Society for Consumer Psychology

The Influence of Culture on Consumer Impulsive Buying Behavior

Author(s): Jacqueline J. Kacen and Julie Anne Lee
Source: Journal of Consumer Psychology, Vol. 12, No. 2 (2002), pp. 163-176
Published by: Society for Consumer Psychology
Stable URL: http://www.jstor.org/stable/1480261 .
Accessed: 15/10/2014 16:01
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Society for Consumer Psychology is collaborating with JSTOR to digitize, preserve and extend access to
Journal of Consumer Psychology.

http://www.jstor.org

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

JOURNALOF CONSUMER PSYCHOLOGY,12(2), 163-176

Copyright? 2002, LawrenceErlbaumAssociates, Inc.

The Influence of Cultureon Consumer

ImpulsiveBuying Behavior
JacquelineJ. Kacen
Departmentof BusinessAdministration
Universityof Illinois at Urbana-Champaign

Julie Anne Lee

Departmentof Marketing

Impulsebuyinggeneratesover$4 billionin annualsalesvolumein theUnitedStates.Withthe

growthof e-commerceandtelevisionshoppingchannels,consumershaveeasyaccessto imbutlittleis knownaboutthissudden,compelling,hedonically
pulsepurchasing
opportunities,
in
behavior
non-Western
cultures.Yetculturalfactorsmoderate
complexpurchasing
manyasnormative
influences,
pectsof consumer'simpulsivebuyingbehavior,includingself-identity,
thesuppression
of emotion,andthepostponement
of instantgratification.
Froma multi-country
surveyof consumersin Australia,UnitedStates,HongKong,Singapore,andMalaysia,our
andindividual
culanalysesshowthatbothregionallevelfactors(individualism-collectivism)
turaldifferencefactors(independent
influence
-interdependent
self-concept)systematically
impulsive purchasingbehavior.

Impulsiveconsumerbuying behavioris a widely recognized

phenomenonin the UnitedStates.It accountsfor upto 80%of
all purchasesin certainproductcategories(Abrahams,1997;
Smith, 1996), andit has been suggestedthatpurchasesof new
products result more from impulse purchasing than from
priorplanning(Sfiligoj, 1996). A 1997 studyfoundthatanestimated $4.2 billion annual store volume was generatedby
impulse sales of items such as candy and magazines
(Mogelonsky, 1998). Paco Underhill,authorof WhyWeBuy:
TheScience of Shopping(1999), affirmsthatmanypurchases
are being made on the premises of stores themselves as customers give in to their impulses. Furthermore,technologies
such as television shoppingchannelsandthe Internetexpand
consumers' impulse purchasing opportunities, increasing
both the accessibility to productsand services and the ease
with which impulse purchasescan be made.
Impulsive buying behavior is a sudden, compelling,
hedonicallycomplex purchasingbehaviorin which the rapidity of the impulse purchase decision process precludes

Requestsfor reprintsshouldbe sent to JacquelineJ. Kacen,Universityof

Illinois at Urbana-Champaign,1206 S. Sixth Street,Champaign,IL 61820.
Email:kacen@uiuc.edu

thoughtful,deliberateconsiderationof all informationand

choice alternatives(Bayley & Nancorrow,1998;Rook 1987;
Thompson,Locander,& Pollio, 1990;Weinberg& Gottwald,
1982). This descriptionis largely based on interviews and
surveys of Westerners.
The growth of e-commerce and the increasing consumer-orientationof many societies aroundthe world offer
expanding occasions for impulse purchasing, but little is
known aboutimpulsive buying behaviorin non-Westernsocieties. Most of the researchon impulse buying focuses on
consumersin the United States.A few studieshave looked at
consumers in Great Britain (Bayley & Nancarrow, 1998;
Dittmar,Beattie, & Friese, 1995; McConatha,Lightner,&
Deaner, 1994), and South Africa (Abratt& Goodey, 1990)
andhave foundthatUnited Statesconsumerstendto be more
impulsive than comparableBritish and South African samples. However,none of these studiesexaminedexplicitly the
effect of culturalfactorson impulse buying behavior.
A recent special issue of the Journal of ConsumerPsychology dealt with culturalissues demonstratingthe growing
interest in cultural differences in consumer behavior and
highlightedthe importanceof understandingthe culturalcontext of consumerbehaviorin an increasingglobalized marketplace (Maheswaran& Shavitt, 2000). We believe that

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

164

KACEN AND LEE

culturalfactorssignificantlyinfluenceconsumers'impulsive
buying behavior. Specifically, the theory of individualism
andcollectivism holdsimportantinsightsaboutconsumerbehaviorthatcan help us to gain a better,morecompleteunderstanding of the impulsive buying phenomenon.Consistent
with this interestin culturaldifferences,this articleexamines
the effect of regionallevel (individualist-collectivist)andindividual difference level (independent-interdependent
self-concept) culturalfactors on consumers' impulsive buying behavior.Utilizing a multi-countrysampleof over a thousand consumersfrom both Westernand Easterncultures,we
investigate how culture systematically moderates impulse
buyingbehavior.This is especially importantas shoppingis a
majorleisureactivityin manyEastAsian countries(Wong &
Ahuvia, 1998), includingSingapore,Hong Kong,andJapan.

IMPULSE BUYING
Impulsebuyingis defined as "anunplannedpurchase"thatis
characterizedby "(1) relatively rapiddecision-making,and
(2) a subjectivebias in favorof immediatepossession"(Rook
& Gardner,1993, p. 3; see also Rook, 1987; Rook & Hoch,
1985). It is describedas more arousing,less deliberate,and
more irresistiblebuying behaviorcomparedto plannedpurchasing behavior. Highly impulsive buyers are likely to be
unreflectivein their thinking,to be emotionally attractedto
the object, and to desire immediate gratification(Hoch &
Loewenstein, 1991; Thompsonet al., 1990). These consumers often pay little attention to potential negative consequences that may result from their actions (Hoch &
Loewenstein, 1991; Rook, 1987; see also O'Guinn& Faber,
1989).
Previous research conducted in the United States and
Great Britain (individualistcultures) has shown that many
factorsinfluence impulsivebuying behavior:the consumer's
mood or emotional state (Donovan, Rossiter,Marcoolyn,&
Nesdale, 1994; Rook, 1987; Rook & Gardner,1993: Weinberg & Gottwald, 1982), trait buying impulsiveness (Puri,
1996; Rook & Fisher, 1995; Weun, Jones, & Beatty, 1998),
normativeevaluation of the appropriatenessof engaging in
impulsebuying(Rook & Fisher, 1995), self-identity(Dittmar
et al., 1995), and demographicfactors, such as age (e.g.,
Bellenger, Robertson,& Hirschman,1978; Wood, 1998).
Several studies demonstrate the effect of consumers'
moods and affective states on impulsive buying behavior.
Rook and Gardner(1993) found that consumers' positive
moods were moreconducive to impulsive buying thannegative moods, although impulse buying occurred under both
types of moods. Beatty and Ferrell (1998) also found that a
consumer's positive mood was associated with the urge to
buy impulsively, while the impulse buyers in Weinbergand
Gottwald's (1982) study were more "emotionalized"than
nonbuyers.Donovanet al. (1994) discovereda positive association between consumers'feelings of pleasurein the shop-

ping environmentand impulse buying behavior.In each of

these studies,pleasurablefeelings led to increasedunplanned
spending.
Cognitive, clinical, social, developmental,and consumer
psychologistshave studiedthe generaltraitof impulsiveness
and impulse control (Eysenck & Eysenck, 1978; Eysenck,
Pearson,Easting,& Allsopp, 1985;Helmers,Young, & Pihl,
1995;Hilgard,1962;Logue & Chavarro,1992;Logue, King,
Cavarro, & Volpe, 1990; Mischel, 1961; Puri, 1996;
Rawlings,Boldero,& Wiseman, 1995;Rook & Fisher, 1995;
Weun et al., 1998). Traitimpulsiveness is characterizedby
unreflectiveactions(Eysencket al., 1985) andis significantly
correlatedwith thrill-seeking(Weun et al., 1998), and the
psychologicalneedto maintaina relativelyhigh level of stimulation (Gerbing,Ahadi, & Patton, 1987). Rook and Fisher
(1995) recentlydeveloped a nine-itemmeasureof traitbuying impulsivenessthat was significantlycorrelatedwith impulse buying behavior. In addition, they found that
consumers' normativeevaluation of the appropriatenessof
engaging in impulse buying in a particularsituationmoderates an individual's trait impulsiveness. Specifically, when
consumersbelieve thatimpulsepurchasingis socially acceptable, they act on theirimpulsivetendencies,but when it is socially unacceptablethese tendenciesmay be thwarted.
The literatureon compulsive shopping (Elliot, 1994),
self-gifts (Mick, DeMoss, & Faber, 1992), and impulse purchases (Dittmaret al., 1995) highlightsthe role of perceived
social image and the expression of self-identity in the purchase decision. Dittmaret al. (1995) hypothesizedthat impulse purchaseswere more likely to be items thatsymbolize
the preferredor ideal self and as such should be affected by
social categories such as gender. They argued that women
value their possessions for emotional and relationship-oriented reasons,whereasmen value theirpossessions for functional and instrumentalreasons. The results of the study
supportedtheirhypothesis:Men reportedmore personal(independent) identity reasons for their purchases whereas
women reportedmore social (relational)identityreasons.
An individual'simpulsive behaviortendencieshave also
been related to demographiccharacteristicssuch as a consumer's age. Based on a national sample of adults in the
UnitedStates,Wood (1998) foundan inverserelationshipbetween age andimpulsebuyingoverall.However,therelationship is non-monotonic- between the ages of 18 and 39
impulse buying increases slightly and thereafterdeclines.
This is consistentwith Bellengeret al. (1978) who foundthat
shoppersunder35 were more proneto impulsebuying comparedto those over 35 yearsold. Researchon traitimpulsiveness indicates that younger individuals score higher on
measuresof impulsivitycomparedto older people (Eysenck
et al., 1985; Helmerset al., 1995; Rawlings et al., 1995) and
demonstrateless self-controlthanadults(Logue & Chavarro,
1992). Because impulsivenessis linkedto emotionalarousal,
this findingconcerningthe relationshipbetween age and impulsiveness is consistent with studies of emotions and emo-

