POSITIVE EMOTION CULTURE 995of positive and not of negative emotions. Miyamoto and colleagues role of positive emotions in depression when comparing European(2010) showed that Japanese participants reported more simulta- Americans with immigrant Asians, then differences between Eu-neous positive and negative emotions than European Americans ropean Americans and Asians (in Asia) were likely to be evenonly in positive and not in negative situations. Similarly, Leu and more significant.colleagues (2009) demonstrated that in positive, but not negative,situations the correlation between positive and negative emotions Present Studywas more negative among European Americans than among Chi-nese or Japanese. These provocative data suggest that in positive Methodssituations, Asians may “find the bad in the good” more thanEuropean Americans, whereas both Asians and European Ameri- Participants. We compared 330 European American, 156cans seem to “find the good in the bad” in negative situations. immigrant Asian, and 147 Asian American (N ϭ 633) collegeThese data further suggest greater cultural variation in the influ- students from a public university in the US who participated forence of positive, rather than negative, emotions on mental health. extra course credit. European American participants were at least third-generation; Asian Americans were all born in the US The Role of Acculturation to immigrant parents; and immigrant Asian participants came to the United States at an average age of 11 years (SE ϭ 0.47). The New evidence using bicultural samples suggests that cultural largest Asian ethnic groups were Chinese, Korean, and Viet-differences in emotion can also occur within a single national namese, although the sample also included South Asians.context. For example, Perunovic and her colleagues (Perunovic, Procedure. All participants completed an hour-long surveyHeller, & Rafaeli, 2008) found that Asian-Canadian college on a computer on measures of perceived stress, emotions, fre-students who were shown Canadian primes reported fewer quency of depression symptoms, and demographics.dialectical emotions than those who were shown Asian primes. MeasuresThere is also evidence that the early emotional socialization of Demographics. Socioeconomic status was measured usingAsian American biculturals may mirror the emotional practices the MacArthur Scale of Subjective Social Status, which capturesof immigrant Asians. Tsai and her colleagues demonstrated that relative social rank (Adler, Epel, Castallazzo, & Ickovics, 2000).young Asian American children preferred low-arousal positive Nativity was measured by participants’ country of birth (i.e., US-emotions (i.e., calmness) over high-arousal positive emotions or Asia-born); culture/ethnicity was self-identified.(i.e., excitement), compared with their European American Dependent variable. The Center for Epidemiologic Studiespeers (Tsai, Louie, Chen, and Uchida, 2007). These studies Depression Scale (CES-D; Radloff, 1977) examines subclinicalraise the interesting question of whether adult Asian Americans depression symptoms among general populations and was usedbehave emotionally like Asians in Asia or like their Western in other studies of positive emotion and depression (Fredrick-peers with regard to positive emotion expression. son et al., 2003). This measure consisted of 20 item statements focusing on frequency of depressed mood, feelings of worth- Implications for Mental Health lessness, and disturbance of appetite and sleep “in the past week,” using a scale from 1 (less than one day) to 4 (5–7 days) Thus, there is reason to predict that culture and acculturation (␣ ϭ 0.89).may influence the role of positive emotions in mental health. Are Independent variables. The Perceived Stress Scale (Cohen,positive emotions as protective among Asians as in Western sam- Kamarck, & Mermelstein, 1983) measured the global perception ofples? There is already evidence that the absence of positive self- stress “over the past month” from 1 (Never) to 5 (Very Often). Sampleesteem is less predictive of depression (Heine et al., 1999) and that items included, “In the last month, how often have you been upsetdialectical emotions are associated with fewer problematic physi- because of something that happened