Feeling close and doing well: The prevalenceand motivational effects of interpersonally

engaging emotions in Mexican and EuropeanAmerican cultural contexts

Krishna Savani1, Ayme Alvarez2, Batja Mesquita3, and Hazel Rose Markus2

1Management Division, Columbia Business School, New York, NY, USA2Department of Psychology, Stanford University, Stanford, CA, USA3Center for Social and Cultural Psychology, University of Leuven, Leuven, Belgium

T wo studies investigate whether interpersonally engaging emotions—those that bring the self closer to others

(e.g., affection, shame)—are central to the model of self and relationships prevalent in Mexican culturalcontexts. Study 1 demonstrated that compared to people in European American contexts, people in Mexicancontexts were more likely to report experiencing interpersonally engaging emotions and less likely to report

experiencing interpersonally disengaging emotions. Study 2 found that interpersonally engaging emotions had asubstantial influence on performance motivation in Mexican contexts—Mexican participants solved more wordsearch puzzles after recalling instances in which they experienced positive interpersonally engaging emotions, and

fewer after recalling negative interpersonally disengaging emotions; in contrast, there were no differences bycondition for European Americans. These findings significantly extend previous research by documenting theimplications of relational concerns (e.g., simpatia, personalismo) for emotion and motivation in Mexican contexts,and are the first to demonstrate the motivational effects of interpersonally engaging emotions.

Keywords: Culture; Emotion; Motivation; Mexico; Latin America.

D eux etudes visaient a examiner a quel point les emotions engageantes sur le plan interpersonnel – c’est-a-dire celles qui nous amenent a etre plus proches des autres (p. ex., l’affection, la honte) – sont centrales pour

les modeles de soi et des relations qui prevalent dans les contextes culturels mexicains. L’etude 1a indique que,

comparativement aux gens des contextes americains europeens, les gens des contextes mexicains ont davantagevecu des emotions engageantes sur le plan interpersonnel et ils ont moins vecu des emotions qui lesdesengageaient sur le plan interpersonnel. L’etude 2a trouve que les emotions engageantes sur le plan

interpersonnel constituaient une influence substantielle sur la motivation a la performance dans les contextesmexicains – les participants mexicains trouvaient plus de mots dans une tache de recherche de mots caches apress’etre rappele des circonstances dans lesquelles ils avaient vecu des emotions positives engageantes sur le planinterpersonnel, mais ils trouvaient moins de mots apres s’etre rappele des emotions negatives qui les

desengageaient sur le plan interpersonnel. En contrepartie, il n’y avait pas de difference dans ces conditions pourles Americains Europeens. Ces resultats representent un apport significatif aux recherches anterieures en nousinformant sur les implications des enjeux relationnels (p. ex., simpatia, personalismo) pour les emotions et la

Correspondence should be addressed to Krishna Savani, Management Division, Columbia Business School, 3022 Broadway

Uris 7L, New York, NY 10027, USA. (E-mail: ks2884@columbia.edu).

This research was supported by a Stanford University Vice-Provost for Undergraduate Education grant to Ayme Alvarezand National Science Foundation grant BCS-0844085 to Hazel Rose Markus. Krishna Savani was supported by a NationalScience Foundation Graduate Research Fellowship and the Beinecke Scholarship. We are most grateful and indebted to AmieBlocker, Jarreau Bowen, Supriya Dhar, Riah Forbes, John Giagnorio, Elizabeth Goodman, Aubrey Gress, Anke Kalaiah,Mbali Mphande, Diana Solorio, Mitali Thakor, Anant Vasudevan, and Wayva Waterman for their invaluable assistance withconducting this research.

International Journal of Psychology, 2013Vol. 48, No. 4, 682–694, http://dx.doi.org/10.1080/00207594.2012.688131

© 2013 International Union of Psychological Science

motivation dans les contextes mexicains. Aussi, ils sont les premiers a demontrer les effets motivationnels desemotions engageantes sur le plan interpersonnel.

S e investiga en dos estudios si las emociones de interaccion interpersonal —aquellas que producen unacercamiento del sı mismo a los demas (por ej., el afecto, la verguenza)— son centrales para el modelo del

self y las relaciones que prevalecen en los contextos culturales mexicanos. El Estudio 1 demostro que, en

comparacion con las personas de contextos europeo-americanos, los individuos de contextos mexicanos eran maspropensos a informar la experiencia de emociones de involucramiento interpersonal y tenıan menoresprobabilidades de informar la experiencia de emociones de desligamiento interpersonal. En el Estudio 2 seencontro que las emociones de involucramiento interpersonal ejercıan una influencia sustancial en la motivacion

del rendimiento en contextos mexicanos— los participantes mexicanos resolvieron mas tareas de ‘‘sopa de letras’’luego de recordar casos en los que experimentaron emociones positivas de interaccion interpersonal y resolvieronmenos despues de recordar las emociones negativas de desligamiento interpersonal; en contraste, no hubieron

diferencias segun condicion para los norteamericanos de origen europeo. Estos hallazgos amplıansignificativamente la investigacion previa al documentar las implicaciones de los aspectos relacionales (por ej.,la simpatıa, el personalismo) para las emociones y la motivacion en contextos mexicanos, y son los primeros en

demostrar los efectos motivacionales de las emociones de involucramiento interpersonal.

The frequency and intensity of various emotionsdiffer across cultural contexts (Barrett, 2006). Onesource of this variation is culturally pervasivemodels of self and of relationships, the commonand desirable ways of being a person and havingrelationships with others (Kitayama & Markus,1994; Mesquita, 2003; Mesquita & Leu, 2007;Oishi & Diener, 2003; Oishi, Koo, & Akimoto,2008; Shweder & Haidt, 2000; Tsai, 2007). Forinstance, the prevalence of happiness in NorthAmerican cultural contexts is linked to the culturalideals of independence from others, success, andpersonal achievement—happiness is supposed toboth help achieve and signal these ideals(D’Andrade, 1987; Hochschild, 1995). In certainMediterranean contexts, the systematic avoidanceof shame may be explained by the high valueplaced on honor, which can only be bestowed anddenied in relationship with others (Abu-Lughod,1986; Antoun, 1968; Ginat, 1979). In this manner,emotions tend to be more prevalent when they fitthe cultural models of self and of relationships,and less prevalent when they are inconsistent withthem (Mesquita, 2003).Recent comparative research on the prevalence

of emotions in Japan and the US has also foundevidence for a match between commonly experi-enced emotions and cultural ideals (Kitayama,Markus, & Kurokawa, 2000; Kitayama, Mesquita,& Karasawa, 2006). In this research, performingmultidimensional scaling on participants’ ratingsof a diverse set of emotion terms, Kitayama andcolleagues found that emotions are defined notonly by their valence and arousal but also by theirrelational outcomes. Interpersonally engagingemotions are those that create a positive linkbetween the self and others (such as affection and

shame), whereas interpersonally disengaging emo-tions are those that create a negative link betweenself and others (such as pride and anger). Toelaborate, a negatively valenced emotion such asshame can be interpersonally engaging because itoften arises from the experiencer’s failure to meetnormative social standards; an acknowledgment ofone’s failure attempts to repair one’s relationshipwith the others who were disappointed. Similarly,a positively valenced emotion such as pride can beinterpersonally disengaging because it indicatesthat one is superior to others, which increases thedistance between the self and others. Kitayamaet al. (2000, 2006) found that consistent with thecultural models of interdependence and indepen-dence, respectively, Japanese report experiencinginterpersonally engaging emotions more fre-quently, whereas US Americans report experien-cing interpersonally disengaging emotions morefrequently.