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

CULTUREAND IMPULSIVEBUYING BEHAVIOR

tional control. Research shows that older individuals

demonstrategreaterregulationof emotionalexpressionthan
do younger adults (Lawton, Kleban, Rajogopal, & Dean,
1992; McConathaet al., 1994; Siegel, 1985). These findings
suggest thatas consumersage, they learnto controltheirimpulsive buying tendencies.
Interestingly,the factorsthathave been linked to impulse
purchasingare also likely to be influenced by culture. The
theory of individualism and collectivism offers several insights into many of the variablesthathave been linked to impulsive buying behavior, including self-identity, normative
influences, the suppressionof emotion, andpostponementof
instantgratification(see Triandis,1995 for a review). In the
next section, we discuss this theoryand demonstratethatit is
well suited to the study of impulse buying.

INDIVIDUALISMAND COLLECTIVISM
Triandis(1995) defined collectivism as a social patternthat
consists of individualswho see themselves as an integralpart
of one or more collectives or in-groups,such as family and
co-workers.People who are more collectivist areoften motivated by normsand duties imposed by the in-group,give priority to the goals of the in-group,and try to emphasize their
connectednesswith the in-group.He defines individualismas
a social patternthat consists of individuals who see themselves as autonomousand independent.People who aremore
individualistare motivatedby their own preferences,needs,
and rights, give priorityto their personalgoals, and emphasize a rational analysis of their relationships with others
(Triandis,1994). These social patternsare expected to influence impulsive purchasingbehaviorthroughtheiraffect on a
person's self-identity, responsiveness to normative influences, and the need (or lack of need) to suppressinternalbeliefs in orderto act appropriately.
The tendencyto focus on grouppreferencesandgroupharmony in collectivist culturesleads to an ability to repressinternal (personal) attributesin certain settings. Accordingly,
people in collectivist cultures often shift their behaviordepending on the context or what is "right"for the situation.
Among collectivists a person is generally seen as more mature when s/he puts personalfeelings aside and acts in a socially appropriatemannerratherthanin a way consistentwith
personal attitudes and beliefs (Triandis, 1995). Consequently, it has been found that attitude-intention(Bagozzi,
Wonge, Abe, & Bergami, 2000; Lee, 2000) and attitude-behavior relationships(Kashima,Siegal, Tanaka,& Kashima,
1992) are weaker in collectivist than individualistcultures.
This patternis likely to carryover to the impulse trait-behavior relationship.
Collectivist culturesalso emphasizethe controland moderation of one's emotions more so than individualisticcultures (Potter, 1988; Russell & Yik, 1996; Tsai & Levenson,
1997). For instance, the maintenanceof harmonywithin the

165

group is dependent on members' ability to manage their

emotions. In short,cultureis likely to impactan individual's
emotional experiences by determiningthe appropriateexpressionof one's feelings (McConatha,1993). Cultureinfluences both "feeling rules,"how an individualinterpretsthe
environment,and "display rules," which emotions are expressed and how they are expressed (Ekman, 1972). For instance, people from Asian (collectivist) cultures have been
foundto controlnegativeemotions and only display positive
emotionsto acquaintances(Gudykunst,1993). Given thatimpulsiveness is related to sensation-seeking and emotional
arousal (Rook, 1987; Weinberg & Gottwald, 1982), it is
likely thatpeople in collectivist cultureslearnto controltheir
impulsive tendencies more than people from individualist
cultures.In fact, children in collectivist cultures are socialized to controltheirimpulses at an early age (Ho, 1994).
In individualistcultures,people often ignore the potential
negative consequences of their impulsive buying behavior
(see Rook, 1987), preferringto focus on the positive consequences of theiractions and on theirown feelings and goals.
This may not be true for people from collectivist cultures,
who aremore likely to focus on the potentialnegativeconsequences of their behaviorand the effect of their actions on
in-groupmembers (Triandis, 1995). The greaterlikelihood
thatpeople in collectivist cultureswill considerthe negative
consequences of their actions makes the suppressionof the
impulse trait-behaviorrelationshipmore probable.
These differencesbetween individualistsandcollectivists
arebest explainedby examiningthe tenantson which the culturalpatternsof individualismandcollectivism arebased.As
Kim,Triandis,Kagiycibasi,Choi,andYoon (1994) explained
Westernindividualistsocieties arebasedon thetenantof liberalism.Inthesesocieties individualsareencouragedto be rational and are given individualrights to define theirown goals
andchoose freely. Conversely,EastAsian collectivist societies are based on Confucianism, which promotes common
goals and social harmonyover individual interests.Within
each society these differences are reinforcedat the cultural
level throughsocial institutionssuch as schools, workplaces,
andfamilies, so thateven very ambitious(i.e., moreindividualist) people who grow up in China are likely to be betterat
controllingtheirimpulses andemotions thanvery family-focused(i.e., morecollectivist)people fromtheUnitedStates.In
collectivist societies individualsare encouragedto suppress
theirown hedonicdesiresin favorof groupinterestsandgoals.
Fromthis we conclude the following:
H la: The relationshipbetween traitbuying impulsiveness
and impulsive buying behavior will be stronger
among people from individualistcultures compared
to people from collectivist cultures.
Several researchershave emphasized the importanceof
examiningthe influence of cultureat the individuallevel as
well as at the nationallevel (e.g., Kim et al., 1994; Schwartz,

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

166

KACENANDLEE

1994; Singelis & Brown, 1995; Triandis, 1994, 1995). Accordingto Triandis(1994), "All of us carrybothindividualist
andcollectivist tendencies;the differenceis thatin some cultures the probability that individualist selves, attitudes,
norms, values, and behaviors will be sampled or used is
higherthanin others"(p. 42). Consequently,people fromcollectivistculturesshouldbe morelikely to rely on a moreinterdependent self-concept and people from individualist
culturesshould be more likely to rely on a more independent
self-concept in any given situation.Singelis (1995) defined
an interdependentself-concept as one emphasizing"(a) external, public features such as statuses, roles, and relationships, (b) belongingandfittingin, (c) occupyingone's proper
place and engaging in appropriateaction, and (d) being indirect in communicationand 'readingothers' minds,"'and an
independentself-concept as one emphasizing "(a) internal
abilities, thoughts, and feelings; (b) being unique and expressingthe self; (c) realizinginternalattributesandpromoting one's own goals; and (d) being directin communication"
(p. 581).
We expect that measuringself-concept at the individual
level across cultures should produce parallel although not
identicalresultsto the cultural(i.e., regional) level analysis.
Although a person's self-concept reveals the partsof culture
thathave been internalizedby thatindividual,it does not fully
explaindifferencesthatmay be due to the influenceof social
institutions,which emphasizethe suppressionof hedonicdesires in favor of group interestsand goals. Thus, at the individual level across societies, we expect to find a similar
patternof trait-behaviorrelationships,althoughthe differential effect of cultureshould be somewhat weakerthan at the
regional-nationallevel.
Hlb: The relationshipbetween traitbuying impulsiveness
andimpulsivebuyingbehaviorwill be strongerfor individualsclassified as havinga moreindependent(individualist)self-concept as comparedto those classified as having a more interdependent(collectivist)
self-concept.
In addition,becausecontrolandmoderationof one's emotions is emphasized more strongly in collectivist cultures,
consumersfromthese culturesaremorelikely to suppressthe
emotional component of their impulse buying experience
thanthose from individualistcultures.
H2: The emotional factors of pleasure and arousal that
characterizeimpulsive buying behaviorwill be more
positively related to impulsive buying behavior
among individualiststhanamong collectivists.
However, pleasureand arousalmay be universalcomponents of spontaneousbuying behavior, and ones sharedby
people in both individualistand collectivist cultures. If so,
thenfeelings of pleasureandarousalwill be positivelyrelated

to impulse buying behavior among both individualistsand

collectivists. Given the lack of researchinto impulse buying
in non-Westernsocieties, one of the objectives of our researchis to determinein whatways impulsivebuyingbehavior differs across culturalcontexts.
Finally, the moderatinginfluenceof age is expectedto affect the impulsivebuyingbehaviorof people fromcollectivist
culturesearlierthanthose from individualistcultures.Given
thatadultsin individualistcultureshave demonstrateda slight
increase in impulsive buying into their late 30's (Wood,
1998), we do not anticipatea decline in impulsivebuyingbehavior for college-aged individualists.Conversely, because
people in collectivist cultureslearnat an earlierage to control
theiremotionsandbehavior,we expect age to negativelyimpact impulse buying once collectivists reachcollege age.
H3: Age will negatively impact impulse buying to a
greaterextent for collectivists comparedto individualists, in theirearly adultyears.