unexpectedly?” and “In thecal symptoms (Miyamoto & Ryff, in press) among Japanese than last month, how often have you felt on top of things?” (re-North Americans. However, this is the only study that directly versed) (␣ ϭ 0.86).examines the relationship between culture, positive emotions, and The Positive and Negative Emotions Schedule–X (PANAS–X;depression among European Americans and a diverse group of Watson, Clark, & Tellegen, 1998) measured how participants feelAsians. “on average” by rating the intensity of positive and negative emotions on a 5-point scale, where 1 ϭ “not at all,” 3 ϭ “mod- Predictions erately,” and 5 ϭ “extremely.” In addition, the measure examined specific positive emo- We predicted that (1) positive emotions would be significant tions such as serenity (i.e., “calm,” “relaxed,” and “at ease”;predictors of depression symptom frequency for European Amer- ␣ ϭ 0.76), joviality (i.e., “happy,” “joyful,” “delighted,”icans, but not for immigrant Asians. On the other hand, we “cheerful,” “excited,” “enthusiastic,” “lively,” and “ener-expected (2) negative emotions to be significantly associated with getic”); ␣ ϭ 0.93); self-assurance (i.e., “proud,” “strong,” “con-depression for all cultural groups. Lastly, we explored the question fident,” “bold,” “daring,” and “fearless”); ␣ ϭ 0.80), and atten-of whether US-born Asian American participants looked more like tiveness (i.e., “alert,” “attentive,” “concentrating,” “determined”,European Americans or foreign-born immigrant Asians. Using ␣ ϭ 0.71).immigrant Asian samples provided us with the additional advan- Importantly, positive emotion ratings were reliable when calcu-tage of serving as a conservative test of cultural variation. We lated only among immigrant Asians and Asian Americans (␣s ϭreasoned that if we could demonstrate cultural differences in the 0.72 to 0.79). The two-factor structure of positive and negative
996 LEU, WANG, AND KOOemotions was replicated among all of our samples, consistent with Culture Moderates the Influence of Positive Emotionspast studies including diverse samples (Watson et al., 1988; on DepressionThompson, 2007). We used a generalized linear regression model to test our hypothesis that culture moderates the effect of positive, but not Results negative, emotions on depression. Specifically, we simultaneously regressed positive emotions, negative emotions, the interaction ofDescriptives culture and emotion variables, and demographics onto the fre- quency of depression symptoms. Increases in negative emotions One-way analyses of variance (ANOVAs) with post hoc Bon- (b ϭ 0.44, SE ϭ 0.05, p Ͻ .001) and decreases in positiveferroni corrections revealed that immigrant Asians (M ϭ 0.86, emotions (b ϭ Ϫ0.33, SE ϭ 0.05, p Ͻ .001) were associated withSE ϭ .04) and Asian Americans (M ϭ 0.72, SE ϭ .04) reported greater depression symptoms. As predicted, an interaction betweenmore frequent depression symptoms than European Americans positive emotions and culture was significant (b ϭ 0.001, SE ϭ(M ϭ 0.63, SE ϭ .03), F(2, 629) ϭ 10.94, p Ͻ .001. Consistent 0.001, p Ͻ .01), whereas there was no significant interactionwith past studies, European Americans (M ϭ 3.59, SE ϭ .03) between negative emotions and culture (b ϭ 0.0001, SE ϭ 0.001,reported greater positive emotion intensity than immigrant Asians p Ͼ .05). To understand the significant culture x positive emotion(M ϭ 3.29, SE ϭ .05) or Asian Americans (M ϭ 3.30, SE ϭ .05), interaction, regression outputs were split by cultural group. AsF(2, 630) ϭ 17.01, p Ͻ .001. Except for depression symptoms, for seen in Table 1, positive emotions were negatively associated withwhich the means were relatively close to the low end of the scale, frequency of depression symptoms among European Americansnone of the variables suffered from ceiling or floor effects. and Asian Americans, but not among immigrant Asians. As seen in Figure 1, the intensity of positive emotion and Does culture moderate positive emotion mediation? Tofrequency of depression symptoms were negatively correlated compare our data with other models of stress and depression thatamong European Americans (r ϭ Ϫ0.46, p Ͻ .001) and Asian have demonstrated positive emotion mediation (Fredrickson et al.,Americans (r ϭ Ϫ0.26, p Ͻ .01), but not among