In the present studies, we extend previousresearch on culture and emotion to Mexicancultural contexts. Drawing on previous research,we predict that interpersonally engaging emotionswill be more desirable and thus more frequent inMexican cultural contexts than in US Americancontexts. We test this hypothesis using a situa-tional sampling paradigm that assesses the fre-quency of different emotions in actualinterpersonal situations, rather than through self-report. We then test for the first time whether,given the significance of interpersonally disenga-ging emotions for independent selves and ofinterpersonally engaging emotions for interdepen-dent selves, these emotions would differentiallymotivate individuals’ performance on unrelatedtasks in the two cultural contexts.

INTERPERSONALLY ENGAGING EMOTIONS 683

INTERPERSONALLY ENGAGING ANDDISENGAGING EMOTIONS

Extensive previous research, including crosscul-tural comparative research, has mapped emotionsin a space defined by the dimensions of valenceand arousal (Russell, 1983; Yik & Russell, 2003).In this research, respondents are typically asked torate the similarity of pairs of emotion terms (andtheir translations), considered in Western lan-guages to be the most common and basic. Byexpanding the commonly used set of emotionterms to include indigenous Japanese emotionwords ‘‘that presuppose the presence of others’’(Markus & Kitayama, 1991, p. 238), Kitayamaet al.’s (2000) multidimensional clustering analysisidentified a novel dimension of emotional experi-ence, the dimension of interpersonal engagement.

The dimension of interpersonal engagementdistinguishes between emotions in ways that arenot captured by either valence or arousal. Forexample, pride and affection are similar in thatthey are both positively valenced emotions.However, the interpersonal significance of theseemotions differs considerably. Pride invokes afeeling of achievement, a sense of self-satisfactionand superiority to others, whereas affectioninvokes a feeling of interdependence, a sense ofcloseness and communion with others. Pride isthus an example of an interpersonally disengagingemotion (an emotions that distances the self fromothers), whereas affection is an example of aninterpersonally engaging emotion (an emotion thatbrings the self closer to others). On the other hand,whereas anger and shame are both intenselynegative emotions, anger distances people fromothers and disrupts interpersonal harmony, and isthus categorized as an interpersonally disengagingemotion. Shame, by acknowledging one’s wrong-doing, attempts to bring the self closer to others,and thus is categorized as interpersonally engaging.Kitayama et al. (2000) found that a cluster analysisof participants’ similarity ratings of 31 emotionalterms yielded three dimensions, corresponding tovalence, arousal, and interpersonal engagement,respectively.

The interpersonal dimension of emotions isintimately related to cultural models of self andrelationships. In addition to finding that thefrequency of interpersonally engaging and disen-gaging emotions differs across cultural contexts,Kitayama et al. (2000) found that general positiveemotions indicative of wellbeing, such as happy,calm, and elated, were positively correlated withinterpersonally disengaging emotions (e.g., pride)in US American contexts, but with interpersonally

engaging emotions (e.g., affection) in Japanesecontexts. Kitayama et al. (2006) replicated these

findings using a daily diary study tapping actualemotional experiences. They found that Japaneseparticipants were more likely to experience generalpositive emotions in the situations where they alsoexperienced positive interpersonally engaging emo-

tions (e.g., affection), whereas American partici-pants were more likely to experience generalpositive emotions in the situations where theyalso experienced positive interpersonally disenga-ging emotions (e.g., pride). These findings demon-strate that the interpersonal dimension of

emotions has substantial implications for people’swellbeing.

EMOTION IN MEXICAN CULTURALCONTEXTS

As in Japanese cultural contexts, in which indivi-duals are conceptualized as interdependent andinterrelated, Mexican cultural contexts foster theidea that ideal relationships are warm andemotionally close (Greenfield & Cocking, 1994;

Triandis, Marin, Lisansky, & Betancourt, 1984).The idea of simpatia in Mexican contexts has beencharacterized as a ‘‘highly valued relational style[that] resembles the search for social harmonycharacteristic of many East Asian cultures butincludes an emphasis on expressive displays of

personal charm, graciousness, and hospitalitymore specific to Latin cultures . . . A person whois simpatico(a) is one who proactively attempts tocreate a highly personable atmosphere as an end initself’’ (Sanchez-Burks, Nisbett, & Ybarra, 2000,p. 175; see also Sanchez-Burks, Bartel, & Blount,

2009). Further, according to the concept ofpersonalismo that is prevalent in Mexican contexts,people should strive to create ‘‘a highly persona-lized communication style that is characterized byinterdependence and cooperation . . . The valuesunderlying personalismo are self-worth, the dignity

of self and others, and respeto (respect shown toothers)’’ (Holloway, Waldrip, & Ickes, 2009,p. 1012). An analysis of people’s spontaneousself-descriptions found that relationships andsimpatia were unique aspects of Mexicans’ self-concepts, not shared by Americans’ self-concepts

(Ramirez-Esparza, Chung, Sierra-Otero, &Pennebaker, 2012). This research collectivelysuggests that Mexican cultural contexts emphasizethe importance of positive and warm forms ofrelating to others more than European Americancultural contexts.