METHOD
Overview
Two studieswere conductedto measurethe influenceof cultureon consumers'impulsivebuying behavior.The preliminarystudyconcentrateson a parsimoniousexplanationof impulsive buying behavior: The basic hypothesis is that
consumerswitha personalitytraitof impulsivenesswill make
more frequentimpulsivepurchases,but thatthis relationship
will be moremoderatein collectivistcultures.The mainstudy
examines this effect but also controlsfor the effect of affective statesandage variableson impulsivebuying behaviorto
betterunderstandthiscomplex buyingbehaviorwithindifferent culturalcontexts.
In thesestudies,surveyswereadministeredto studentsand
non-studentsin highly individualistand highly collectivist
countries purposefully selected from their positioning on
Hofstede's (1991) ranking of individualismto include the
UnitedStates(individualismscore= 91) andAustralia(90) as
highly individualisticcountriesandSingapore(20), Malaysia
(26), and Hong Kong (25) as highly collectivist countries.In
each study,culturaldifferenceswere comparedat two levels
of analysis:culturalregion(WesternIndividualistvs. Eastern
Collectivist) and individual level (independentvs. interdependent self-concept). Using both levels helps to address
some of the more common criticisms associated with
cross-culturalresearch.Althoughusing culturalregion as an
indicatorof individualismandcollectivism offers the advantage of capturingthe morecomplex natureof the construct,it
also includes the disadvantageof adding between-country
variance to the often problematicwithin-countryvariance
found in cross-culturalresearch. On the other hand, using
people's self-conceptas an indicatorof theirlevel of individ-

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

CULTUREAND IMPULSIVEBUYING BEHAVIOR

ualism and collectivism measures the within-countryvariances, recognizingthateach personinternalizesnationaland

institutionalinfluencesto a greateror lesser extent,butit fails
to pick up the morecomplex natureof the construct.Using an
individuallevel measureof culturein additionto a regionalor
national level measure adds confidence that the results are
due to the constructof cultureregardlessof its measurement
(see Maheswaran& Shavitt,2000; Schwartz, 1994; Singelis
& Brown, 1995; Triandis, 1994, 1995).

PRELIMINARYSTUDY
Participants and Measures
A survey was administeredto 706 studentsand non-students
in fourcountries,two individualisticcountries(Australiaand
United States) and two collectivist countries(Singaporeand
Malaysia). As partof a largerstudy, participantswere asked
to complete a questionnaireconcerning a recent impulsive
purchase,defined in this questionnaireas, "one in which you
experience a sudden unexpectedurge to buy somethingthat
you cannotresist. Impulsepurchasesoccur while a personis
in the store and involve rapiddecision making."The survey
included questions on impulsive purchasing behavior, respondents' independent and interdependent self-concept
(Singelis, 1994), traitbuying impulsiveness(Rook & Fisher,
1995), anddemographicitems includingthe countryin which
respondentcurrentlylives, andwhetherthiscountryis theone
she or he has lived in most of his or her life.
A single measureof impulsive buying behaviorwas used
for this preliminarystudy. The item, "How often do you buy
things on impulse?"was measuredon a 4-point scale from 1
(almost every day) to 4 (almost never). The summarystatistics for this item are reportedin Table 1.
Respondentswere classified into culturalgroupingsbased
both on theirculturalregion andtheirindividualself-concept
score as follows. At the cultural region level, respondents
were classified into two groupsbasedon theircountryof residence: individualist(Australiaand United States) or collectivist (Malaysiaand Singapore).Those respondentswho had
not lived in theircountryfor most of theirlife were excluded
from the sample at this level of analysis. The resultingsamples comprised245 respondentsfromthe individualistregion
(n = 131 from Australiaand n = 114 from the United States)
and 344 respondentsfrom the collectivist region (n = 160
from Malaysia and n = 184 from Singapore). The cultural
classification proceduredeveloped by Triandis(1995) was
used to group participantsat the individuallevel of culture.
Respondentsindicatedtheirlevel of agreementwith 12 independence and 12 interdependenceself-concept statements
found in Singelis (1994) on 9-point strongly disagree-stronglyagree scales. These two self-concept scales
have been used frequently in cross-culturalresearch with
consistentresultsandin this studyeach scale receiveda satis-

167

factorylevel of reliabilityin each sample, using Nunnally's

(1978) criteriaof a > .70. The 12 interdependentitems and 12
independentitems wereeach averaged,andrespondentswere
trichotomizedwitha scoreof 1,2, or 3 on each. The independent score was reversedand addedto the interdependentscore
to producea measureof independence-interdependence
on a
scale of 2 to 6. Those respondentswho scored a 2 or 3 were
classified as independent(n = 217), those who scoreda 5 or 6
were classified as interdependent(n = 207) and those who
scored 4 were classified as neither and removed from the
analysis at the individualself-concept level.
The personalitytraitof buyingimpulsivenesswas initially
measured using Rook and Fisher's (1995) nine trait-buying-impulsiveness-scaleitems, measuredon 5-point strongly
disagree-stronglyagree scales. This scale achieved satisfactory levels of reliability(rangingfrom .79 to .92) in each of
the countrysamples (Nunnally, 1978). However, it is possible thatthe nine items measureslightlydifferentconstructsin
differentculturesand if so, the impulse model may fit poorly
(see Marsh & Byrne, 1993). All nine of the traitimpulsiveness items were factor analyzed across culturalregions and
reducedto a sub-scale of four items1that were most consistent across differentcultures(for details see Appendix A).
Of course, we predictthatpeople with highertraitbuying
impulsivitywill make more frequentimpulse purchases,but
we also hypothesizethat for consumersin individualistculturesthe trait-behaviorrelationshipwill be strongerthanfor
consumersin collectivist cultures.To test this, a comparison
of correlationsand their variationacross culturalgroupings
was conducted.

Results
As seen in the top portionof Table 2, the correlationbetween
traitandbehavioralimpulsivenessequals0.64 for individualist culturesand 0.40 for collectivists. Similarly,the correlation between traitand behavioralimpulsivenessequals 0.59
for independentand 0.46 for interdependentself-conceptsof
culture.All of these correlationsare significantlypositive at
the .001 level as expected.
As hypothesized,the buyingimpulsivenesstraitwas more
stronglyassociatedwith impulse buying behaviorfor the individualistthan for the collectivist groups. Fisher's z-transformations revealed that the correlations differed
significantlyin the expecteddirectionat both the culturalregion (z = 3.87, p < .001) and the individualself-concept(z =
1.93,p < .05) levels of analysis(see Table2). As expected,the
effect was more distinctat the regionallevel thanat the individual level of measurementof culture.Notably, the results

'Justdo it' describesthe way I buy things;(2) 'I see it, I buy it' describes
me;(3) 'Buy now, thinkaboutit later'describesme;and(4) 1 oftenbuy things
withoutthinking.

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

168

KACENANDLEE

showed strongersupportfor our hypothesisat both levels of

analysis when we comparedthe partialcorrelationsaftercontrollingfor variances(culturalregion,z = 6.27, p < .001; individual difference,z = 4.40, p < .001).

Discussion
The results from our preliminaryinvestigation indicate a
strongerrelationshipbetweentraitbuyingimpulsivenessand
impulsive buying behavior for individualists compared to
collectivists, which suggests thatcollectivists are less driven
thanindividualiststo act on their traitbuying impulsiveness

by making an impulse purchase. Our finding is consistent

with otherresearchthat indicatesthe attitude-behaviorrelationshipis weakerin collectivist thanin individualistcultures
(Kashimaet al., 1992). This evidence suggests that culture
does moderatethe impulse trait-behaviorrelationship.
Although this finding highlights a significant difference
between consumersin Westernversus Easterncultures,it is
importantto examine othervariablesthat may also differentially affect the impulsive buying behaviorof individualists
comparedto collectivists. The impulsebuyingliteraturesuggests thatconsumers'emotionalstatesandtheirage influence
theirimpulsivebuyingbehavior,yet the theoryof individualism and collectivism would predicta less importantrole for

TABLE1
Descriptionof Measures and SummaryStatistics for Scales
IndividualistRegion (n = 230)

Collectivist Region (n = 318)

SD

SD

Impulsivebuying behavior(4-point)*
Traitbuying impulsive sub-scale (5-point)**

2.03

0.72

2.09

0.52

2.53

0.97

2.59

0.81

"Justdo it" describesthe way I buy things

2.90

1.17

3.03

1.06

"I see it, I buy it" describes me

2.42

1.13

2.57

1.02

"Buy now, thinkabout it later"describes me

2.34

1.12

2.37

1.03

I often buy things withoutthinking

2.46

1.11

2.36

0.94

Description of Measures
PreliminaryStudy

Caucasian (n = 167)