immigrant Asians 2003; Kiecolt-Glaser et al., 2002), we assumed a causal relation-(r ϭ Ϫ0.03, p Ͼ .05). Negative emotion intensity and depression ship between stress and emotions and between emotions andwere positively correlated for all groups (rEA ϭ 0.57, p Ͻ .001; depression in our data despite the obvious limitations of usingrIA ϭ 0.43, p Ͻ .001; rAA ϭ 0.52, p Ͻ .001). Evidence of a cross-sectional data. We found that positive emotion mediateddialectical relationship between positive and negative emotions stress and frequency of depression symptoms only among Euro-was found among immigrant Asians but not Asian Americans pean Americans and not immigrant Asians. On the other hand, we(rIA ϭ 0.24, p Ͻ .01; rAA ϭ 0.11, p Ͼ .05). found that negative emotions mediated stress and frequency of European Americans Asian Americans Immigrant Asians Figure 1. Scatter plots of the correlation between positive emotions and frequency of depression symptoms among European Americans (r ϭ Ϫ.46 ,)ءءءAsian Americans (r ϭ Ϫ.26 ,)ءءand Immigrant Asians (r ϭ Ϫ.03) ( ءp Ͻ .05, ءءp Ͻ .01, ءءءp Ͻ .001).
POSITIVE EMOTION CULTURE 997Table 1 Exploring an Alternative ExplanationRegression of Demographics, Positive, and Negative Emotions Tsai (2007) suggests that low-arousal positive emotions (e.g.,on Mean Frequency of Depression Symptoms serenity) are preferred by East Asians, whereas high-arousal pos- Model 1 Model 2 Model 3 itive emotions (e.g., joviality, self-assurance, attentiveness, and surprise) are preferred by European Americans. Therefore, by European Asian Immigrant using a summary positive affect variable that conflated both high- American American Asian and low-arousal positive emotions, we may have overlooked the Variables B (SE) B (SE) B (SE) possibility that low-arousal positive emotions predict depression among our Asian samples. ءIntercept 0.66 (0.41) 1.05 (0.43) 0.27 (0.55) To test this alternative hypothesis, we reran the original regres-Sex 0.04 (0.04) 0.001 (0.06) Ϫ0.07 (0.09) sion analysis, replacing the summary positive affect variable withAge Ϫ0.01 (0.02) Ϫ0.02 (0.02) 0.01 (0.02)SES 0.03ء (0.01) Ϫ0.01 (0.02) 0.02 (0.03) high- or low-arousal positive affect subscales in separate models.Negative Emotion 0.40ءءء (0.04) 0.45ءءء (0.05) 0.35ءءء (0.06) We simultaneously regressed high-/low-arousal positive emotions,Positive Emotion Ϫ0.26ءءء (0.04) Ϫ0.22ءءء (0.05) Ϫ0.10 (0.06) negative emotions, the interaction of culture and emotion vari-ء ءءء ables, and demographics onto the frequency of depression symp- p Ͻ .05. p Ͻ .001. toms. We found a significant interaction effect between culture and the high-arousal subscale (b ϭ 0.001, SE ϭ 0.001, p Ͻ .01) ondepression symptoms for all groups. Although not our main pre- depression symptoms. However, there was no interaction betweendiction, these analyses (reported in detail below) complement the culture and the low-arousal subscale (serenity). For both analyses,regression analyses reported in Table 1 in suggesting that culture there was no interaction between culture and negative affect.may moderate the role of positive, but not negative, emotions in Therefore, inconsistent with ideal affect theory, there was nodepression symptom expression. evidence that culture moderated the association between low- Using a common four-step criteria test for mediation (Baron & arousal positive emotions and depression symptoms. ConsistentKenny, 1986), we found evidence that positive emotions mediated with ideal affect theory, however, when we split the regressionthe relationship between perceived stress and depression symp- outputs for the high-arousal subscale by cultural group, we foundtoms only among European Americans and not among either Asian evidence that increases in high-arousal positive emotions (i.e.,group. For example, among European Americans, perceived stress joviality, self-assurance, attentiveness, and surprise) were as-was correlated with positive emotions (b ϭ Ϫ0.31, SE ϭ .05, ␤ ϭ sociated with decreases in depression symptoms among Euro-Ϫ0.33, p Ͻ .001). Perceived stress was then correlated with pean Americans (b ϭ Ϫ0.31, SE ϭ 0.04, p Ͻ .001), but notfrequency of depression symptoms (b ϭ 0.48, SE ϭ .03, ␤ ϭ 0.63, immigrant Asians (b ϭ Ϫ0.12, SE ϭ 0.04, p Ͼ .05). Interest-p Ͻ .001). Positive