684 SAVANI ET AL.

If emotions that fit with cultural models of selfand relationships are more prevalent, we wouldexpect that people in Mexican cultural contextswould be more likely than European Americans toexperience interpersonally engaging emotions, andless likely to experience interpersonally disenga-ging emotions; some crosscultural research sug-gests this might be the case. In one of the firstcrosscultural analyses of US American andMexican contexts from a psychological perspec-tive, Triandis et al. (1984) found that people fromvarious Hispanic backgrounds were more likelythan European Americans to expect positiveinteractions in various interpersonal situationsand less likely to expect negative interpersonalbehaviors; if these expectations are grounded inreality, then warm, agreeable interactions might bemore prevalent in Hispanic American settings thanin non-Hispanic European American settings.Further, Sanchez-Burks et al. (2000) found thatMexican participants preferred workgroups thathad a socioemotional focus in addition to a taskorientation, whereas European Americans pre-ferred exclusively task-oriented workgroups.These socioemotionally oriented groups giveimportance to positive relational interactionsbetween group members in addition to efficienttask performance, and hence can be expected toinduce positive engaging emotions.Research by Holloway et al. (2009, Studies 1

and 2) also suggests that positive interpersonallyengaging emotions may be more prevalent inMexican contexts than in European Americancontexts. Using the 20 statement test, they foundthat when compared to European Americanstudents, Mexican American students were morelikely to describe themselves using simpatia-relatedterms, such as likable, sympathetic, polite, plea-sant, amiable, congenial, agreeable, easygoing,understanding, gracious, tolerant, respectful, con-siderate, friendly, and courteous, many of whichrefer to positive interpersonally engaging emo-tions. Further, Holloway et al. (2009, Studies 3aand 3b) found that in dyadic interactions, MexicanAmerican participants were more likely to expresspositive interpersonally engaging behaviors (e.g.,smiling, partner-directed gazing) than EuropeanAmerican and African American participants, andpartners of Mexican Americans reported havingbetter quality interactions compared to partners ofAfrican Americans and European Americans.These findings suggest that overall, people enga-ging in Mexican cultural contexts might be morelikely to experience and express interpersonallyengaging emotions than non-Hispanic USAmericans.

Ramirez-Esparza et al. (2012) provided addi-tional support for the idea that warm relationshipsare central to self-concepts in Mexican culturalcontexts. They collected spontaneous self-descrip-tions of Mexican and American participants andanalyzed them according to the meaning extractionmethod, a text-analytic procedure designed forextracting self-concepts from open-ended self-descriptions (Chung & Pennebaker, 2008). Theresearchers performed factor analyses (within eachculture) on the co-occurrence of the 40 mostfrequent self-descriptors to identify the key dimen-sions of the self-concept in each culture. Mostlystrikingly, they found that the first and mostprominent factor in Mexicans’ self-concept wasrepresented by the concept of relationships (e.g.,parents, friends, family, love, hope, pain). Anotherfactor referred to simpatia (e.g., affectionate,responsible, devoted). However, both these dimen-sions were absent in Americans’ self-concepts,indicating that warm, positive relationships areparticularly salient in Mexican cultural contexts.

Although Japanese and Mexican culturesappear similar in their emphasis on the interperso-nal dimension of emotion, there is some suggestionthat Japanese have avoidance goals whereasMexicans have approach goals. Research suggeststhat Japanese are particularly motivated byavoiding negative engaging emotions, such asshame. For example, Japanese are more motivatedafter failure than after success feedback (Heineet al., 2001), and are more likely to learn frommistakes (Stevenson & Stigler, 1994). Further,whereas Asian Americans tend to experience asimilar degree of positive and negative emotions,European Americans experience more positivethan negative emotions, and Hispanic Americansexperience even more positive emotion thanEuropean Americans (Oishi, Diener, Scollon, &Biswas-Diener, 2004). Together, the above find-ings suggest that whereas Japanese might have thegoal of avoiding negative emotions, particularlynegative interpersonally engaging emotions,Mexicans might have the goal of approachingpositive emotions, particularly positive interper-sonally engaging emotions.

While previous research has found that inter-personally engaging self-concepts and behaviorsare more common in Mexican cultural contextsthan in European American cultural contexts, inthe present research, we tested for the first timewhether Mexicans and European Americans differin how often they experience interpersonallyengaging and disengaging emotions. The researchreviewed above suggests that Mexicans might bemore likely than European Americans to

INTERPERSONALLY ENGAGING EMOTIONS 685

experience interpersonally engaging emotions andless likely to experience interpersonally disenga-ging emotions.

STUDY 1

Study 1 aimed to assess the relative prevalence ofsocially engaging vs. socially disengaging emotionsin Mexican and European American contexts bysampling emotion-inducing situations from bothcultural contexts. We hypothesized that Mexicanswould report experiencing interpersonally enga-ging emotions more often than EuropeanAmericans, but that European Americans wouldreport experiencing interpersonally disengagingemotions more often than Mexicans.

Previous research assessed crosscultural varia-tion in the frequency of interpersonally engagingand disengaging emotions by asking participantsto rate the extent to which they experienced anumber of engaging and disengaging emotions,which were selected to be representative of theemotion domain in Japanese and English lan-guages (e.g., Kitayama et al., 2000, 2006). Instead,we asked respondents to report the emotions thatthey experienced in recent emotion-eliciting situa-tions that occurred in their lives. We made thismethodological choice for two reasons. First,when using rating scales, researchers need to pre-select a list of emotions; however, we did not wishto make assumptions about the most frequentemotions in Mexican contexts. Second, cross-cultural comparisons with rating scales sufferfrom a number of biases due to variable referencepoints, social comparisons, and contrast effectsdue to deprivation (Heine, Lehman, Peng, &Greenholtz, 2002; Peng, Nisbett, & Wong, 1997).

To alleviate these concerns, we asked partici-pants to describe recent situations in which theyexperienced positive and/or negative emotions,and asked them to describe the emotion(s) thatthey experienced in each situation; we then codedthe described emotions on the dimensions ofvalence and interpersonal engagement. The pre-sent methodology allows us to sample emotionsthat individuals experience in their everyday lives,thus avoiding possible experimenter biases asso-ciated with preselecting a list of emotions. Further,by assessing the relative frequency of differentemotions based on their occurrence in a largesample of everyday situations, we avoid confoundsassociated with crosscultural comparisons usingrating scales, such as the reference group effect(Heine et al., 2002) and social comparison anddeprivation (Peng et al., 1997). However, because

culture shapes what information is salient inmemory (Masuda & Nisbett, 2001; Oishi, 2002),participants might not recall a random or repre-sentative sample of affectively laden situations thatthey experienced in their daily lives, but a sampleof situations that is more salient in memory.Therefore, the situations that participants recallmight reflect the salience of specific emotionalevents in memory, as well as actual emotionalexperience.

Method

Participants

A total of 74 European American students atStanford University (42 women, 32 men, mean age21.3 years) and 80 Mexican students atUniversidad de Guadalajara, Mexico participatedin this study. Due to a clerical error, gender andage data are available for only 41 Mexicanparticipants (16 women, 25 men; mean age 20.0years).

Materials

Participants responded to a situation samplingprompt asking them to describe recent emotion-inducing situations that occurred in their lives. Toobtain a diverse sample of situations, we askedparticipants to describe situations involving posi-tive affect and situations involving negative affect.The complete prompt was:

Please take a moment to reflect upon the last 7

days and describe all the situations in which you

felt good or bad. Feel free to describe any type of

situation that you encountered and describe your

feeling in that situation. Please be specific when

describing both the situation and the feeling. In

the following pages you will find space to list

many situations. While it is not necessary to fill all

the spaces, we do request that you fill in as many

as you can. Please continue to think about the

situations for the next 15 minutes and write down

as many situations as you can. The experimenter

will inform you once 15 minutes have passed. If

you have any questions at all, please ask the

experimenter.

