Contextual and Cultural Differences in Positive Thinking

Positive thinking has long fascinated the public (e.g., Peale, 1952^; Smith, 2013). This makes sense given that thinking positively has been associated with a variety of favorable life outcomes (Gallagher et al., 2013^; Scheier et al., 1989^, 2001). Past research often compares positive thinking or optimism across cultures while focusing on culture main effects (Chang, 1996^; Hardin & Leong, 2005^; Lee & Seligman, 1997^; You et al., 2009), with minimal attention on the influence of context. To fill the gap, this paper examines how contextual factors (i.e., positive vs. negative events) may moderate cross-cultural differences in positive thinking.

We define positive thinking as the tendency to have a preponderance of positive (pleasant, favorable) thoughts relative to negative (unpleasant, unfavorable) thoughts. Positive thinking may emerge because people consider both positive and negative thoughts, but their thought processes favor positive over negative thoughts (e.g., having but dismissing negative thoughts), or because people generate more positive than negative thoughts in the first place. Similarly, positive thinking can reference any subject or timeframe. For example, optimism and pessimism may be considered specific instances of positive (versus negative) thinking that reference oneself in the future (Bandura, 1986^; Dember et al., 1989^; Scheier & Carver, 1985). Similarly, Taylor and Brown’s (1988^, 1994) work specifically addresses overly positive self-relevant thoughts (“positive illusions”). However, positive thinking is just a preponderance of thoughts of positive versus negative valence, and the construct need not be constrained to first-person, future-oriented thoughts. Positive thinking can also be captured in thoughts about other people or about outcomes very broadly (e.g., thoughts about the fate of humanity) and thoughts about the past or the present. In this work, we focus on future timeframes because the future is inherently ambiguous, creating a lot of variance in responses. However, our studies consider both thoughts about the self and impersonal contexts to ensure generalizability of our claims.

Cultural Variation in Positive Thinking

Broadly speaking, most research has indicated that people from Western cultures strongly prefer positivity in many aspects of their experiences, whereas people from Eastern cultures have a more mixed set of preferences. Most research treats these preferences or tendencies for increased Western positivity as fixed (dispositional) differences in underlying traits. For example, compared to Americans, Chinese are less optimistic (but see Chang, 1996^; You et al., 2009), and/or more pessimistic (Chang, 1996^; Hardin & Leong, 2005^; Lee & Seligman, 1997). Relatedly, Japanese show less unrealistic optimism than Canadians (Heine & Lehman, 1995). Euro-Americans and Euro-Canadians (i.e., North Americans of European ancestry) expend more effort than Asian-Americans and Asian-Canadians (i.e., North Americans of Asian ancestry) to experience positive feedback and enjoyment and avoid tasks that give negative feedback (Falk et al., 2010^; Oishi & Diener, 2003). Furthermore, Asian-Canadians focus more on negative social comparison information (i.e., negative role models); whereas European-Canadians focus primarily on positive social comparison targets (Lockwood et al., 2005). Finally, life satisfaction is less closely linked to a dense association of positive constructs among Koreans and Asian-Americans versus Euro-Americans (Koo & Oishi, 2009). Thus, East Asians may have weaker inclinations for positive thinking relative to Western people.

Various explanations have been used to explain these cultural differences. A highly influential theory implicates self-enhancement. Eastern individuals, motivated to self-improve, may show less positive thinking at least about themselves; Western individuals are motivated to self-enhance, thus they tend to think more positively about themselves (Gierlach et al., 2010^; Heine et al, 2001). In addition, cultural differences in values such as uncertainty avoidance, individualism, and/or power distance (Hofstede, 1980), may account for cultural variance in positive thinking (Chang, 1996^; Fischer & Chalmers, 2008^; Jacobson et al., 2012). For example, Chinese people might engage in less positive thinking due to their higher uncertainty avoidance: “by anticipating the worst, some Asians may paradoxically gain a sense of control in their uncertainty about future outcomes” (Chang, 1996, p. 821). These perspectives suggest stable cultural differences in positive thinking by pointing to cultural values that should remain intact regardless of context. However, a purely “fixed cultural difference” perspective has some issues. For instance, one recent meta-analysis (Fischer & Chalmers, 2008) offers no support for overall cultural differences in optimism among mainland Chinese, Canadians, and Americans. This null effect suggests that cultural differences in positive thinking may be variable across context.

