RUNNING HEAD: NORMS ACROSS CULTURES 1
Norms Across Cultures: A Cross-Cultural Meta-Analysis of Norms Effects in the
Theory of Planned Behavior
Updated & corrected version: https://psyarxiv.com/v2wz6/ NORMS ACROSS CULTURES 2
Abstract
Norms have emerged as a central concept across various fields of psychology.
In social psychology, norms have been important to predict intentions and behavior, but cultural variability has not been examined. In cultural psychology, norms have also played a central role in explained cultural differences. In contrast, to date, variability in norm-intention and norm-behavior relationships has not been systematically investigated. Any systematic variability may be challenging to both social and cultural psychology. We re-analyzed effect sizes taken from five previously published meta-analyses using a fixed-effects model and demonstrate that the relative strength of norm-intention and norm-behavior correlations in this sample of previously published studies are systematically higher in less economically developed societies. We also found significant, but weaker, effects for individualism, tightness– looseness and monumentalism vs flexibility. Meanwhile, behavior domain effects also emerged, which suggests that norms are behavior specific. Norms effects systematically vary across previously published studies, implying that more attention is needed to investigate culturally conditioned domain and behavior effects.
Note: A previous version of this paper has been retracted from publication by the authors. The analyses and results reported in this pre-print are corrected and updated.
The retracted previous version featured incorrect results. NORMS ACROSS CULTURES 3
Norms Across Cultures: A Cross-Cultural Meta-Analysis of Norms Effects in the
Theory of Planned Behavior
Our behavior is shaped by social norms; our perceptions of what others do and approve strongly shapes our own behavior (Cialdini, Reno, & Kallgren, 1990). This insight is not new, with researchers both implicitly and explicitly studying norms at least since the beginnings of modern psychology, sociology, and anthropology in the late 19th century. Yet, to date, there is little attention to whether norms are equally effective across cultures. This is somewhat surprising because norms are by definition shared between group members and differ across groups (e.g., Jackson, 1966), raising the distinct possibilities that norms might also operate differently across cultural groups. Indeed, Triandis (1995), in his now classic definition of individualism- collectivism, proposed that norms are more important for individuals in collectivistic contexts. The study of norms has experienced a revival over the last two decades and within the realm of cultural psychology; for example, the concept of intersubjective descriptive norms (e.g., Chiu, Gelfand, Yamagishi, Shteynberg, & Wan, 2010;
Fischer, 2006; Wan et al., 2007) has gained much prominence for explaining cultural differences. This is also evidenced in a recent special issue in this journal. The goal of our article is to engage with the concept of norms from a cross-cultural comparative angle, by presenting a meta-analysis of norm research drawing on the theory of planned behavior (Ajzen, 1991). The insights from the meta-analysis allow us to critically assess how norms may emerge and how they are operating within and across cultural contexts. NORMS ACROSS CULTURES 4
Norms in Social Psychology
Social psychologists distinguish between two different types of norms. These are descriptive norms, referring to the behaviors people typically do, and injunctive norms, which are what people should do (implying a punitive element for deviating from the norm), based on what others approve or disapprove of (Cialdini et al., 1990).
Intervention studies promoting pro-environmental behaviors in primarily Western societies have given us several insights into norms. Specifically, norms exert the strongest influence on behavior when they are salient (Cialdini et al., 1990), and when both descriptive and injunctive norms are aligned to promote the same behavior
(Cialdini et al., 2006).
The most widely used norm concepts across cultures have been in the context of the TPB (Ajzen, 1988, 1991), which extends the earlier theory of reasoned action
(TRA; Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975). The main thrust is that behavioral intentions are determined by one’s evaluation of a behavior as positive
(attitude), and one’s perception that people one cares about want one to do the action
(subjective norm), as well as the belief that the one is able to perform the behavior
(perceived behavioral control). Subjective norms in this theoretical framework are conceptualized as ‘perceived social pressure to perform or not to perform the behavior’ (Ajzen, 1991, p.88). White et al. (2009) argued that subjective norms in the theory capture social injunctive norms, as they are capturing perceptions of whether significant others approve a specific behavior and imply an ‘ought’ component.