Participants were then provided with two pagesin which they could describe up to 16 situations.For each situation, participants were asked tospecify the emotions that they experienced inthat situation. The questionnaire was preparedin English, translated into Spanish, and then

686 SAVANI ET AL.

back-translated to ensure comparability with theoriginal English version.

Results

Mexican participants generated 10.4 situations onaverage, for a total of 822 situations, whileAmerican participants generated 12.5 situationson average, for a total of 913 situations. It ispossible that cultural differences in the practice ofcompleting surveys, or perhaps cultural differencesin the tendency to reflect on one’s internal states,might have contributed to this difference. Of thesesituations, we took a random sample of 200situations from each culture (as in Kitayama,Markus, Matsumoto, & Norasakkunkit, 1997).For each of these situations, we noted all theemotions that the participant mentioned, resultingin a total of 707 emotional terms, 315 generatedby Mexican participants and 392 generated byAmerican participants).Below are some sample situations.

Sample Mexican situationsMy newborn nephew opened his eyes after

sleeping and he fixated a look on my face for 15

seconds. I felt affection and harmony between

unequals.

This Sunday I was very hungover and began to

think about what I had done the night before.

I felt guilt and shame.Sample American situations:Monday was my birthday and my best friend at

school didn’t get me a gift or even say ‘‘happy

birthday.’’ I felt bad.

Baking cookies with a roommate at 10 pm to put

off a problem set for a class we’re both in, others

wandered in and out of the kitchen happy

that we were making cookies. I felt pride,

accomplishment.

A Spanish-English bilingual research assistant,born in Mexico and raised in the US, translated allthe emotion terms listed by Mexican participantsfrom Spanish to English. The research assistantwas asked to take particular care to ensure thatany possible indigenous emotion terms were notmistranslated. The second author, who was alsoborn in Mexico and raised in the US, back-translated the emotion terms from English toSpanish, and discrepancies were resolved bydiscussion.Thereafter, using the classification guidelines

employed by Kitayama et al. (2000, 2006), twotrained research assistants, one Mexican Americanand one Native American, independently coded

the English version of each of these emotions onthe dimensions of valence (i.e., positive or nega-tive) and interpersonal engagement (i.e., interper-sonally engaging, interpersonally disengaging, orneutral). Emotions were independently coded onthese two dimensions. An emotion was coded asinterpersonally engaging if it created a positive linkbetween the self and others (e.g., affection,gratitude, sympathy), or attempted to resolve anegative link between the self and others (e.g.,shame, embarrassment, guilt). An emotion wascoded as interpersonally disengaging if it high-lighted the self as autonomous and separate fromothers (e.g., pride, superior, confident), or if itdistanced the self from others (e.g., angry,frustrated, annoyed). Intercoder agreement was99.7% for the valence of the emotion and 95.2%for the interpersonal engagement of the emotion.Disagreements were resolved by discussion.

A 2 (Culture)� 2 (Valence) chi-square testrevealed that Mexicans and European Americansreported experiencing a similar proportion ofpositive and negative emotions in the situationsthat they described, w2 (df¼ 1, N¼ 707)¼ .63,p4 .22; 49% of the emotions generated byAmericans and 51% of the emotions generatedby Mexicans were positive. It is possible that bothgroups described a very similar proportion ofpositive and negative emotions because they wereinstructed in the prompt to recall situations inwhich they felt ‘‘good or bad,’’ that is, to generateabout equal numbers of positive and negativesituations. Notably, this finding ensures thatcultural differences in valence of the emotionsexperienced do not account for any possiblecultural differences in the interpersonal engage-ment of the emotions.

To assess cultural variation in the interpersonaldimension of emotion, we first conducted a 2(Culture)� 3 (Interpersonally engaging vs. disen-gaging vs. neutral) chi-square test of independence,which yielded a significant effect, w2 (df¼ 2,N¼ 257)¼ 12.27, p¼ .002, Cramer’s V¼ .13. Tospecifically compare the relative prevalence ofinterpersonally engaging and disengaging emo-tions, we excluded emotion terms that were codedas interpersonally neutral (e.g., good, bad; 68% ofUS-generated emotions, and 58% of Mexico-generated emotions). We next conducted a 2(Culture)� 2 (Interpersonally engaging vs. disen-gaging) chi-square test of independence, whichyielded a significant effect, w2 (df¼ 1,N¼ 257)¼ 5.48, p5 .02, Cramer’s V¼ .15.European American participants reported moreinterpersonally disengaging emotions thanMexican participants, M¼ 75% vs. 62%,

INTERPERSONALLY ENGAGING EMOTIONS 687

whereas Mexican participants reported moreinterpersonally engaging emotions than EuropeanAmericans, M¼ 38% vs. 25% (see Figure 1).

Discussion

Study 1 found that overall, both groups reportedexperiencing more disengaging emotions thanengaging emotions in the situations that theyrecalled, perhaps because all participants were incompetitive college environments, probably awayfrom home for the first time, which might fosterdisengaging emotions (e.g., pride in one’s ownsuccess, jealousy of others’ successes). However,consistent with the greater emphasis on warm,agreeable relationships in Mexican cultural con-texts, Mexicans were more likely to experienceinterpersonally engaging emotions than EuropeanAmericans, and consistent with the greater empha-sis on independence in European Americancultural contexts, European Americans weremore likely to experience interpersonally disenga-ging emotions than Mexicans.

Whereas previous research assessed culturalvariation in prevalence of interpersonally engagingand disengaging emotions using participants’ratings of emotion frequency as the dependentmeasure (Kitayama et al., 2000, 2006), we con-ceptually replicated these findings using a differentparadigm that sampled emotions experienced invarious everyday situations. These findings suggestthat the greater cultural emphasis on closerelationships in Mexican contexts compared toEuropean American contexts (Holloway et al.,2009; Ramirez-Esparza et al., 2012; Sanchez-Burkset al., 2000, 2009) is also reflected in people’sreports of their emotional experiences.

STUDY 2

Study 1 documented crosscultural variation in theprevalence of interpersonally engaging emotions,which we hypothesized derive from differencesbetween the models of self and of relationshipsprevalent in American and Mexican contexts.These models prescribe how to feel and also howto act. In Study 2 we took a step further to test forthe first time whether the emotions that areprevalent in particular context might also impactpeople’s motivation in that context. The facilita-tion of adaptive behaviors is often considered to bean important function of affect (Frijda, 1999).That is, emotions are functional in that they inducepeople to engage in necessary and appropriatebehaviors. For example, stimuli eliciting negative

affect facilitate avoidance behaviors, whereas

stimuli eliciting positive affect facilitate approach

behaviors (Chen & Bargh, 1999), and negative

emotional states amplify startle responses, whereas

positive emotional states attenuate them (Lang,

Bradley, & Cuthbert, 1990).Previous research on the relationship between

emotion and motivation has examined how

affective valence influences appetitive motivation.