Relatedly, most research tends to characterize optimism, which we classify as a specific type of positive thinking, as a fixed trait (Abramson et al., 1978^; Carver & Scheier, 2009^; Peterson & Seligman, 1984^; Scheier et al., 2001). However, optimism can be situationally determined. For example, people may be less optimistic when anticipating feedback (vs. not), when outcomes are important (vs. unimportant), when negative outcomes are more accessible (vs. less accessible), when outcomes are uncontrollable (vs. controllable), and when they have low (vs. high) self-esteem (Sweeny et al., 2006). Positive thinking more generally may also be situationally determined.

Fluctuating contextual factors may be responsible for the inconsistency in previous findings on optimism. Although optimism has been treated as a trait in most research, it can change over time in response to novel information (but see Carver & Scheier, 2014^; Malouff & Schutte, 2017^; Shifren & Hooker, 1995). This opens the door to situational pressures being able to shift how positively people think about the future. We propose that seemingly cross-cultural differences in positive thinking—often assumed to be fixed—may actually appear, disappear, or reverse depending on situational factors. In the present research, we examine one such contextual variable: the valence of the context or events to which people are responding.

Culture, Cognition, and Positive Thinking in Context

Culture shapes people’s thinking style and beliefs (Nisbett, 2003). Nisbett et al. (2001) have demonstrated a tendency for different thinking styles between East Asians and Westerners (i.e., people of European descents). Specifically, Westerners tend to think analytically, attending primarily to an object and the categories it belongs to, and using rules including formal logic to understand its behaviors. In contrast, East Asians tend to think holistically, attending to the field and context, making less use of categories and formal logic, and relying on dialectical reasoning (i.e., tolerating and accepting contradictions; Peng & Nisbett, 1999). For instance, Chinese are more likely to believe in a form of dialecticism whereby opposing statements are resolved by finding a “middle way” or accepting contradictions. In contrast, Americans think in an Aristotelian manner: when information conflicts, Americans seek to determine which is true and which is false, henceforth endorsing only the true proposition (Peng & Nisbett, 1999).

Interestingly, these beliefs can help account for cross-cultural differences. For example, Spencer-Rodgers et al. (2004) found that higher ambivalence in the experience of positive and negative self-relevant thoughts (among Asian-Americans versus European-Americans) were mediated by “naive dialecticism in the domain of self-perception” (p. 1423). Similarly, Spencer-Rodgers et al. (2010) showed that higher rates of simultaneous positive and negative emotion (among Chinese versus Americans) depended on dialectical thinking.1 Underlying lay beliefs may also help to explain cultural differences in positive and negative thinking.

One important principle underlying Chinese dialectical thinking is their belief in change (Peng & Nisbett, 1999). Chinese tend to hold stronger beliefs in change, especially cyclical change, compared to North Americans (Ji, 2005^; Ji et al., 2001). If events and people are constantly changing, then contradiction becomes inevitable. Thus, believing in constant change may lead people to accept contradictions. Ji et al. (2001) have shown that Chinese and Americans hold different lay theories of change—beliefs about how events develop over time (Ji, 2005; also see Ji et al., 2008^; 2009). Chinese are more likely than North Americans to believe that the world is constantly and cyclically changing (e.g., what goes up will come down, and what comes down will go up), with this gap developing as early as middle childhood (Ji, 2008). For example, Chinese participants were more likely than North Americans to expect two people fighting currently to become lovers one day and to expect a dating couple to break up after graduating from university (Ji et al., 2001). Likewise, Chinese were more likely to view personal happiness and unhappiness to transform into each other (because life is expected to change), whereas Americans tended to see happiness as something stable (because life is expected not to change; Ji et al., 2001). Interestingly, Koreans are also more likely than Americans to anticipate changes in future events, consistent with the prediction that holistic (vs. analytic) cultures will be more likely to anticipate change (Choi et al., 2007).

Cultural differences in lay theories of change, therefore, suggest that Chinese are more likely than North Americans to expect good and bad events to convert into each other. Due to such differences in lay theories of change, one would expect cultural differences in positive thinking, at least about the future, to vary depending on the context. In response to positive events, Chinese may have more negative (less positive) thoughts about the future than do North Americans; in response to negative events, Chinese may have more positive (less negative) thoughts than do North Americans.

