent self-
concepts. In fact, they seem to have more resources with which to
self-affirm than monocultural East Asians who can affirm them-
selves only using interdependent self-concepts.
In future research, it would be interesting to demonstrate the
versatility of such bicultural East Asians by showing that interde-
pendent self-affirmation is useful when their independent selves
are threatened. It is not difficult to imagine that the resourcefulness
of biculturally identified individuals could have a number of
implications for their mental health and that such resourcefulness
can lead to intriguing investigations in the field of health psychol-
ogy among immigrants.
Another interesting area of future cross-cultural research would
be affective reactions in cognitive dissonance. Although both
Westerners and East Asians seem to experience psychological
discomfort, the constellation of emotions that they experience
around their discomfort or dissonance arousal could be quite
different, as the situations in which these two cultural groups
experience cognitive dissonance have been found to be different in
our research. Whereas Westerners might experience more self-
related emotions, East Asians might experience more interpersonal
emotions. Because very little research on affective reactions has
been done within the framework of cognitive dissonance in gen-
eral, such cross-cultural investigations may prove to be beneficial.
309
CULTURAL GUISES OF COGNITIVE DISSONANCE

Conclusions
Our findings point to an important perspective on cross-cultural
research. Many psychological and social phenomena that seem to
be cross-culturally variable on the surface might have the same
underlying mechanisms; many other phenomena that seem to be
cross-culturally similar on the surface may have cross-culturally
different psychological functions or may arise from culture-
specific factors. Thus, when examining cross-cultural aspects of
psychology, it is important to consider each phenomenon with its
underlying mechanism or mechanisms and carefully examine the
role of culture—not just on the surface, but at its deeper level as
well.
In this research, we have examined the self-system of East
Asians and Westerners from two different angles, that is, self-
threat in the cognitive dissonance processes and culturally ideal
self-image maintenance in the self-affirmation processes, and dem-
onstrated both cross-cultural similarity and variation. We believe
that through cross-cultural research our work contributes to a more
thorough understanding of cognitive dissonance. Moreover, we are
unaware of other research demonstrating that self-affirmation the-
ory, which was developed in North America, is cross-culturally
viable. We believe that our research contributes to advancement in
the understanding of self-affirmation theory as well as in the
understanding of the East Asian self-system.
References
Brehm, J. W. (1956). Postdecision changes in the desirability of alterna-
tives. Journal of Abnormal and Social Psychology, 36, 384 –389.
Cross, S. E., & Madson, L. (1997). Models of the self: Self-construals and
gender. Psychological Bulletin, 122, 5–37.
Fein, S., & Spencer, S. J. (1997). Prejudice as self-image maintenance:
Affirming the self through derogating others. Journal of Personality and
Social Psychology, 73, 31–44.
Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA:
Stanford University Press.
Heine, S. J., & Lehman, D. R. (1997). Culture, dissonance, and self-
affirmation. Personality and Social Psychology Bulletin, 23, 389 – 400.
Heine, S. J., Lehman, D. R., Markus, H. R., & Kitayama, S. (1999). Is there
a universal need for positive self-regard? Psychological Review, 106,
766 –794.
Hong, Y-Y., Morris, M. W., Chiu, C-Y., & Benet-Martinez, V. (2000).
Multicultural minds: A dynamic constructivist approach to culture and
cognition. American Psychologist, 55, 709 –720.
Hoshino-Browne, E., Zanna, A. S., Spencer, S. J., & Zanna, M. P. (2004).
Investigating attitudes cross-culturally: A case of cognitive dissonance
among East Asians and North Americans. In G. Haddock & G. R. Maio
(Eds.), Contemporary perspectives on the psychology of attitudes (pp.
375–397). New York: Psychology Press.
Kitayama, S., Conner Snibbe, A., Markus, H. R., & Suzuki, T. (2004). Is
there any free choice? Self and dissonance in two cultures. Psycholog-
ical Science, 15, 527–533.
LaFramboise, T., Coleman, H. L. K., & Gerton, J. (1993). Psychological
impact of biculturalism: Evidence and theory. Psychological Bulletin,
114, 395–412.
Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications
for cognition, emotion, and motivation. Psychological Review, 98, 224 –
253.
McFarland, C., & Ross, M. (1982). Impact of causal attributions on
affective reactions to success and failure. Journal of Personality and
Social Psychology, 43, 937–946.
Nel, E., Helmreich, R., & Aronson, E. (1969). Opinion change in the
advocate as a function of the persuasibility of his audience: A clarifica-
tion of the meaning of dissonance. Journal of Personality and Social
Psychology, 12, 117–124.
Norton, M. I., Monin, B., Cooper, J., & Hogg, M. A. (2003). Vicarious
dissonance: Attitude change from the inconsistency of others. Journal of
Personality and Social Psychology, 85, 47– 62.
Rosenthal, R. (1991). Meta-analytic procedures for social research. New-
bury Park, CA: Sage.
Sakai, H. (1981). Induced compliance and opinion change. Japanese
Psychological Research, 23, 1–8.
Sakai, H., & Andow, K. (1980). Attribution of personal responsibility and
dissonance reduction. Japanese Psychological Research, 22, 32– 41.
Spencer, S. J., Josephs, R. A., & Steele, C. M. (1993). Low self-esteem:
The uphill struggle for self-integrity. In R. F. Baumeister (Ed.), Self-
esteem: The puzzle of low self-regard (pp. 21–36). New York: Plenum
Press.
Steele, C. M. (1988). The psychology of self-affirmation: Sustaining the
integrity of the self. In L. Berkowitz (Ed.), Advances in experimental
social psychology (Vol. 21, pp. 261–302). San Diego, CA: Academic
Press.
Steele, C. M., & Liu, T. J. (1983). Dissonance processes as self-
affirmation. Journal of Personality and Social Psychology, 45, 5–19.
Steele, C. M., Spencer, S. J., & Lynch, M. (1993). Self-image resilience
and dissonance: The role of affirmational resources. Journal of Person-
ality and Social Psychology, 64, 885– 896.
Triandis, H. C. (1989). The self and social behavior in differing cultural
contexts. Psychological Review, 96, 506 –520.
Triandis, H. C. (1996). The psychological measurement of cultural syn-
dromes. American Psychologist, 51, 407– 415.
Received November 21, 2003
Revision received April 18, 2005
Accepted April 22, 2005 䡲
310
HOSHINO-BROWNE ET AL.