Winkielman, Berridge, and Wilbarger (2005)

manipulated affective valence by subliminally

exposing participants to faces with either smiling

expressions or frowning expressions, and then gave

participants the opportunity to consume a pleasant

beverage. They found that participants’ appetitive

motivation for the beverage was enhanced after

subliminal priming with positive valence and

dampened after subliminal priming with negative

valence: Participants consumed more beverages,

wanted the beverage to a greater extent, and were

willing to pay more money for the beverage in the

positive valence condition than in the negative

valence condition.We theorized that given the importance of

warm, agreeable relationships for the self-concept

in Mexican cultural contexts (Holloway et al.,

2009; Ramirez-Esparza et al., 2012; Sanchez-Burks

et al., 2000), positive interpersonally engaging

emotional states will be functional, and therefore

more motivating, in Mexican cultural contexts.

Specifically, we hypothesized that positive inter-

personally engaging emotions, such as affection

and closeness, would signal that culturally norma-

tive goals of connectedness have been met, and

thus make individuals feel more motivated.

However, given that positive interpersonally enga-

ging emotions are less relevant to the independent

self-concept prevalent in European American

contexts, these emotional states would be less

Figure 1. Proportion of interpersonally engaging and disenga-

ging emotions recalled by participants in Study 1.

688 SAVANI ET AL.

functional and thus would not influencemotivation.Negative engaging emotions obstruct approach

motivation in relational contexts; they signal thatinterpersonal relationships have been disruptedbecause of the individuals’ actions, and that theindividual can no longer approach others. Forexample, shame highlights that one has violatedothers’ expectations and compels one to intro-spect. Therefore, we expected Mexicans to beparticularly demotivated by negative engagingemotions because the culturally mandated inter-dependent self is threatened. In contrast, asdisruptions of interpersonal relationships are lessself-relevant for more independent EuropeanAmericans, we hypothesized that negative enga-ging emotions would not influence EuropeanAmericans’ motivation.Given the importance of independence for the

self-concept in European American cultural con-texts, positive interpersonally disengaging emo-tional states are likely to be more functional andtherefore, more motivating. Specifically, positivedisengaging emotions, like pride, signal that theperson has succeeded in the goal of distinguishingthemselves from others and in creating a positiveimage of the self. Therefore, we hypothesized thatpositive disengaging emotions would motivateEuropean Americans. However, given that posi-tive interpersonally disengaging emotions are notas relevant for interdependent selves, which aremore prevalent in Mexican contexts, we hypothe-sized that they would not influence Mexicans’motivation.Negative disengaging emotions are more rele-

vant to the independent self; they arise whenpersonal goals have been blocked. For example,frustration occurs when an individual is unable todo what they want to do. Therefore, we expectedEuropean Americans to be particularly demoti-vated by negative engaging emotions because theculturally pervasive independent self is threatened.In contrast, given that the blocking of personal

goals is not as threatening to interdependentselves, we hypothesized that negative disengagingemotions would not influence Mexicans’ motiva-tion (see Table 1 for a depiction of the hypotheses).

We measured performance motivation as theoutcome variable in our study, given its ecologicalvalidity in the educational context. Most testsadministered in educational settings measureperformance, e.g., how many questions studentsanswer correctly in a fixed duration. Anotheroption would have been to measure persistence,e.g., how long students work on a task without anytime limit. We chose to measure performance overpersistence given the greater ecological validity ofperformance in academic settings, and given thatboth performance and persistence are equally validindicators of motivation (e.g., Eccles, Wigfeld, &Schiefele, 1998; Maehr, 1984).

Method

Participants

Participants were 198 Mexican students from LaUniversidad de Guadalajara in Jalisco, Mexico (99women, 99 men; mean age 19.9 years, range 17–31;95 with at least one college-educated parent), and199 European American students from StanfordUniversity in northern California and WakeForest University in North Carolina (115women, 84 men; mean age 19.1 years, range17–24; 195 with at least one college-educatedparent). There were no significant effects ofparticipants’ gender or parental education.

Procedure

Participants first completed an ‘‘EmotionStudy’’ in which the types of emotion that theywere asked to recall were manipulated betweensubjects in a 2 (valence: positive vs. negative)� 2(interpersonal engagement: engaging vs. disenga-ging) design. We used the emotional recall taskbased on past research showing that the reliving ofemotional situations is an effective way to elicitemotions (Levenson, Carstensen, Friesen, &Ekman, 1991). Participants were instructed (infor-mation in italics was varied between subjectsaccording to the assigned conditions):

Please take a moment to reflect upon your last few

weeks and describe all situations in which you felt

good/bad towards another person/about yourself.In the following pages you will find space to list

many situations. While it is not necessary to fill all

the spaces, we do request that you fill in as many

TABLE 1Depiction of hypotheses for Study 2

Condition European Americans Mexicans

Positive engaging 0 þPositive disengaging þ 0

Negative engaging 0 –

Negative disengaging – 0

‘‘þ’’ indicates a hypothesized increase in motivation; ‘‘–’’

indicates a hypothesized decrease in motivation; ‘‘0’’ indicates a

hypothesized null effect on motivation

INTERPERSONALLY ENGAGING EMOTIONS 689

as you can. Please continue to think about the

situations for the next 10 minutes and write down

as many situations as you can. The experimenter

will inform you once 10 minutes have passed. If

you have any questions at all, please ask the

experimenter.

Participants were provided with space todescribe up to nine situations. In each condition,to give participants a concrete idea of the types ofemotion that we wanted them to recall, weprovided participants with examples of emotionsfrom the target category drawn from Kitayamaet al.’s (2000) classification.1 The questionnairewas prepared in English and translated intoSpanish using the standard back-translationprocedure.

Ten minutes later, the experimenter interruptedparticipants and asked them to proceed to a ‘‘wordsearch puzzle,’’ which served as a measure ofperformance motivation. Participants were pre-sented with a word search matrix of 15� 15 lettersand provided with a list of 20 target words (allnouns) that were presented in the matrix. Theywere given 5min to identify the target words in thematrix. All words were easily translatable fromEnglish to Spanish and all but one had the samenumber of letters in both languages.

Results

We counted the number of the 20 target words thatparticipants identified in the word search matrix,which served as the dependent measure. Overall,Mexicans identified 8.2 words and EuropeanAmericans identified 11.4 words. Initial analysesindicated that age was positively correlated withperformance for Mexicans, r¼ .17, p¼ .02, but notfor Americans, r¼�.01, p4 .91. Therefore, wecontrolled for age in the following analyses.