Asian (n = 233)

Main Study
Impulsivebuying behavior(numberof times in last
month)
Traitbuying impulsivenesssub-scale (7-point)

4.68

4.51

3.29

2.89

When I go shopping, I buy things that I had not intendedto purchase***

4.23

1.57

4.12

1.30

4.18

1.87

4.17

1.59

I am a person who makes unplannedpurchases***

When I see somethingthat really interestsme, I buy it
withoutconsideringthe consequences****

4.38

1.92

3.98

1.68

3.66

1.98

3.98

1.82

I avoid buying things that are not on my shoppinglist

(r)****
Arousalscale (8-point semanticdifferentialscales)

4.70

1.63

4.35

1.65
0.96

5.49

1.04

5.07

Stimulated-relaxed(r)

5.63

1.64

4.78

1.85

Calm-excited

5.41

1.76

5.22

1.67

Frenzied-sluggish(r)
Unaroused-aroused

5.01

1.16

4.76

1.10

5.91

1.23

5.55

1.12

6.26

0.91

5.75

0.94
1.32

Pleasurescale (8-point semanticdifferentialscales)

Happy-unhappy(r)

6.49

1.09

5.98

Annoyed-pleased
Unsatisfied-satisfied

6.38

1.22

5.86

1.28

6.39

1.16

5.91

1.21

Contented-melancholic(r)

5.81

1.33

5.24

1.20

Note. Itemswith an (r) arenegativelywordedandarescoredinversely.Itemswithone asteriskaremeasuredas 1) almostevery day, 2) often, 3) sometimes,4)

never. Itemswith two asterisksare measuredas 1) stronglydisagree,2) disagree,3) neither,4) agree,5) stronglyagree. Itemswith threeasterisksaremeasuredas
1) very rarely,4) sometimes, 7) very often. Items with four asterisksare measuredas 1) stronglydisagree,4) neither,7) stronglyagree.

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

CULTUREAND IMPULSIVEBUYING BEHAVIOR

169

TABLE
2
Correlations of Trait Buying Impulsiveness Subscale With Impulsive Buying Behavior and Entire Trait Buying Impulsiveness Scale
Correlation With
Impulse Buying
Behavior

.64***

.75

Collectivist

.40***

.42

Independent

.59***

.68

.46***

.49

PreliminaryStudy
Individualist

Interdependent
Main Study
Individualist

.49**

.54

Collectivist

.42**

.45

Caucasian

.51**

.57

Asia

.36**

.38

Independent

.53**

.59

Interdependent

.34**

.36

Zfor Partial
Correlations

Correlation With
Entire Trait Scale

Sample Size

3.87***

6.27***

.91***

237

.87***

336

1.93**

4.40***

.90***

212

.88***

201

Z=

- Z2

.83

1.35*

1.94**
1.88**

1.67**
2.08**

.99***

194

.97***

164

.99***

173

.98***

241

.99***

127

.97***

157

*p <.10. **p < .05. ***p < .001 or smaller, one-tailed.

these variablesin the behaviorof collectivists due to cultural

differences.We examine the effect of consumers'age andaffective feelings in the main study, describednext.
MAINSTUDY
Due to limitationsinherentin using only one studyto uncover
the buying impulsivenesstrait-behaviorrelationship,albeit a
large survey in four differentcountries,we conducteda conceptualreplicationand extension. While this main studywas
carriedout in a mannersimilar to the preliminarystudy, we
variedthe majormeasuresusinga differentbuyingimpulsiveness traitscale anda differentmeasureof impulsivebuyingbehaviorto attemptto generalizeourfindings,as recommended
by Lehmann (2001). Using a different but equally reliable
measureof traitbuying impulsiveness provides more assurance thatour resultsreveal culturalinfluences and arenot an
artifactof the measuresused.In addition,anattemptwas made
to bettermatchthe samplesby using only Universitystudents.
We also extended our analysis to include the effects of consumers' age and affective feelings, as these variableshave
been characterizedas distinctcomponentsof impulsivebuying behaviorin previous (individualist)studies.

ist thanindividualistdue to the mixed culturalbase (Kashima

et al., 1995). Studentswere chosen in all five countriesin an
attempt to achieve demographicallymatched samples and
thus minimize the variance in terms of age, education, and
other potentialconfounds. The mean age across all respondents was 24 years (range 19 to 45 years).
As part of a larger study, participantswere asked to respondto questionsconcerninga recentspontaneousclothing
purchase,and then to answera more generalset of questions
relevantto this study regardingtheirtraitand behavioralimpulsiveness.2The introductionto the questionnairewas similar to that of the preliminarystudy, except that the words
"impulsive purchase" were replaced with "spontaneous
clothing purchase."The word "impulsive" may be more
value-ladenin some culturesso the more neutral"spontaneous" was used, consistent with the historical
operationalizationof impulsive buying as "a purchasecharacterizedby spontaneity"(O'Guinn& Faber, 1989, p. 150).
In this study,impulsivebuyingbehaviorwas measuredas,
"Howoften have you boughtthingsspontaneouslyin the last
month?"Trait-buyingimpulsivenesswas measuredwith the
five-item scale fromWeun et al. (1998), ratherthanthe Rook
and Fisher (1995) scale. Respondents' affective feelings
when engaging in impulsive buying behaviorwere assessed

Participants and Measures

A survey was administeredto 481 studentsat large universities in five countries: Australia (n = 61), the Midwestern
United States (n = 168), Singapore (n = 72), Malaysia (n =
53), andHong Kong (n = 62). Additionaldatagatheredin Hawaii (n = 65) was separatedfrom the Midwestern United
Statessample,as Hawaiihas been foundto be morecollectiv-

2Respondentswere firstqueriedabouta recentspontaneousclothing purchase beforebeing asked moregeneralquestionsabouttheirimpulsivebuying behavior in orderto increase the saliency of the topic (see Sudman &
Bradbur 1982). The productcategory of clothing was chosen for several
reasons:(1) clothing was one of the most likely consumergoods to be purchasedon impulseaccordingto the Dittmaret al. study(1995), (2) clothingis
a productpurchasedand wornby consumersof both sexes and all ages, and
(3) it is a productthatis readily availablein each of the countriesstudied.

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

170

KACENANDLEE

by eight semantic-differentialscales measuringpleasureand

arousalbased on Russell and Pratt(1980). According to the
Russell andPrattmodel, two majordimensions,pleasureand
arousal can represent all affective states.3 These pleasure-arousalscales have been used by others to assess consumers' affective responses (see Dawson, Bloch, &
Ridgway, 1990; Donovan & Rossiter, 1982; Donovan et al.,
1994; Havlena& Holbrook,1986; Holbrook& Batra,1987).
As in the preliminary study, independent-interdependent
self-concept was measuredusing Singelis' (1994) scales. In
addition,several demographicitems were included,such as
the countryin which respondentcurrentlylives, whetherthis
countryis the one she or he has lived in most of his or herlife,
ethnicity, age, and sex. See Table 1 for scale items and summarystatistics.
Respondents were classified into three cultural groupings: cultural region, individual difference (independent-interdependent self-concept), and self-reported
ethnicity. As in the preliminarystudy, at the culturalregion
level respondentswere classified into two groups based on
their country of residence: individualist (Australia and
United States - excluding Hawaii) and collectivist (Singapore, Malaysia, and Hong Kong). The resultingsample size
was 197 respondentsfrom the individualistregion including
Australia (n = 42) and the United States (n = 155), and 174
from the collectivist region including Singapore (n = 66),
Malaysia (n = 50) and Hong Kong (n = 58). The reliabilities
for Singelis' (1994) independence-interdependencescales
were satisfactory in each country sample and the respondents were classified into two individual difference groups
following the same procedureas in the preliminarystudy.At
the ethnicity level, respondents were classified into two
groups based on their self-reportedethnicity (coded as Caucasian and Asian). Table 3 details the percentageof respondents from each country allocated to the cultural region,
individual difference, and ethnicity groups.
Respondentsfromthe morecollectivist countries(Singapore, Malaysia, and Hong Kong) were almost all of Asian
ethnicity, and more often classified as interdependentthan
as independent.As expected, the Hawaiian sample shared
the characteristics of collectivist countries. Respondents
from the more individualist countries (Australia and the
United States) showed more variationin both ethnicity and
at the individual self-concept level. This is due to the larger
internationalstudentpopulationat the universities sampled
in Australiaand the United States. We expect this to impact
our culturalregion level analysis, but not the ethnicity or in-

The Russell-Prattmodel representsa modification of Mehrabianand

Russell's (1974) Pleasure-Arousal-Dominanceparadigm.The thirddimension, dominance,has been found to be less robustand not always replicable
across studies (Russell, 1978; Watson & Tellegen, 1985). Russell deleted
dominancein his laterworkarguingthatit requiresa cognitive interpretation
and is not applicableto purelyaffective responses.

dividual difference comparisons. For this reason

reliabilities and equivalencies were assessed using the ethnicity groupings.
The five-itemtraitbuyingimpulsivenessscale fromWeun
et al. (1998) achieved satisfactorylevels of reliability(ranging from .68 to .88) in each country sample, although the
reliabilitieswere noticeably lower for the collectivist countries thanfor the individualistcountries.Next, all five items
were submittedto a maximumlikelihood exploratoryfactor
analysis for each ethnicitygroup.Following the same procedureas in the preliminarystudyproduceda 4-item scale4accountingfor 64%of the variancein the Caucasiansampleand
52%of the variancein theAsian sample(see Table 1 for items
includedin this subscale and Appendix A for details of the
factoranalysis).
Using the same method, the four pleasure items and the
four arousal items were submittedto maximum likelihood
exploratoryfactoranalyses for each culturalregion (see Table 1 for the items and summary statistics). These constructs were found to be consistent across cultures, as
demonstratedin Appendix B.