emotions (b ϭ Ϫ0.23, SE ϭ 0.04, ␤ ϭ Ϫ0.28, ingly, Asian Americans looked more similar to European Amer-p Ͻ .001) were significantly correlated with frequency of depres- icans than immigrant Asians (b ϭ Ϫ0.25, SE ϭ 0.05, p Ͻ .001).sion symptoms when controlling for perceived stress. The Sobeltest was significant (Z ϭ 4.55, p Ͻ .001), indicating partialmediation by positive emotions. For immigrant Asians, although General Discussionperceived stress and frequency of depression symptoms were In a large survey of college students, we found evidence thatcorrelated (bIA ϭ 0.52, SE ϭ 0.06, ␤ ϭ 0.59, p Ͻ .001), perceived culture may moderate the role of positive emotions on depression.stress and positive emotions were not significantly correlated We used two kinds of analyses to demonstrate this. First, we(bIA ϭ Ϫ0.12, SE ϭ 0.09, ␤ ϭ Ϫ0.10, p Ͼ .05) and mediation was showed that increased positive emotions were directly related tonot run. For Asian Americans, perceived stress and frequency of decreases in depression symptoms among European Americans,depression symptoms were correlated (bAA ϭ 0.50, SE ϭ 0.04, but not among immigrant Asians. Also as predicted, increased␤ ϭ 0.69, p Ͻ .001). Positive emotions were not significantly negative emotions were associated with increases in depressioncorrelated with frequency of depressive symptoms when control- symptoms across all groups. These findings are consistent withling for perceived stress (b ϭ Ϫ0.002, SE ϭ 0.05, ␤ ϭ Ϫ0.003, research on dialectical emotions, where cultural differences arep Ͼ .05). greater in positive than negative situations. Also as predicted, we found evidence that negative emotions Next, we tested whether positive feelings explained the influ-mediated the relationship between perceived stress and depression ence of stress on depression among both European Americans andsymptoms among all three cultural groups. Perceived stress was Asians. There are limitations of assuming causality in data thatcorrelated with negative emotions (bEA ϭ 0.59, SE ϭ .04, ␤ ϭ only measures one point in time, so we need be cautious of0.65, p Ͻ .001; bIA ϭ 0.62, SE ϭ .08, ␤ ϭ 0.53, p Ͻ .001; bAA ϭ generalizing this evidence. Still, we found more evidence that0.45, SE ϭ .07, ␤ ϭ 0.48, p Ͻ .001). Perceived stress and culture influences the role of positive feelings in depression. Thefrequency of depression symptoms were significantly correlated influence of stress on depression was partially explained by the(see positive emotion mediation results). Negative emotions intensity of positive feelings for only European Americans. How-(bEA ϭ 0.23, SE ϭ 0.05, ␤ ϭ 0.27, p Ͻ .001; bIA ϭ 0.13, SE ϭ ever, negative emotions partially explained the effect of stress on0.06, ␤ ϭ 0.17, p Ͻ .05; bAA ϭ 0.19, SE ϭ 0.05, ␤ ϭ 0.25, p Ͻ depression for all groups. Altogether, these findings suggest that.001) were significantly correlated with frequency of depression culture may moderate the role that positive emotions play insymptoms when controlling for perceived stress. A significant mental health.Sobel test (ZEA ϭ 4.82, p Ͻ .001; ZIA ϭ 2.08, p Ͻ .05; ZAA ϭ 3.26, One may wonder if our findings are the result of positivep Ͻ .001) indicated partial mediation by negative emotions. emotions being correlated with negative emotions, such that the
998 LEU, WANG, AND KOOassociation between positive emotions and depression symptoms is sality. We used subclinical depression symptoms, which havean “epiphenomon” of the correlation between negative emotion previously been used to demonstrate the protective role of positiveand depression. However, the regression analyses (see Table 1) emotions among Western samples (Fredrickson et al., 2003). Pre-show that this is not the case. We also know that the positive sumably, cultural differences among clinical samples may be evenemotion ratings from both Asian samples were reliable, so the more pronounced, although a clinical sample is ultimately needed.validity of the PANAS–X scale is not the reason for failing to find Despite these limitations, our findings raise the question ofevidence for the effects of positive emotions among Asians. We whether positive psychology interventions (i.e., optimistic think-further ruled out the possibility that low-arousal positive emotions ing or replaying