We submitted this measure to a 2 (Culture)� 2(Valence)� 2 (Engagement) ANCOVA, with amain effect of age and a Culture�Age interactionas covariates. We found a main effect of Culture,F(1, 382)¼ 69.41, p5 .001, a Culture�Valenceinteraction, F(1, 382)¼ 7.42, p5 .01, and aCulture�Valence�Engagement interaction, F(1,382)¼ 9.93, p5 .005, Cohen’s f¼ .15.

To decompose the three-way interaction, wesplit the data by culture and conducted separate 2(Valence)� 2 (Engagement) ANCOVAs (with age

as a covariate) for each culture. None of the effects

were significant for European American partici-

pants, F values5 2.41, p values4 .12, indicating

that this particular manipulation of valence and

interpersonal engagement did not influence their

performance (see Figure 2). A separate ANOVA

including university (Stanford vs. Wake Forest) as

a predictor found a main effect of university, F(1,

191)¼ 13.49, p5 .001, M¼ 10.08 (Wake Forest)

vs. 12.17 (Stanford), but university did not interact

with valence, engagement, or the valence� en-

gagement interaction, F values5 2.55, p

values4 .11. Follow-up t-tests found that the

number of words American participants solved in

the two disengaging conditions did not differ,

t(95)¼ .36, p4 .70, but Americans solved margin-

ally more words in the negative engaging condition

than in the negative disengaging condition,

t(93)¼ 1.70, p¼ .09, d¼ .35.In contrast, for Mexican participants, we found

a main effect of age, F(1, 197)¼ 3.79, p¼ .05, a

main effect of valence, F(1, 197)¼ 9.59, p5 .005,

and the predicted Valence�Engagement interac-

tion, F(1, 197)¼ 7.80, p¼ .006, Cohen’s f¼ .20.

Supporting our hypothesis, Mexican participants

found more words in the positive engaging

condition than in the negative engaging condition,

M values¼ 9.47 vs. 6.45, t(98)¼ 4.97, p5 .0001,

d¼ 1.0, while there was no significant difference

between the positive and negative disengaging

conditions, M values¼ 8.34 vs. 8.35, t5 1,

p4 .75 (see Figure 3).We then conducted planned orthogonal con-

trasts to test our hypothesis more precisely with

the data from Mexican participants. Contrast 1

tested whether participants solved more words in

Figure 2. Number of word search puzzles solved by European

American participants in Study 2. Error bars represent

standard errors of the mean.

1The examples were friendly towards another person, happy for another person, close to another person, and respect foranother person in the positive interpersonally engaging condition; sad for another person, indebted to another person, andembarrassed in the negative engaging condition; superior, proud, happy for yourself, and self-satisfied in the positivedisengaging condition; and frustrated, mad at yourself, sad, and moody in the negative disengaging condition.

690 SAVANI ET AL.

the two positive valence conditions compared tothe two negative valence conditions; Contrast 2tested whether participants solved more words inthe positive engaging condition compared to thepositive disengaging condition; finally, Contrast 3tested whether participants solved fewer words inthe negative engaging condition compared to thenegative disengaging condition (see Table 2 forcontrast coefficients). We also controlled for age inthis analysis.Contrast 1 was significant, b¼ .76, t(192)¼ 3.10,

p¼ .001 (one-tailed), indicating that participantssolved more words in the positive valence condi-tions than in the negative valence conditions.Contrast 2 was significant, b¼ .57, t(192)¼ 1.66,p5 .05 (one-tailed), indicating that participantssolved more words in the positive engagingcondition than in the positive disengaging condi-tion. Finally, Contrast 3, was also significant,b¼�.80, t(192)¼ 2.28, p5 .02 (one-tailed), indi-cating that participants solved fewer words in thenegative engaging condition than in the negativedisengaging condition.Finally, we tested whether there were cultural

differences in performance across the four condi-tions. Americans outperformed Mexicans in thepositive disengaging condition, t(94)¼ 8.14,p5 .001, the negative disengaging condition,t(97)¼ 3.88, p5 .001, the negative engaging con-dition, t(94)¼ 2.20, p5 .001, and marginally so inthe positive engaging condition, t(97)¼ 1.80,p5 .08.

Discussion

The results document that interpersonally relevantaffective states have substantial effects on Mexicanparticipants’ performance motivation: Positiveinterpersonally engaging emotions motivatedMexican participants to perform on an unrelatedtask, whereas negative interpersonally engagingemotions undermined their motivation. Positiveand negative interpersonally disengaging emotionsdid not influence participants’ performance, indi-cating that emotions that have implications for theculturally sanctioned self are more likely toinfluence motivation. The rather large effect sizeobserved for the difference between the twoengaging conditions, Cohen’s d¼ 1.0, suggeststhat interpersonally engaging emotions might beimportant antecedents of motivation in Mexicancultural contexts.

Contrary to predictions, European Americanparticipants’ performance motivation did not varyby condition. The overall better performance byEuropean Americans compared to Mexicansmight be due to greater practice with crosswordpuzzles. It is possible that European Americansmight have reached a ceiling effect in the numberof words that they could reasonably solve in thegiven time, although theoretically, they could havesolved eight more words. Alternatively, maybeEuropean Americans were overall more motivatedat baseline than Mexicans, and thus their motiva-tion had less room to be influenced by themanipulation. It might also be possible that forsome reason, European Americans’ performancemotivation is not as sensitive as Mexicans’performance motivation to the particular emotionmanipulation that we employed. Future researchwould be needed to explore these possibilities.

GENERAL DISCUSSION

Based on previous research highlighting theimportance of warm interpersonal relationshipsin Mexican cultural contexts, as conveyed by theconcepts of personalismo and simpatica, we exam-ined cultural differences in the prevalence andimpact of interpersonally engaging and disenga-ging emotions. Sampling everyday emotion-elicit-ing situations in Mexican and European Americancultural contexts, Study 1 found that Mexicanswere more likely than European Americans toreport experiencing interpersonally engaging emo-tions, whereas European Americans were morelikely than Mexicans to report experiencing inter-personally disengaging emotions. These findings

TABLE 2Depiction of planned contrasts used in Study 2

Condition Contrast 1 Contrast 2 Contrast 3

Positive engaging 1 1 0

Positive disengaging 1 �1 0

Negative engaging �1 0 �1

Negative disengaging �1 0 1

Figure 3. Number of word search puzzles solved by Mexican

participants in Study 2. Error bars represent standard errors of

the mean.

INTERPERSONALLY ENGAGING EMOTIONS 691

demonstrate that interpersonally engaging emo-tions are more common in non-Western culturesbeyond Japan. Notably though, both groupsreported experiencing more disengaging emotionsthan engaging emotions.