Existing evidence partially backs up this reasoning. In response to real life negative events such as the SARS outbreak or the COVID-19 pandemic, Chinese participants were more likely than Euro-Canadians to mention positive changes brought to their life and the world by the negative event (Ji et al., 2004^; 2020). These findings appear at odds with some previous findings that Chinese engage in less positive thinking than North Americans. We argue that the context may explain the contrasting findings. In response to positive events, Chinese may engage in less positive thinking (i.e., will expect bad outcomes to arise from good events) than North Americans do. But in response to negative events (like SARS), Chinese may show more positive thinking (i.e., will expect good outcomes to arise from bad events) than North Americans do.

The Present Work

The present research focuses on a contextual factor (i.e., positive vs. negative events) as a moderator of cross-cultural differences in positive thinking. As we derive our predictions based on past research involving East Asian (including Chinese, Japanese and Koreans) and European Americans or Canadian samples, the present research compares an East Asian group (i.e., Chinese) with Euro-Canadians. Based on previous research on cultural differences in dialectical thinking and especially lay theories of change, we predicted that cultural differences in endorsing positive thinking should be qualified by context. Specifically, compared to Euro-Canadians, Chinese participants should endorse less positive thinking about the future in positive contexts and more positive thinking in negative contexts. Furthermore, such an effect may be accounted for by Chinese people’s (vs. Euro-Canadians’) greater belief in change.

We conducted three studies to test the predictions. In each study, we provided Chinese and Euro-Canadian participants with either positive and/or negative hypothetical events and then assessed the positivity of their thinking. In all studies, pilot testing determined that both cultural groups recognized the positive (negative) events as positive (negative), supporting that differences here indicate distinctions in the groups’ thinking processes about events rather than initial judgment of events themselves. Study 1 presented participants with both positive and negative events, in response to which participants rated the perceived likelihood of given positive and negative thinking responses. Study 2 replicated Study 1 without using rating scales. Study 3 asked participants to generate their own outcomes for positive and negative events, another approach to measure positive and negative thinking. In addition, Study 3 investigated the underlying mechanism for cultural differences in positive (or negative) thinking in context, namely, lay theories of change.

Study 1

In Study 1, participants read about several events, half of which were positive, and the other half negative. Participants indicated their endorsements of responses reflecting positive and negative thinking. The study had a 2 (Culture: Chinese vs. Euro-Canadian) × 2 (Context: positive vs. negative event) × 2 (Response: positive vs. negative thinking) mixed design, with the latter two variables as within-participant factors. We hypothesized that Chinese, compared to Euro-Canadians, would show less positive thinking in response to positive events but more positive thinking in response to negative events.

Results and Discussion

We conducted a 2 (Culture: Chinese vs Euro-Canadian) × 2 (Event Valence: positive vs. negative) × 2 (Response: positive vs. negative thinking) ANOVA with the latter two variables as within-participant variables. The key three-way interaction was significant, F(1, 285) = 34.85, p < .001, η_p² = 0.11. As shown in Figure 1, for positive events, the two-way interaction between culture and response was significant, F(1, 285) = 11.24, p = .001, η_p² = 0.04. Chinese endorsed less positive thinking, F(1, 285) = 4.92, p = .027, η_p² = 0.02, and more negative thinking, F(1, 285) = 10.55, p < .001, η_p² = 0.04, than did Euro-Canadians.

Open multimedia modal

The opposite pattern was found for negative events, where once again the two-way interaction between culture and response was significant, F(1, 285) = 18.53, p < .001, η_p² = 0.06. When given negative events, Chinese endorsed positive thinking more, F(1, 285) = 12.64, p < .001, η_p² = 0.042, and negative thinking less, F(1, 285) = 15.08, p < .001, η_p² = 0.050, than Euro-Canadians did. Thus, supporting our hypothesis, compared to Euro-Canadians, Chinese showed less positive thinking given positive events and more positive thinking given negative events. We also found main effects of event valence and response, and two-way interactions between event valence and culture as well as between event valence and response, Fs(1, 285) > 3.99, ps < 0.05.

Study 2

In Study 2, we conceptually replicated Study 1. Instead of using continuous rating scales, participants made binary choices reflective of positive or negative thinking. This choice approach minimizes concerns regarding potential cultural differences in extreme response bias (e.g., Chen et al., 1995). For example, some readers might wonder if the effects in Figure 1 were driven by Chinese endorsing less extreme responses (but see Ji et al., 2000). Such a criticism would not explain a binary-judgment in which gradations of “extreme” response do not exist. The study followed a 2 (Culture) × 2 (Context: positive vs. negative event) between-participant design.