positive experiences) which typically alleviatewould be significantly predictive of depression among immigrant depression symptoms for Westerners (Sin & Lyubomirsky, 2009),Asians but not European Americans (Tsai, 2007). will be similarly effective for Asians to practice. Instead, therapies We suggest that one reason why positive emotions may not be which encourage individuals to embrace both positive and nega-as “positive” for Asians as for European Americans is because of tive emotions may be more effective with Asian clients (e.g.,cultural differences in the meaning assigned to positive, but not Hayes, Strosahl, & Wilson, 1999). Future longitudinal data andnegative, emotions. At least two experimental studies (Leu et al., experimental studies with diverse clinical samples will contribute2009; Miyamoto et al., 2010) have previously shown that East- to a needed body of research that addresses how culture influencesWest differences in emotion reports are greatest in positive, and the “good” of positive emotions, and more broadly, the role ofnot negative, situations. While maximizing positive emotions may culture and positive psychology in mental health.be a cultural goal in Western contexts, emotion moderationthrough balancing positive emotions with negative ones may be a Referencescultural goal in Asian contexts. Adler, N. E., Epel, E., Castallazzo, G., & Ickovics, J. (2000). Relationship of subjective and objective social status with psychological and physi-Significance of Acculturation ological functioning in preliminary data in healthy white women. Health Psychology, 19, 586 –592. The inclusion of two Asian samples (American-born and immi- Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variablegrant) was important to explore whether acculturation (i.e., nativ- distinction in social psychological research: Conceptual, strategic, andity) influences the role of positive emotions in depression. In the statistical considerations. Journal of Personality and Social Psychology,current study, there was some evidence that Asian Americans were 51, 1173–1182.similar to both European Americans and immigrant Asians, de- Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure ofpending on the analysis. Greater positive emotions were associated perceived stress. Journal of Health and Social Behavior, 24, 385–396.with fewer depression symptoms among Asian Americans, similar Fredrickson, B. L., & Cohn, M. A. (2008). Positive emotions. In M. Lewis,to European Americans. Also, the correlation between positive J. M. Haviland-Jones, & L. Feldman Barrett (Eds.), Handbook of emo-emotions and depression was significantly negative among Asian tions (pp. 777–796). New York: Guilford Press.Americans, although the size of the correlation was smaller for Fredrickson, B. L., Tugade, M. M., Waugh, C. E., & Larkin, G. R. (2003). What good are positive emotions in crisis? A prospective study ofAsian Americans than for European Americans. However, there resilience and emotions following the terrorist attacks on the Unitedwas a lack of evidence for positive emotion mediation of stress and States on September 11th, 2001. Journal of Personality and Socialdepression among Asian Americans, similar to immigrant Asians. Psychology, 84, 365–376.Further studies need to be conducted to better understand the role Hayes, S. C., Strosahl, K., & Wilson, K. G. (1999). Acceptance andof acculturation in the emotional behavior of biculturals. commitment therapy: An experiential approach to behavior change. New York: Guilford Press. Heine, S. J., Lehman, D. R., Markus, H. R., & Kitayama, S. (1999). Is therePractical Significance, Limitations, and Future a universal need for positive self-regard? Psychological Review, 106,Directions 766 –794. This study directly examined the relationship between culture, Kiecolt-Glaser, J., McGuire, L., Robles, T., & Glaser, R. (2002). Emotions, morbidity, and mortality: New perspectives from psychoneuroimmunol-positive emotions, and depression among European Americans and ogy. Annual Review of Psychology, 53, 83–107.a diverse group of Asians. Our immigrant Asian sample suggests Kitayama, S., Markus, H. R., & Kurokawa, M. (2000). Culture, emotion,that positive emotions may play a limited role in the mental health and well-being: Good feelings