Extending previous research on the relationshipbetween affect and motivation (Winkielman et al.,2005), Study 2 documented for the first time thatinterpersonally relevant emotional states canenhance performance motivation on unrelatedtasks. Mexican participants solved more wordsearch puzzles after recalling situations in whichthey experienced positive interpersonally engagingemotions, and fewer puzzles after recalling nega-tive interpersonally engaging emotions, whereaspositive and negative disengaging emotions had noinfluence. This finding shows that Mexicans’motivation is highly sensitive to the interpersonalsignificance of their current affective states. Incontrast, European Americans’ motivation wasnot influenced by the emotion manipulations,indicating that their performance motivation isless influenced by their emotional states than thatof Mexicans.

While the present studies were an initial attemptto examine the prevalence and motivationalconsequences of interpersonally relevant affectivestates in Mexican contexts, many questionsremain. Although Study 1 documented thatMexicans are more likely than EuropeanAmericans to recall situations in which theyexperienced interpersonally engaging emotions, itis possible that overall, the two groups experiencea similar frequency of engaging in disengagingemotions but that engaging situations are rela-tively more salient for Mexicans, and disengagingemotions less salient. Future research might testthis possibility by providing both groups ofparticipants with the same set of situations andasking them to rate the extent to which they wouldexperience different engaging and disengagingemotions in each situation (e.g., Kitayama et al.,2006).

The results of Study 2 are provocative but futureresearch needs to more closely examine the typesof motivations that are influenced by interperson-ally relevant affective states. Together withWinkielman et al. (2005), our study shows thatemotions can influence a wide range of outcomesranging from appetitive motivation to perfor-mance motivation, which is not surprising giventhe functional overlap between emotion andmotivation at the psychological level (Frijda,1999). Future research might cross differentemotion manipulations with different motivationaloutcomes to document which types of

motivational outcomes are most influenced bywhich emotional states. Our finding that thevalence and interpersonal engagement emotionsdid not influence European American participants’performance, particularly when contrasted withthe large effect observed among Mexicans, raisessome questions. Is it the case that the manipulationwas less effective in arousing the target emotions inEuropean Americans than in Mexicans, or areinterpersonally relevant affective states simply lessmeaningful for European Americans? Might thesame manipulation influence some other measureof motivation for European Americans? Futureresearch might explore these intriguingpossibilities.Although this study was conducted with stu-

dents in Mexico, the findings are relevant toMexican American students because even second-and third-generation Mexican Americans engagewith the ideas and practices of Mexican contexts intheir families (Valdes, 1996). Educators in the USoften try to motivate students by making them feelgood about themselves, proud of themselves,confident of themselves, and so on; they seek toevoke these positive interpersonally disengagingemotions because these emotions highlight the selfas distinct from others and better than others, as isnormative in European American settings.However, the above findings suggest that studentsfrom Mexican backgrounds might instead be moremotivated by emotions that highlight their positiverelationships and closeness with others. Similarly,negative interpersonally engaging emotions, suchas shame and embarrassment, might particularlyundermine the motivation of Mexican students. Toclose the substantial gap between MexicanAmerican and European American students inacademic performance and high school graduationrates (Casserly, Petteruti, & Williams, 2006;Suarez-Orozco & Suarez-Orozco, 2001), educatorsin the US need to engage in culturally sensitivepedagogy by using different strategies to motivateand to avoid demotivating students from differentcultural backgrounds (Gutierrez & Rogoff, 2003;Markus, 2008; Trumbull, Rothstein-Fisch, &Greenfield, 1999).While the present research demonstrated the

motivating effects of interpersonally engagingemotions by using performance on a verbal taskas the dependent measure, to ensure that thefindings are broadly applicable to the educationalcontext, future research might further test whetherinterpersonally engaging emotions have similareffects on Mexican students’ performance onspatial and math tasks. Future research mightalso design an intervention to train teachers in

692 SAVANI ET AL.

school districts with large Latin American popula-tions to use positive interpersonally engagingemotions to motivate students, while minimizingthe use of negative interpersonally engagingemotions.Whereas researchers have focused on valence

and arousal as the two most important dimensionsof emotion, the present research highlights inter-personal engagement as perhaps an equally impor-tant dimension—one that is sensitive to culturalnuances and one that has significant motivationalconsequences.

Manuscript received June 2011

Revised manuscript accepted March 2012

First published online June 2012

REFERENCES

Abu-Lughod, L. (1986). Veiled sentiments. Berkeley,CA: University of California Press.

Antoun, R. T. (1968). On the modesty of women inArab Muslim villages. American Anthropologist, 70,671–696.

Barrett, L. F. (2006). Are emotions natural kinds?Perspectives on Psychological Science, 1, 28–58.

Casserly, M., Petteruti, A., & Williams, A. (2006).Beating the odds: A city-by-city analysis of studentperformance and achievement gaps on state assess-ments. Results from the 2004–2005 school year.Washington, DC: Council of the Great City Schools.

Chen, M., & Bargh, J. A. (1999). Consequences ofautomatic evaluation: Immediate behavioral predis-positions to approach or avoid the stimulus.Personality and Social Psychology Bulletin, 25,215–224.

Chung, K. C., & Pennebaker, J. W. (2008). Revealingpersonality in natural language: Using an automatedmeaning extraction method on open-ended self-descriptions. Journal of Research in Personality, 42,96–132.

D’Andrade, R. G. (1987). A folk model of the mind.In D. Holland, & N. Quinn (Eds.), Cultural models inlanguage and thought (pp. 112–149). Cambridge, UK:Cambridge University Press.

Eccles, J. S., Wigfeld, A., & Schiefele, U. (1998). Thedevelopment of achievement motivation.In N. Eisenberg (Ed.), Handbook of child psychology(5th ed., Vol. IV, pp. 1017–1095). New York, NY:Wiley.

Frijda, N. H. (1999). Emotions and hedonic experience.In D. Kahneman, E. Diener, & N. Schwarz (Eds.),Well-being: The foundations of hedonic psychology(pp. 190–210). New York: Russell Sage Foundation.

Ginat, J. (1979). Illicit sexual relationships and familyhonor in Arab society. Israeli Studies in Criminology,10, 179–202.

Greenfield, P. M., & Cocking, R. R. (1994). Cross-cultural roots of minority child development. Hillsdale,NJ: Lawrence Erlbaum Associates.

Gutierrez, K., & Rogoff, B. (2003). Cultural ways oflearning: Individual traits or repertoires of practice.Educational Researcher, 32, 19–25.

Heine, S. J., Kitayama, S., Lehman, D. R., Takata, T.,Ide, E., Leung, C., et al. (2001). Divergent motiva-tional consequences of success and failure in Japanand North America. Journal of Personality and SocialPsychology, 81, 599–615.