Results
Two traditionalmethods, Fisher's z-test and moderatedregression analysis, were used to assess moderationfor this
study.The first was done to maintaincomparabilitywith our
preliminarystudy; we tested the association between trait
buying impulsivenessand impulsive buying behaviorat the
culturalregion (individualist-collectivist),ethnicity(Caucasian-Asian), and individual difference (independent-interdependentself-concept)levels to determinewhetherthe buying impulsivenesstraitexplains as much of the variationin
impulsive buying behaviorfor one group as it does for the
other.However,it is possible thatthe correlationof impulsive
buying behavior with just a personalitytrait fails to distinguish the impact of trait from situational variables (like
mood) or demographicvariables (like age) that also vary
within the sample. To deal with this a moderatedregression
analysiswas also conductedto determinethe differentialcontributionof traitbuyingimpulsiveness,arousal,pleasure,and
age across cultures.
First, the Fisher's z-test results of the cultural differences in trait-behaviorimpulsiveness correlationsare consistent with the first study (see Table 2). A Fisher's
z-transformation revealed that the correlations differed
significantly at the ethnicity (z = 1.94, p < .05) and individual difference (z = 1.88, p < .05) levels, but not at the cul-

(1) When I go shopping,I buy thingsthatI hadnot intendedto purchase;

(2) I am a personwho makesunplannedpurchases;(3) WhenI see something
that really interestsme, I buy it withoutconsideringthe consequences;and
(4) I avoid buying things thatare not on my shoppinglist.

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

ANDIMPULSIVE
CULTURE
BUYINGBEHAVIOR 171

TABLE
3
andSelf-Concept
of Respondents'Countryby Ethnicity
Cross-Tabulation
Ethnicity

Self-Concept

Caucasian(%)

Asian(%)

SampleSize

Australia
UnitedStates
Hawaii

78
92
11

Singapore
Malaysia
HongKong
Collectivist
Individualist

2
1
94

22
8
89
100
100
98
99
6

55
137
57
71
51
56
165
171

tural region level, most likely due to the larger percentage

of internationalstudents in our samples from Australiaand
the United States (although directional supportfor our hypothesis was found). For the more individualist groups, the
buying impulsiveness trait was more strongly associated
with how often a consumer bought something on impulse
than it was for the corresponding collectivist groups, supporting our hypothesis. After controlling for the variances,
these results increased in significance and the cultural region comparison reached significance at the p < .10 level
(see Table 2).
Second, a moderatedregressionanalysiswas conductedto
assess the influence of all the relevantindependentvariables
on consumers'reportedimpulsivebuyingbehavior.As previously stated,we expected the regionallevel variable,ethnicity, to capturethe moderatingeffect of cultureon impulsive
buying behavior, but we also expected some within culture
variation.To capturethis, we includedrespondents'level of
independentself-concept in the regression model. Respondents' level of independence,a continuousvariable,was calculatedby averagingthe scores from the independencescale
items and dividing that score by a sum of the averageof respondents'scores on the independencescale items and interdependence scale items. Values closer to I indicate high
independencewhile scores closer to zero indicate low independence.The influence of affective feelings of pleasureand
arousal,andage on impulsebuyingwas also assessedthrough
regression analysis. The following multivariateregression
equationwas used to test the moderatingeffect of cultureon
impulsive buying behavior.Ethnicitygrouping(CULTURE,
with I = Caucasian,0 = Asian) was treatedas a moderatorof
the effects of traitbuying impulsiveness(TRAITIMP),independence (INDEP), age (AGE), pleasure (PLEAS), and
arousal (AROUS) on impulsive buying behavior
(IMPBEHAVIOR):
IMPBEHAVIOR= a + b, CULTURE + b2TRAITIMP+ b.
INDEP + b4AGE (1)

(%) Interdependent
(%)
Independent
44
62
29
32
26
47
34
60

56
38
71
68
74
53
66
71

SampleSize
36
104
45
38
31
34
96
124

x CULTURE)
+ b7(TRAITIMP
+ b8
+ b5PLEAS+b6AROUS
(INDEP x CULTURE)
+ b9(AGE x CULTURE)+ bo (PLEAS x CULTURE)+ b,
(AROUS x CULTURE)+ error.

Table 4 summarizesthe empirical results. As expected

fromthe preliminarystudy,in eitherculturea personwho has
moretraitbuyingimpulsivenesshas made morespontaneous
purchasesin the previousmonth.The coefficients for the interaction terms, b7-bll,indicating the moderatingeffect of
Caucasian ethnicity on impulsive buying behavior can be
found in Table 4 underthe column heading "Differencebetween Caucasiansand Asians."
We predicteda positive parameterfor the TRAITIMPx
CULTUREterm.The resultssupportourhypothesiswiththis
parameterestimatebeing positive and significant(b7= 0.88, t
= 3.54, p = .001) indicatingthatthe Caucasian-Asianculture
variabledoes moderatethe influence of trait buying impulsiveness. Thatis, while the coefficient for traitbuyingimpulsiveness (TRAITIMP)is positive and significant for both
Caucasians (b2 + b7 = 1.51, t = 8.36, p = .001) and Asians (b2
=0.63, t = 3.61, p = .001) the influence of trait buying impul-

siveness on impulsive buying behavior is strongerfor the

Caucasiansample.
The independence-interdependencecultural variableallows us to check for an additionalself-concept effect within
ethnicity group. The independence (INDEP) parameteris
positive and significant for Caucasians(b3 + bs = 8.17, t =
2.11, p = .04) but not significantfor Asians. Among Asians a
more independentsense of self does not contributeto more
impulsive buying behaviorbut it does positively impactimpulse buyingamongCaucasians.Inotherwords,Asians regulate their impulsive buying behavior regardless of their
individuallevel of independentself-concept. This finding is
consistentwithpreviousresearchthathas foundthatwhile individualsin Asian culturesmaintainbothan independentand
an interdependentsense of self (Markus& Kitayama,1991),

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

172

KACENANDLEE

4
TABLE
Moderated
RegressionCoefficients
BuyingBehavior:Unstandardized
RegressionAnalysisforImpulsive

Variable

Caucasian

t-Value

t-Value

Asian
.63***

1.51***

8.36

Independence

8.17**

2.11

3.42

Age
Pleasure

-.02

-.49

-.09**

-.29

-.91

.12

.06

.24

-3.89

-1.35

TraitImpulsiveness

Arousal
Intercept
Adj R2

-1.98

.88***

3.61

4.75

.78

.52**

.25

Difference Between
Caucasians and
Asians

t- Value
3.54
.81

-2.15

.07

.95

.49

-.41

-1.03

2.35

-.46

-1.35

-.69

-1.91

-.47

.25

Note. To get the t-statisticsfor the two cultures,two equivalentversionsof equation(1) were estimated,the first using a dummyvariablecoding of "1"for
Asianand"0"forCaucasian(see the "Caucasian"column)anda secondreversingthecoding,using"1"forCaucasianand"0"forAsian(see the "Asian"column).
N =414.
*p < .10. **p < .05. ***p < .01 or smaller.

they areable to suppresstheirindependentself in certainsituations, puttingaside theirown preferencesin orderto act appropriately(Triandis,1995).
The moderatinginfluence of age was expected to negatively affect collectivists' impulsive buying behavior at a
youngerage thanindividualists.As predicted,the age coefficient is negative and significant for the Asian sample (b4 =
-.09, t = -2.15, p = .03), suggesting thatimpulse buying decreases with age among college-aged Asian students. Although directionally consistent, the age coefficient for
Caucasiansis nonsignificant.This is consistentwithprevious
research,which found an increasein impulse buying among
United Statesconsumersin their20s, with declines occurring
afterthe mid-30s (Bellengeret al., 1978;Wood, 1998).Ninety
percentof the Caucasiansample was under30 years old.
The contributionof affective feelings to impulsivebuying
behavior was surprising. While the arousal coefficient
(AROUS) is nonsignificantfor the Caucasiansample, it is
positive andsignificantfor Asian sample(b6= .52, t = 2.35, p
= .02), suggesting thatAsians' impulsive buying behavioris
responsiveto feelings of excitementor arousal,while Caucasians' behavioris not. No significant effect of pleasurewas
producedfor either group. In addition,the interactionterms
between pleasure and culture and arousal and culturewere
not significant.This is discussed furtherin the next section.
Discussion
Theresultsof the mainstudyprovidefurtherevidencethatculture affects the relationshipbetween trait buying impulsiveness and impulsivebuying behaviorat the culturalgrouping
and individual difference levels. Comparedto Caucasians,
Asiansengagein less impulsebuyingbehaviordueto traitbuying impulsiveness.In addition,independence-an individual
level measureof culturalinfluence-was not shownto impact
Asians' impulsivebuyingbehavior,althoughit did contribute