Results

Positive thinking

For our primary analysis, we conducted binomial logistic regression in R (R Core Team, 2013). Choice of negative thoughts (scored 0) versus positive thoughts (scored 1), the dependent variable, was regressed onto Culture (Chinese = +0.5, Euro-Canadian = −0.5), Event Valence (Positive = +0.5, Negative = −0.5), and their interaction term. The critical interaction of Culture X Event Valence was significant, B = −2.02, Z = −2.73, p = .006, OR = 0.13 [0.03, 0.55], as depicted in Figure 2. In the positive event condition, fewer Chinese (M = 73.3%, SD = 44.6%) than Euro-Canadians (M = 85.7%, SD = 35.3%) endorsed positive thinking, although this difference was non-significant, B = −0.78, Z = −1.68, p = .092, OR = 0.46 [0.18, 1.12]. In the negative event condition, significantly more Chinese (M = 25.0%, SD = 43.7%) endorsed positive thinking than did Euro-Canadians (M = 9.7%, SD = 29.8%), B = 1.14, Z = 2.17, p = .030, OR = 3.11 [1.16, 9.32]. Critically, however, the interaction effect means that the group differences differed significantly as a function of event type (positive versus negative). Additionally, a simple effect of condition, B = 3.07, Z = 8.78, p < .001, OR = 21.49, CI_95% = [11.14, 44.12] revealed, unsurprisingly, that more positive thoughts were chosen in the positive (M = 79.5%, SD = 38.9%) than in the negative (M = 17.3%, SD = 38.9%) condition.5

Open multimedia modal

Study 3

In Study 3, we moved to an open thought-listing/rating task whereby Chinese and Euro-Canadian participants generated their own ideas about how certain events would unfold. We predicted that Chinese would be more likely than Euro-Canadians to generate negative outcomes for positive events and to generate positive outcomes for negative events. Furthermore, we hypothesized that these differences would be accounted for by lay theories of change. The study had a 2 (Culture) × 2 (Context: positive vs. negative event) design, with Context as a within participant factor.

Results and Discussion

Positive thinking: Proportion of positive thoughts

On average, Chinese (M = 4.46, SD = 1.71) generated fewer outcomes per event than Euro-Canadians did (M = 5.72, SD = 1.88), t(571) = −8.39, p < .001, d = 0.71 [0.54, 0.88]. Therefore, for each event, we created a proportional measure by dividing the number of all positive outcomes listed (i.e., outcomes rated +1 to +5) by the sum of all negative (i.e., outcomes rated −1 to −5) and positive outcomes combined. We then averaged this proportion index across the three positive events and across the three negative events, respectively. On this proportion measure (ranging between 0 and 1), scores over 0.5 indicate positive thinking (generating a greater number of positive than negative outcomes), and numbers under 0.5 indicate negative thinking (generating a greater number of negative than positive outcomes).

A 2 (Culture: Chinese vs Euro-Canadian) × 2 (Event Valence: positive or negative) mixed ANOVA on positive thinking revealed a significant interaction, F(1, 569) = 261.50, p < .001, η_p² = 0.32. As Figure 3a shows, for positive events, although both cultural groups showed significant positive thinking, Chinese thought less positively (M = 0.72, SD = 0.20) than did Euro-Canadians (M = 0.80, SD = 0.17), M_diff = −0.08 [−0.11, −0.05], p < .001. In contrast, for negative events, Chinese thought more positively (M = 0.52, SD = 0.18) than did Euro-Canadians (M = 0.25, SD = 0.19), M_diff = 0.27 [0.24, 0.30], p < .001. These results support our hypothesis. This analysis also revealed a significant main effect of culture, F(1, 569) = 68.59, p < .001, η_p² = 0.11 and a significant main effect of event valence, F(1, 569) = 1197.94, p < .001, η_p² = .68.

Open multimedia modal

Positive thinking: Valence of thoughts

We predicted that the outcomes associated with positive events should be less positive for Chinese than for Euro-Canadians, and the outcomes associated with negative events should be less negative for Chinese than for Euro-Canadians. A 2 (Culture) × 2 (Event Valence) mixed ANOVA on participants’ ratings of the self-generated outcomes, averaged across events, revealed a significant interaction, F(1, 571) = 308.90, p < .001, η_p² = 0.351.