in Japan and the United States. Cognitionof Asians in Asia compared with Western samples. With Asians and emotion, 14, 93–124.comprising more than 60% of the global population, these findings Leu, J., Mesquita, B., Ellsworth, P. C., Zhang, Z., Huijuan, Y., Buchtel, E.,suggest the need to proceed with some caution in our claims about Karasawa, M., & Masuda, T. (2009). Situational differences in dialec-the benefits of positive emotions for everyone. Incorporating more tical emotions: Boundary conditions in a cultural comparison of Northculturally diverse samples in our research on positive emotions not Americans and East Asians. Cognition & Emotion, 24, 419 – 435.only addresses concerns about generalizability, but may potentially Lyubomirsky, S., King, L. A., & Diener, E. (2005). The benefits ofreveal the cultural and structural nature of positive emotions them- frequent positive affect: Does happiness lead to success? Psychological Bulletin, 131, 803– 855.selves. For example, these findings add to the literature suggesting Mesquita, B., & Karasawa, M. (2002). Different emotional lives. Cognitionthat positive and negative emotions are not simply bipolar oppo- and Emotion, 16, 127–141.sites and often have different predictors and consequences. Miyamoto, Y., & Ryff, C. (in press). Cultural differences in the dialectical Our data were cross-sectional, limiting our ability to interpret and non-dialectical emotional styles and their implications for health.the meditational analyses. Future studies should include a longi- Cognition and Emotion.tudinal design to gain greater confidence in the direction of cau- Miyamoto, Y., Uchida, Y., & Ellsworth, P. C. (2010). Culture and mixed
POSITIVE EMOTION CULTURE 999 emotions: Co-occurrence of positive and negative emotions in Japan and psychological perspective on mental-health. Psychological Bulletin, the United States. Emotion, 10, 404 – 415. 103, 193–210.Perunovic, W. Q. E., Heller, D., & Rafaeli, E. (2007). Within-person Thompson, E. R. (2007). Development of an internationally reliable short- changes in the structure of emotion: The role of cultural identification form of the Positive and Negative Affect Schedule (PANAS). Journal of and language. Psychological Science, 18, 607– 613. Cross-Cultural Psychology, 38, 227–242.Radloff, L. S. (1977). The CES–D scale: A self report depression scale for Tsai, J. L. (2007). Ideal affect: Cultural causes and behavioral conse- research in the general population. Applied Psychological Measurement, quences. Perspectives on Psychological Science, 2, 242–259. 1, 385– 401. Tsai, J. L., Louie, J. Y., Chen, E. E., & Uchida, Y. (2007). Learning whatRichman, L. S., Kubzansky, L. K., Maselko, J., Kawachi, I., Choo, P., & feelings to desire: Socialization of ideal affect through children’s story- Bauer, M. (2005). Positive emotion and health: Going beyond the books. Personality and Social Psychology Bulletin, 33, 17–30. negative. Health Psychology, 24, 422– 429. Uchida, Y., & Kitayama, S. (2009). Happiness and unhappiness in east andSchimmack, U., Oishi, S., & Diener, E. (2002). Cultural influences on the west: Themes and variations. Emotion, 9, 441– 456. relation between pleasant emotions and unpleasant emotions: Asian Uchida, Y., Norasakkunkit, V., & Kitayama, S. (2004). Cultural construc- dialectic philosophies or individualism-collectivism? Cognition and tions of happiness: Theory and empirical evidence. Journal of Happiness Emotion, 16, 705–719. Studies, 5, 223–239.Sin, N. L., & Lyubomirsky, S. (2009). Enhancing well-being and allevi- Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and ating depressive symptoms with positive psychology interventions: A validation of brief measures of Positive and Negative Affect: The practice-friendly meta-analysis Journal of Clinical Psychology: In Ses- PANAS scales. Journal of Personality and Social Psychology, 54, sion, 65, 467– 487. 1063–1070.Spencer-Rodgers, J., Williams, M. J., & Peng, K. (2010). Cultural differ- ences in expectations of change and tolerance for contradiction: A decade of empirical research. Personality and Social Psychology Re- Received April 5, 2010 view, 14, 296 –312. Revision received June 18, 2010Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being—A social Accepted June 28, 2010 Ⅲ