to impulsivebuying among Caucasians.For Caucasians,the

moreindependenttheirself-concept,the moreimpulsivebuying behaviorthey are likely to engage in. This findingreveals
the powerfulandconsistentinfluenceof cultureat boththeethnicity level and the individuallevel. Furthermore,this difference occursdespitesimilarlevels of traitbuyingimpulsiveness
amongourdifferentculturalgroups.Althoughthereis no significantdifferencein traitbuyingimpulsivenessbetweenCaucasiansandAsiansthereis a differencein theirimpulsivebuying behavior.5
In addition,at the college-age level (range 18-45 years,
mean24 yearsin both samples),getting olderreducesimpulsive buyingamongourAsian samplebutdoes not have a significant effect on our Caucasiansample. This is consistent
bothwith ourhypothesisfor the Asian sampleandwithprevious researchon Caucasians(Bellenger et al., 1978; Wood,
1998). A separateregressioninvestigating the influence of
age on traitbuying impulsivenessindicatedthattraitbuying
impulsivenessdecreaseswith age amongAsians butnot Caucasians. Ourresultsconfirmthatthe moderationof behavior
occurs earlierfor Asians comparedto Caucasians.
Unexpectedly, affective feelings played a greaterrole in
the impulsivebuyingbehaviorof Asians comparedto Caucasians.Foroursampleof Asians, heightenedarousalpromoted
impulsive buying. One possibility is thattraitbuying impulsiveness capturesthe contributinginfluenceof affective feelings of pleasureandarousalon impulsivebuyingbehaviorfor
Caucasians.It should be noted that the reliabilities for the
pleasureand arousalscales were not high for either Caucasians or Asians. The affective feelings associatedwith impul5The differencein traitimpulsivenessfor Caucasians(see Table 1 where
M = 4.23, SD = 1.57) andAsians (M = 4.12, SD = 1.30), t = .78, is not significant at 5% with 331 degrees of freedom.Caucasiansengage in significantly
more impulse buying (M = 4.68, SD = 4.51) than Asians (M = 3.29, SD =
2.89), since t = 3.56, has ap < .001 for 271 observations.

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

CULTUREAND IMPULSIVEBUYING BEHAVIOR

sive buying behaviormay be difficult to assess retroactively

through a survey instrument.It is also possible that these
scales do not accuratelymeasurethe affective states Asians
experience while making impulse purchases.Clearly, additional researchis called for to help clarify these issues.
GENERAL DISCUSSION
Attempts to understandconsumer impulse buying behavior
based solely on a Western point-of-view is incomplete, at
best. The Western-individualistemphasison the self, individual needs anddesires,andhedonisticpleasureencouragesimpulsive buying behavior. However, Eastern-collectivistnotions of the self, which emphasize interdependence,
emotional control and moderation,and group needs and desires would seem to discourageimpulsebuyingbehavioras it
is practicedand describedin the West. Due considerationof
these differences is warranted.
As Rook (1987) correctlystated,buying impulsesarepresumedto be largely universalin nature,but local marketconditions, systems of exchange and variousculturalforces will
impact how consumersoperateon impulse. The findings of
the articles in the recent special issue of the Journal of Consumer Psychology confirmed that culturaldifferences are a
significantfactorandneed to be takeninto accountin ourtheories of consumer behavior. Our findings demonstratethat
culturedoes have an influence on impulse buying behavior.
Although previous researchershave explored the influence of mood and emotions (Donovan et al., 1994; Rook &
Gardner,1993), trait impulsiveness (Rook & Fisher, 1995;
Weun et al., 1998), norms (Rook & Fisher, 1995), and
self-identity(Dittmaret al., 1995) on consumerimpulsebuying behavior,none have looked at culturalinfluencessuch as
collectivist and individualisttendencies, or independent-interdependentself-concept. As shown in this article,cultural
factorsdo moderateconsumerimpulsive buying behavior.
Overall, Asian collectivist consumersengage in less impulsive buying than Caucasianindividualistconsumers,despite the highly developed shoppingculturein East Asia. In
addition,there is a weakercorrelationbetween self-reported
trait buying impulsiveness and the frequency of impulsive
buying behaviorfor collectivists comparedto individualists.
This finding supports and extends previous research that
found thatcollectivists are able to maintaininconsistentattitude-behaviorrelationships(Kashimaet al., 1992) andto put
theirown feelings aside in orderto act in an appropriatemanner (Triandis,1995). Althoughcollectivists possess the buying impulsiveness traitin equal measurewith individualists,
they suppressthis traitimpulseandactin a mannerthatis consistentwith culturalnorms,in this case, reducingtheirimpulsive buying behavior, which has been characterizedas a
highly individualistic,emotionally chargedbehavior.
In the United States,it is assumedthatimpulsepurchasing
is correlatedwith personalitytraits,such as variety seeking,
sensation seeking, and risk aversion. If the correlations

173

among these variables are weaker in certain countries or

regions, theremust be otherdriving factorsdifferentiallyaffecting the amount of impulse purchasingthat occurs. One
suchfactor,identifiedby RookandFisher(1995), is the social
acceptabilityof impulse purchasing.An interestingarea for
future research would be to investigate the interactionbetween cultureandthe appropriatenessof engaging in impulse
buying in differentsituations.Althoughthe ability to control
the trait-behaviorrelationshipappearsto differ by culture,it
is also likely thatthe appropriatenessof the behaviorwould
influencethe desireandthusthe extentof control.One limitation of this studywas thatthe appropriatenessof impulsebuying was not investigated,althoughthe countriesincludedare
all consideredto be "shoppingcultures."
Many other aspects of culture are also likely to interact
with impulsiveness, at least as it has been characterizedby
Westernresearch,including hedonism, risk avoidance, perceived consequences,and the influence of others.Futureresearch of a qualitative nature is needed to uncover the
antecedents and consequences of buying impulsiveness
across cultures.
It has been suggested(Beatty& Ferrell,1998) thatprofiles
of highly impulsive shoppers be identified, so that promotions can be targetedat these individuals.Previous research
using consumerimpulsivity as a lifestyle traithas identified
thatpeople varyin theirimpulsebuyingsusceptibility(Rook,
1987;Rook & Fisher, 1995;Rook & Gardner,1993).Workin
this vein can identify high-, medium-,and low-impulse consumers.Ourresearchsuggests thatthese profiles may in fact
be different in other cultures. Furtherresearch needs to be
conductedinto whatfactorsarereliableindicatorsof impulse
buyersin culturesoutside of the United States.
The marketingfactorsthatencourageimpulsepurchasing
also need renewedattention.It would be useful to investigate
in detail how various marketingfactors supportimpulsive
purchasing,and which ones exert the strongest influence
withindifferentculturalcontexts. Access to the WorldWide
Web may well be an importantfactor.The Internetmost certainlyhas changedthe access to impulsepurchasingopportunitiesforthose frommoreremoteareas.The global electronic
marketplaceis makingit increasinglyimportantto study the
processes that may differentially affect people from other
countries, regions or areas of the world. One future study
wouldbe to investigatethe interactionof situationalvariables
withindifferentculturalsettings and among consumerswith
differentlevels of the impulsivenesstrait.
Our researchuncovers anotherarea where scales developed in the United States are not valid for use in othercountries, highlighting the difficulty of cross-culturalresearch.
The results of the factor analyses indicate that there may in
fact be morethanone dimensionto the buying impulsiveness
trait.We found that the Weun et al. (1998) scale produced
betterresultsthanthe Rook andFisher(1995) scale, although
even this scale had to be modified to achieve equivalence
across cultures. Furtherresearch needs to be conducted to

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

174

KACEN AND LEE

provide evidence of the nomonological validity of the trait

impulsiveness subscales.
This research has the traditionallimitations associated
with self-report survey research. Using a single-item measure of impulse purchasingmay be problematicdue to error
in the measurementof this construct.Yet, due to the unique
nature of impulse buying researchers often rely on single-item self-reports to measure this behavior (e.g., Beatty
& Ferrell, 1998; Rook & Fisher, 1995). Our use of two different items in two different investigations of impulsive
buying behavior produced the same hypothesized results
providing more confidence that measurementerrordid not
produceerroneousor misleadingresults.One concernis that
the cultural differences in impulsive buying behavior that
we found were due to cultural biases in response style.
Membersof EasternAsian cultureshave been shown to use
the midpointin ratingscales more thanmembersof Western
Caucasiancultures (Chen, Lee, & Stevenson, 1995) so it is
possible that our Caucasian-individualistrespondentswere
more likely to use the extreme values of the rating scales.
However, Chen et al. (1995) found that even when response-style biases due to cultureare accountedfor, they do
not significantly change the outcome of between-group
comparisons. Furthermore,no difference was found between the self-reportedbuying impulsiveness traitlevels in
each culture.
The use of sub-scalescould be problematicin capturingall
of the aspects of the impulsiveness trait.While this may be
problematicin bothof ourstudies,thereis some evidencethat
these sub-scales are highly correlatedwith the originalscale
(r < .9). In addition,the sub-scaledid notproducedifferentresults from the same analysis using the originalscale, lending
confidence to our findings.
Researchersand practitionersneed to be awareof cultural
differenceswhen applyingUnited States-basedresearchfindings to marketingstrategiestargetingnon-UnitedStatesconsumers. Although we surveyed individuals from highly
societies whereample opportunitiesto enconsumer-oriented
gage in impulsepurchasingexist, thereareessentialunderlying
differencesbetweenconsumersin Westernindividualistsocieties andthosein Easterncollectivistcultures.Althoughthe publishedresearchon impulsivebuyingbehaviorhas been helpful
in unravelingtheimpulsebuyingphenomenonin individualistic
societies, this articleclearlyhighlightsthe impactthatculture
hason thiscomplexconsumerbehaviorandsuggestswe needto
takeintoaccountthe interactionof cultureandconsumersin order to betterunderstandimpulsivebuyingbehavior.
ACKNOWLEDGMENTS
We gratefullyacknowledgethe financialsupportof the Center for InternationalBusiness, Educationand Researchat the
University of Hawaii and the University of Hawaii RFDC.
We thankSharonShavitt,Jim Hess, and Rich Lutz for their
comments on earlierdrafts,the editor and the reviewersfor