As Figure 3b reveals, Chinese thought less positively (M = 1.71, SD = 1.47) than did Euro-Canadians (M = 2.53, SD = 1.27) for positive events, M_diff = −0.82 [−1.05, −0.59], p < .001. Furthermore, Chinese thought less negatively (M = 0.07, SD = 1.27) than did Euro-Canadians (M = −1.88, SD = 1.39) for negative events, M_diff = 1.94 [1.72, 2.16], p = .001. This pattern conceptually replicates the previous analysis, as well as Studies 1-2, but substitutes a free thought-listing task. Thus, cross-cultural differences emerged when positivity/negativity were not even mentioned until after thoughts were already generated. In addition, we identified a significant main effect of culture, F(1, 214) = 6.92, p = .009, η_p² = 0.03, and a significant main effect of event valence, F(1, 214) = 742.71, p < .001, η_p² = 0.67.

Open multimedia modal

Mediation: Lay theory of change (LTC)

We established the LTC’s factor structure by separately examining Chinese and Euro-Canadian data with respect to scree plots, fit indices (SRMR, RMSEA), and interpretability of resulting solutions, detailed in Supplementary-2.7 Chinese endorsed greater lay theories of change (M = 2.83, SD = 0.64, α = 0.62) than did Euro-Canadians (M = 1.80, SD = 0.69, α = 0.72), t(570) = 18.42, p < .0001, d = 1.56 [1.37, 1.75]. To investigate if lay theories of change accounted for cultural differences in positive thinking, we conducted mediation analyses. Although mediation analysis cannot prove a causal order given our partially cross-sectional design (MacKinnon et al., 2000), it can reveal if the data pattern is consistent with a mechanistic interpretation. Because regression requires a single outcome score, we created difference scores for our two dependent variables as noted below.

We subtracted the proportion of positive thoughts generated for negative events from positive thoughts generated for positive events to create an index representing relatively positive thoughts in relatively positive contexts. This index represented the tendency to engage in thinking whose valence matches the context (i.e., positive thinking in a positive context, negative thinking in a negative context), so we labeled it congruent outcome expectancy. Alternatively, low scores can be understood as “expected change” from the event to the outcome.

To test if culture influenced congruent outcome expectancy via differences in LTC, we ran Model 4 (simple mediation) in Process 3.3 with SPSS (Hayes, 2017). As seen in Figure 4, culture predicted LTC, B = 1.03 [0.92, 1.14], t(568) = 18.58, p < .001, and LTC predicted congruent outcome expectancy, B = −0.04 [−0.07, −0.01], t(567) = −2.38, p = .017. A 95% percentile-based bootstrapped confidence interval based on 10,000 bootstrap samples was entirely below zero [−0.08, −0.01], indicating a significant indirect effect of culture on congruent outcome expectancy through lay theories of change. Thus, Chinese people’s increased tendency to believe in the lay theory of change partially accounted for their weaker tendency for positive thinking in positive contexts and negative thinking in negative contexts (i.e., congruence between the valence of the event and the outcome). In sum, Study 3 replicated early findings, such that cultural differences in positive thinking depended on positive/negative context, and it provided evidence for a psychological mechanism: Chinese (vs Euro-Canadian) people’s increased belief in the lay theory of change explained some of the cultural differences in positive thinking.

Open multimedia modal

General Discussion

Psychological research often documents benefits of positive thinking (Carver & Scheier, 2009^; Taylor & Brown, 1988), but relatively little work explores when and how cultural groups differ in optimism. The present research identifies a contextual moderator that shapes cross-cultural differences in positive thinking. Consistent with cultural differences in lay theories of change (Ji, 2005^; Ji et al., 2001), we found that relative to Euro-Canadians, Chinese thought more negatively in positive contexts and more positively in negative contexts. The findings were obtained through diverse approaches to evaluating positive thinking. Studies 1-2 concerned the likelihood of positive/negative outcomes following from specific hypothetical events occurring to the respondent him/herself, whereas Study 3 targeted “people” in general (e.g., the outcomes of “a person getting mugged,” emphasis added) rather than the respondents themselves. Similarly, Study 1 used rating scales of fixed outcomes, Study 2 used a forced-choice between fixed outcomes, and Study 3 used an open thought listing/rating task. The breadth of measurement types in our work provides confidence that the effect observed is not localized to one scale or stimulus set but is valid across paradigms (see Judd et al., 2012). This breadth matches our conceptualization of positive thinking, which is not specific to self-relevant thoughts or particular response formats.

Conclusion

The present work re-examines positive thinking through a Culture × Context interactive lens. We have shown more positive thinking among Chinese than among Euro-Canadians in response to negative events, but less positive thinking among Chinese than among Euro-Canadians in response to positive events. The present research highlights how culture and situation can work together to influence people’s psychological responses.