Heine, S., Lehman, D., Peng, K., & Greenholtz, J.(2002). What’s wrong with cross-cultural compar-isons of subjective Likert scales? The reference-groupeffect. Journal of Personality and Social Psychology,82, 903–918.

Hochschild, J. L. (1995). What is the American dream?In J. L. Hochschild (Ed.), Facing up to the Americandream: Race, class and the soul of the nation(pp. 15–38). Princeton, NJ: Princeton UniversityPress.

Holloway, R. A., Waldrip, A. M., & Ickes, W. (2009).Evidence that a simpatico self-schema accounts fordifferences in the self-concepts and social behavior ofLatinos versus Whites (and Blacks). Journal ofPersonality and Social Psychology, 96, 1012–1028.

Kitayama, S., & Markus, H. R. (Eds.) (1994). Emotionsand culture: Empirical studies of mutual influence.Washington, DC: American PsychologicalAssociation.

Kitayama, S., Markus, H. R., & Kurokawa, M. (2000).Culture, emotion, and well-being: Good feelings inJapan and the United States. Cognition and Emotion,14, 93–124.

Kitayama, S., Markus, H. R., Matsumoto, H., &Norasakkunkit, V. (1997). Individual and collectiveprocesses in the construction of the self: Self-enhancement in the United States and self-criticismin Japan. Journal of Personality and SocialPsychology, 72, 1245–1267.

Kitayama, S., Mesquita, B., & Karasawa, M. (2006).Cultural affordances and emotional experience:Socially engaging and disengaging emotions inJapan and the United States. Journal of Personalityand Social Psychology, 91, 890–903.

Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1990).Emotion, attention, and the startle reflex.Psychological Review, 97, 377–395.

Levenson, R. W., Carstensen, L. L., Friesen, W. V., &Ekman, P. (1991). Emotion, physiology, and expres-sion in old age. Psychology and Aging, 6, 28–35.

Maehr, M. (1984). Meaning and motivation: Toward atheory of personal investment. In R. Ames, &C. Ames (Eds.), Research on motivation in education(Vol. 1, pp. 115–144). New York: Academic Press.

Markus, H. R. (2008). Identity matters: Ethnicity, race,and the American Dream. In M. Minow,R. A. Shweder, & H. R. Markus (Eds.), Just schools:Pursuing equality in societies of difference(pp. 63–100). New York, NY: Russell SageFoundation.

Markus, H. R., & Kitayama, S. (1991). Culture and theself: Implications for cognition, emotion, and moti-vation. Psychological Review, 98, 224–253.

Masuda, T., & Nisbett, R. E. (2001). Attendingholistically vs. analytically: Comparing thecontext sensitivity of Japanese and Americans.Journal of Personality and Social Psychology, 81,911–934.

Mesquita, B. (2003). Emotions as dynamic culturalphenomena. In R. Davidson, H. Goldsmith, & K.R. Scherer (Eds.), The handbook of the affective

INTERPERSONALLY ENGAGING EMOTIONS 693

sciences (pp. 871–890). New York, NY: OxfordUniversity Press.

Mesquita, B., & Leu, J. (2007). The cultural psychologyof emotion. In S. Kitayama, & D. Cohen (Eds.), Thehandbook of cultural psychology (pp. 734–759).New York, NY: Guilford Press.

Oishi, S. (2002). The experiencing and remembering ofwell-being: A cross-cultural analysis. Personality andSocial Psychology Bulletin, 28, 1398–1406.

Oishi, S., & Diener, E. (2003). Culture and well-being:The cycle of action, evaluation, and decision.Personality and Social Psychology Bulletin, 29,939–949.

Oishi, S., Diener, E., Scollon, C., & Biswas-Diener, R.(2004). Cross-situational consistency of affectiveexperiences across cultures. Journal of Personalityand Social Psychology, 86, 460–472.

Oishi, S., Koo, M., & Akimoto, S. (2008). Culture,interpersonal perceptions, and happiness in socialinteractions. Personality and Social PsychologyBulletin, 34, 307–320.

Peng, K., Nisbett, R. E., & Wong, N. (1997). Validityproblems comparing values across cultures andpossible solutions. Psychological Methods, 2,329–344.

Ramirez-Esparza, N., Chung, C. K., Sierra-Otero, G., &Pennebaker, G. W. (2012). Cross-cultural construc-tions of self-schemas: Americans and Mexicans.Journal of Cross-Cultural Psychology, 43, 233–250.

Russell, J. A. (1983). Pancultural aspects of the humanconceptual organization of emotions. Journal ofPersonality and Social Psychology, 45(6), 1281–1288.

Sanchez-Burks, J., Bartel, C. A., & Blount, S. (2009).Performance in intercultural interactions atwork: Cross-cultural differences in response tobehavioral mirroring. Journal of AppliedPsychology, 94, 216–223.

Sanchez-Burks, J., Nisbett, R. E., & Ybarra, O. (2000).Cultural styles, relationship schemas, and prejudiceagainst out-groups. Journal of Personality and SocialPsychology, 79, 174–189.

Shweder, R. A., & Haidt, J. (2000). The culturalpsychology of the emotions: Ancient and new.In M. Lewis, & J. M. Haviland-Jones (Eds.),Handbook of emotions: Second edition(pp. 397–414). New York, NY: Guilford Press.

Stevenson, H., & Stigler, J. W. (1994). The learning gap:Why our schools are failing and what we can learnfrom Japanese and Chinese education. New York,NY: Simon & Schuster.

Suarez-Orozco, C., & Suarez-Orozco, M. (2001).Children of immigration. Cambridge, MA: HarvardUniversity Press.

Triandis, H. C., Marin, G., Lisansky, J., & Betancourt,H. (1984). Simpatia as a cultural script of Hispanics.Journal of Personality and Social Psychology, 47,1363–1375.

Trumbull, E., Rothstein-Fisch, C., & Greenfield, P. M.(1999). Bridging cultures in our schools: Newapproaches that work. San Francisco, CA: WestEd.

Tsai, J. L. (2007). Ideal affect: Cultural causes andbehavioral consequences. Perspectives inPsychological Science, 2, 242–259.

Valdes, G. (1996). Con respecto: Bridging the distancesbetween culturally diverse families and schools. NewYork, NY: Teachers College Press.

Winkielman, P., Berridge, K. C., & Wilbarger, J. L.(2005). Unconscious affective reactions to maskedhappy versus angry faces influence consumptionbehavior and judgments of value. Personality andSocial Psychology Bulletin, 31, 121–135.

Yik, M. S. M., & Russell, J. A. (2003). Chinese affectcircumplex: I. Structure of momentary affect. AsianJournal of Social Psychology, 6, 185–200.

694 SAVANI ET AL.