their helpful comments, and MarkPattonfor his kind assistance with datacollection.
REFERENCES
Abrahams,Ben. (1997). It's all in the mind. Marketing,March27, 31-33.
Abratt,Russell, & Goodey, StephenDonald (1990). Unplannedbuying and
in-storestimuliin supermarkets.Managerialand Decision Economics,
11(2), 111-121.
Bagozzi, RichardP., Wong, Nancy, Abe, Shuzo, & Bergami, Massimo.
(2000). Culturaland situationalcontingencies and the theory of reasoned action:Applicationto fast food restaurantconsumption.Journal
of ConsumerPsychology, 9(2), 97-106.
Bayley, Geoff, & Nancarrow,Clive. (1998). Impulsivepurchasing:A qualitative explorationof the phenomenon.QualitativeMarketResearch,
1(2), 99-114.
Beatty, Sharon E., & Ferrell, M. Elizabeth. (1998). Impulse buying:
Modeling its precursors.Journal of Retailing, 74(2), 169-191.
Bellenger,DannyN., Robertson,Dan H., & Hirschman,ElizabethC. (1978).
Impulsebuyingvariesby product.Journalof AdvertisingResearch,18,
15-18.
Chen, Chuansheng,Lee, Shin-ying, & Stevenson, HaroldW. (1995). Responsestyle andcross-culturalcomparisonsof ratingscales amongEast
Asians and North American Students. Psychological Science, 6(3),
170-175.
Dawson, Scott, Bloch, Peter H., & Ridgway, Nancy (1990). Shoppingmotives, emotional states, and retail outcomes. Journal of Retailing, 66,
408-427.
Dittmar,Helga, Beattie, Jane, & Friese, Susanne. (1995). Gender identity
and materialsymbols: Objectsand decision considerationsin impulse
purchases.Journalof EconomicPsychology, 16, 491-511.
Donovan,RobertJ., & Rossiter,JohnR. (1982). Storeatmosphere:An environmentalpsychology approach.Journal of Retailing,58, 34-57.
Donovan, RobertJ., Rossiter,John R., Marcoolyn,Gilian, & Nesdale, Andrew. (1994). Store atmosphereand purchasingbehavior.Journal of
Retailing, 70(3), 283-294.
Ekman,Paul.(1972). Universalandculturaldifferencesin facial expression
of emotion.InJ. R. Cole (Ed.),Nebraskasymposiumon motivation(pp.
207-283). Lincoln:Universityof NebraskaPress.
Elliott,Richard.(1994). Addictiveconsumption:functionandfragmentation
in postmodernmity.
Journalof ConsumerPolicy, 17, 159-179.
Eysenck, Sybil B., & Eysenck, HansJ. (1978). Impulsivenessand venturesomeness: Their position in a dimensional system of personalitydescription.Psychological Reports,43, 1247-1255.
Eysenck,Sybil B. G., Pearson,PaulR., Easting,G., & Allsopp, J. F. (1985).
Age normsfor impulsiveness,venturesomenessandempathyin adults.
Personalityand IndividualDifferences,6(5), 613-619.
Gerbing,David W., Ahadi, StephanA., & Patton,Jim H. (1987). Towarda
conceptualizationof impulsivity:Componentsacrossthe behavioraland
self-reportdomains.MultivariateBehavioralResearch,22, 357-379.
Gudykunst,William B. (Ed.). (1993). Communicationin Japan and the
UnitedStates. Albany:State Universityof New York Press.
Havlena, William J., & Holbrook,MorrisB. (1986). The varieties of consumption experience: Comparingtwo typologies of emotion in consumerbehavior.Journal of ConsumerResearch, 13, 394-404.
Helmers,KarinF., Young, Simon N., & Pihl, Robert0. (1995). Assessment
of measuresof impulsivityin healthymale volunteers.Personalityand
IndividualDifferences, 19, 927-935.
Hilgard,ErnestR. (1962). Impulsiveversus realisticthinking:An examination of the distinction between primaryand secondary processes in
thought.Psychological Bulletin,59, 477-488.
Ho, David Y. F. (1994). Cognitive socialization in confucian heritagecultures. In Patricia M. Greenfield & Rodney R. Cocking (Eds.),
Cross-cultural roots of minority child development (pp. 285-313).
Hillsdale,NJ: LawrenceErlbaumAssociates, Inc.

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

CULTUREAND IMPULSIVEBUYING BEHAVIOR

Hoch, StephenJ., & Loewenstein,George F. (1991). Time-inconsistentpreferences and consumerself control.Journalof ConsumerResearch, 17,
492-507.
Hofstede, Geert. (1991). Culturesand organizations:Softwareof the mind.
London:McGraw-Hill.
Holbrook,MorrisB., & Rajeev, Batra.(1987). Assessing the role of emotions as mediators of consumer responses to advertising.Journal of
ConsumerResearch, 14, 404-420.
Kashima,Yoshihisa, Yamaguchi,Susumu, Kim, Uichol, Choi, Sang-Chin,
Gelfand,Michele J., & Yuki, Masaki.(1995). Culture,gender,and self:
A perspective from individualism-collectivismresearch. Journal of
Personalityand Social Psychology, 69(5), 925-937.
Kashima, Yoshihisa, Siegal, Michael, Tanaka, Kenichiro, & Kashima,
Emiko S. (1992). Do people believe behavioursareconsistentwith attitudes? Towardsa culturalpsychology of attributionprocesses. British
Journal of Social Psychology, 31, 111-124.
Kim, Uichol, Triandis,HarryC., Kagitcibasi,Cigdem, Choi, Sang-Chin,&
Yoon, Gene. (1994). Introduction.In U. Kim, H. C. Triandis, C.
Kagitcibasi,S. Choi, & G. Yoon (Eds.), Individualismand collectivism:
Theory,method,andapplications(p. 1-16). ThousandOaks,CA: Sage.
Lawton,M. Powell, Kleban,MortonH., Rajagopal,Doris, & Dean,Jennifer.
(1992). Dimensions of affective experience in three age groups. Psychology and Aging, 7(2), 171-184.
Lee, JulieAnne. (2000). AdaptingTriandis'smodel of subjectivecultureand
social behaviorrelations to consumerbehavior. Journalof Consumer
Psychology, 9(2), 117-126.
Lehmann, Donald R. (2001). Quantitativeempirical generalizationsand
progress toward knowledge: Pushing the meta-analysisenvelope. In
Abbie Griffin& JamesD. Hess (Eds.), Proceedingsof the 15th Paul D.
ConverseSymposium,(p. 2-21). Chicago: AmericanMarketingAssociation.
Logue, A. W., & Chavarro,Adolfo. (1992). Self-controland impulsiveness
in preschool children. The Psychological Record, 42, 189-204.
Logue, A. W., King, George R., Chavarro,Adolfo, & Volpe, Joseph S.
(1990). Matchingand maximizingin a self-controlparadigmusing human subjects. Learningand Motivation,21(3), 340-368.
Maheswaran,Durairaj,& Shavitt,Sharon.(2000). Issues andnew directions
in globalconsumerpsychology.Journalof ConsumerPsychology, 9(2),
59-66.
Markus,Hazel Rose, & Kitayama,Shinobu.(1991). Cultureandthe self: Implications for cognition, emotion, and motivation.Psychological Review, 98, 224-253.
Marsh,HerbertW., & Byme, BarbaraM. (1993). Confirmatoryfactoranalysis of multitrait-multimethodself-concept data: Between group and
within-group invariance constraints. Multivariate Behavioral Research, 28, 313-349.
McConatha,JasminT. (1993). Culturalvariationsin emotionalexpression.
In Magill's Surveyof Social Science: Psychology (p. 887-892). Pasadena, CA: Salem.
McConatha,Jasmin T., Lightner,Eileen, & Deaner, Stephanie L. (1994).
Culture, age, and gender as variables in the expression of emotions.
Journal of Social Behavior and Personality, 9, 481-488.
Mehrabian,Albert,& Russell,James.(1974). Anapproachto environmental
psychology. Cambridge,MA: MIT Press.
Mick, David Glen, DeMoss, Michelle, & Faber,RonaldJ. (1992). A projective studyof motivationsand meaningsof self-gifts: Implicationsfor retail management.Journal of Retailing,68, 122-144.
Mischel, Walter.(1961). Preferencefor delayedreinforcementandsocial responsibility.Journal of Abnormaland Social Psychology, 62, 1-7.
Mogelonsky, Marcia. (1998). Keep candy in the aisles. American Demographics, 20, 32.
Nunnally,Jum C. (1978). Psychometrictheory. New York:McGraw-Hill.
O'Guinn, Thomas C., & Faber, Ronald J. (1989). Compulsive buying: A
phenomenological exploration. Journal of Consumer Research, 16,
147-157.

175

Potter,Sulasmith H. (1988). The culturalconstructionof emotion in rural

chinese social life. Ethos, 16, 181-208.
Puri,Radhika.(1996). Measuringand modifying consumerimpulsiveness:
A cost-benefitaccessibility framework.Journalof ConsumerPsychology, 5(2), 87-113.
Rawlings,David, Boldero,Jenny,& Wiseman,Felicity. (1995). The interactionof age with impulsivenessandventuresomenessin the predictionof
adolescentsexualbehavior.Personalityand IndividualDifferences,19,
117-120.
Rook, Dennis W. (1987). The buying impulse. Journal of ConsumerResearch, 14, 189-199.
Rook, Dennis W., & Fisher,RobertJ. (1995). Normativeinfluences on impulsivebuyingbehavior.Journalof ConsumerResearch,22, 305-313.
Rook, Dennis W., & Gardner,Meryl.(1993). Inthe mood: Impulsebuying's
affective antecedents.Research in Consumer Behavior (Vol. 6, pp.
1-28). Greenwich,CT: JAI Press.
Rook, DennisW., & Hoch, StephenJ. (1985). Consumingimpulses. In Morris Holbrookand ElizabethHirschman(Eds.), Advances in Consumer
Research (Vol. 12, pp. 23-27). Provo, UT: Association for Consumer
Research.
Russell,JamesA. (1978). Evidenceof convergentvalidityon the dimensions
of affect. Journal of Personality and Social Psychology, 36,
1152-1168.
Russell, JamesA., & Pratt,Geraldine.(1980). A descriptionof the affective
quality attributedto environments.Journal of Personality and Social
Psychology, 38, 311-322.
Russell, JamesA. & Yik, SulamithA. (1996). Emotionamong the Chinese.
In Michael H. Bond (Ed.), The handbookof chinese psychology (pp.
166-188). Hong Kong, China:Oxford UniversityPress.
Schwartz,ShalomH. (1994), Are thereuniversalaspectsin the structureand
contentsof humanvalues?Journal of Social Issues, 50(4), 19-45.
Sfiligoj, Eric. (1996, June 30). Helping the little guy to merchandise.Periscope, 20.
Siegel, J. M. (1985). The measurementof anger as a multidimensionalconstruct.In MargaretA. ChesneyandRay H. Rosenman(Eds.),Angerand
hostility in cardiovascular and behavioral disorders (pp. 59-62).
Washington,DC: Hemisphere.
Singelis, TheodoreM. (1994). The measurementof independentandinterdependentself- construals.Personality and Social Psychology Bulletin,
20(5), 580-591.
Singelis, TheodoreM., & Brown, William J. (1995), Culture,self, and collectivist communication:Linking culture to individualbehavior,Human CommunicationResearch, 21(3), 354-389.
Smith, Don. (1996, February).The joy of candy. National PetroleumNews
Supplement,S2.
Sudman, Seymour, & Bradbur, Norman M. (1982). Asking questions: A
practical guide to questionnairedesign. San Francisco:Jossey-Bass.
Thompson,CraigJ., Locander,William B., & Pollio, HowardR. (1990). The
lived meaning of free choice: An existential-phenomenologicaldescriptionof everydayconsumerexperiences of contemporarymarried
women. Journal of ConsumerResearch, 17, 346-361.
Triandis, Harry C. (1994). Culture and social behavior. New York:
McGraw-Hill.
Triandis, HarryC. (1995). Individualismand collectivism. Boulder, CO:
Westview.
Tsai, JeanneL., & Levenson,RobertW. (1997). Culturalinfluenceson emotional responding:Chinese Americanand EuropeanAmericandating
couples duringinterpersonalconflict. Journal of Cross-CulturalPsychology, 28, 600-625.
Underhill,Paco. (1999). Whywe buy: The science of'shopping.New York:
Simon & Schuster.
Watson,David, & Tellegen, Auke. (1985). Towarda consensualstructureof
mood. Psychological Bulletin,98, 219-235.
Weinberg,Peter,& Gottwald,Wolfgang. (1982). Impulsiveconsumerbuying as a resultof emotions. Journal of Business Research, 10, 43-57.

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions

176

KACENANDLEE

Weun,Seungoog,Jones, MichaelA., & Beatty, SharonE. (1998). Thedevelopment and validationof the impulse buying tendencyscale. Psychological Reports,82, 1123-1133.
Wong, Nancy Y., & Ahuvia,AaronC. (1998). Personaltasteandfamilyface:
Luxuryconsumptionin confucian and western societies. Psychology
and Marketing,15(5), 423-441.
Wood, Michael. (1998). Socio-economic status, delay of gratification,
and impulse buying. Journal of Economic Psychology, 19, 295-320.

Accepted by Dawn Iacobucci.

APPENDIX A
Analysis of Trait Buying Impulsiveness
Scales
Inthepreliminarystudy,all nineRookandFisher's(1995) trait
impulsivenessitems were submittedto a maximumlikelihood
exploratoryfactor analysis for each of the culturalregions.
Using theeigenvaluegreaterthan1 andthepercentof variance
accountedfor as initial indicators,one factoremergedfor the
individualistregion accountingfor 53% of the variance,but
two factorsemergedfor the collectivistregionaccountingfor
26% and 18%of the variance,respectively.In addition,some
of the communalitieswere as low as 0.17. After successive
eliminationof the items with low communalitiesa set of four
potentially equivalent items were left. These items loaded
greaterthan0.6 on thefactorandheldcommunalitiesof greater
than0.4 in each culturalregion.In addition,the X2for the exploratoryfactor analysis supporteda one-factorsolution in
bothregions(individualist,X2= .413, df= 2, p = .82 andcollectivist,X2= 4.98, df= 2, p = .08). Next, the dataweresubmitted
to a maximum-likelihoodmulti-groupconfirmatoryfactor
analysisusingLISREL8 to simultaneouslyassess the equality
of factor structuresacross the two regions. The four factor
loadings and four errorvarianceswere found to be invariant
acrossthe groups(X2= 15.90,df= 12,p = .20; RMSEA= .03),
indicatingmetricequivalence.Please see Table 1 for the individualsub-scaleitemsandtheirsummarystatisticsandTable2
for the subscalecorrelationswith the full nine-itemscale.For
the mainstudy,all five Weun et al. (1998) trait-buyingimpul-

siveness itemsweresubmittedto the same analysis.TheX2for

each region supporteda one-factorsolution(Caucasian,X2=
1.04,df= 2, p = .60; Asian,X2= .46, df= 2, p = .80). Whenthe
data were submittedto the maximum-likelihoodmulti-group
confirmatoryfactoranalysisthe four factorloadingsand four
errorvarianceswerefoundto be invariantacrossthegroups(X2
= 15.65, df= 12,p = .21), indicatingmetricequivalence.
APPENDIX B
Analysis of Pleasure and Arousal Scales
The four pleasure items and the four arousal items were
submittedto maximum likelihood exploratoryfactor analyses for each culturalregion (see Table 1 for the items and
summarystatistics). The pleasure scale accounted for 47%
of the variance in the Caucasian sample and 45% in the
Asian sample.
In addition,the X2for each region supporteda one-factor
solution (Caucasian,X2= 2.27, df= 2, p = .32; Asian, %2=
5.50, df= 2, p = .06). Whenthese items were submittedto the
maximum-likelihoodmulti-groupconfirmatoryfactoranalysis the four factorloadingsand fourerrorvariancesproduced
a slightly higherchi-square(%2= 23.64, df= 12,p = .02), but
an adequatefit using otherindices thatare less influencedby
samplesize (e.g., RMSEA= .07). As an addedprecaution,the
invariantmodel was comparedwith other less constrained
models to assess metric equivalence across groups. The x2
differencebetweenthe invariantfactormodel andthreeother
models (allowing factor loadings and errorvariancesto differ) producedan insignificantdifference in each case. The
arousalscale also produceda one-factorsolution(Caucasian,
X2= 1.28, df= 2, p = .53; Asian, X2= 1.50,df= 2, p = .47), althoughit accountedfor less variancein each group (35% in
the Caucasiansample and 25% in the Asian sample). When
these items were submitted to the maximum-likelihood
multi-groupconfirmatoryfactoranalysisthe fourfactorloadings and four error variances were found to be invariant
across the groups (X2= 14.75, df= 12, p = .26), indicating
metricequivalence.

This content downloaded from 221.120.220.12 on Wed, 15 Oct 2014 16:01:45 PM

All use subject to JSTOR Terms and Conditions