Armitage and Conner (2001) argued that social norms are perceptions of the social pressure by significant others combined with a motivation to comply with the normative expectations of those groups or individuals. This again suggests that it is more of an injunctive norm component. However, common measures of social norms NORMS ACROSS CULTURES 5 included the subjective experience of both descriptive and injunctive norms of those referent groups (see Ajzen & Klobas, 2013). The exact measurement of norms and who is considered as the norm reference group varies from study to study (see Ajzen,
1991). Therefore, within the context of the TPB, subjective norms can probably best be seen as a combination of descriptive and injunctive norms.
To date, the TRA and the TBP have been applied widely; this research has been summarized in several meta-analyses examining health-related behaviors such as exercise and food choices, technology adoption, and a variety of other behavioral domains. We found 74 meta-analyses that have demonstrated that this framework has emerged as one of the major theories within social psychology. In these meta- analyses, the general pattern is that the subjective norm component is typically a weaker predictor of behavioral intentions compared with attitudes (e.g., Ajzen, 1991;
Armitage & Conner, 2001; McDermott et al., 2015; Paquin & Keating, 2017). The relative weakness of the normative component within the theory have even led some researchers to drop it from the analyses (see Armitage & Conner, 2001). Yet, these discussions have not generally considered possible cultural differences in the importance of social norms. Given the importance of attitudes for informing behavioral decisions in the more individualistic contexts typically studied in psychological research, relevant effects may have been overlooked in previous research. In one of the rare meta-analytic investigations that explicitly considered cultural differences, Zhang, Zhu and Liu (2012) examined the influence of culture in terms of a West versus East difference on the association between subjective norms and perceived usefulness of mobile commerce. They found somewhat stronger norm correlations in Eastern cultures, although they did not find that culture dichotomized into East-West influenced the subjective norm-behavioral intention relationship. NORMS ACROSS CULTURES 6
Hence, their treatment of the data may have been suboptimal and obscured relevant patterns because only a generic East-West difference was used. What cultural, economic, or social variables may be relevant for understanding norm effects across contexts?
Norms Across Cultural Contexts
A number of cultural dimensions as well as ecological factors seem relevant for understanding norm effects. Here, we focus on some classic variables that are typically considered in cross-cultural psychology. First, individualism-collectivism is one of the core constructs of cross-cultural psychology. Triandis (1995) argued that one of the defining components is that collectivists are supposed to be more strongly guided by norms compared with individualists. Groups take on greater importance in more collectivistic contexts; hence, individuals are likely to pay more attention to normative information when deciding on actions instead of paying attention to their own impulses and inclinations. This conceptual distinction of individualism- collectivism along the attitude versus normative influence route has been supported when using a theory-driven measure of descriptive individualism-collectivism norms
(Fischer et al., 2009). Smith (2017) also found a significant effect of descriptive norms on helping behavior, but not emotion regulation. These two studies suggest that individuals in collectivistic environments pay attention to what other people are doing and then adjust their behavior accordingly. A number of diverse studies and methods have also provided more direct support for the central claim that (descriptive) norms are more important than attitudes or other self-centered psychological attributes, with normative pressures showing stronger relationships with intentions to stay with a company, and with behavioral intentions and willingness to participate in a study NORMS ACROSS CULTURES 7 without further pay in more collectivistic samples (e.g. Abrams, Ando, & Hinkle,
1998; Bontempo & Rivero, 1992; Cialdini, Wosinska, Barrett, Butner, & Gornik-
Durose, 1999; Fischer & Mansell, 2009). Matsumoto, Yoo, and Fontaine (2008) also demonstrated that the variability of emotion expression is related to individualism- collectivism, indicating that there are tighter descriptive norms around emotional display in collectivistic countries (but see Smith, 2017). The previous work has primarily focused on descriptive norms; however, given the importance of maintaining harmony in social relations, we would expect that social norms as perceived social pressure will show similar effects. Based on these findings, we predict that norms are more strongly related to both intentions and behaviors in more collectivistic contexts.
A second cultural variable that appears important is tightness-looseness
(Gelfand et al., 2011; Gelfand & Harrington, 2015). In loose cultures, there is a wide range of acceptable behaviors, and behavioral transgressions of weakly delineated and enforced norms are tolerated. In contrast, tight cultures restrict behavioral expression and only allow a narrower range of behaviors for individuals to engage in within specific situations. There is also a lower tolerance for those individuals who deviate from the social norm. Tight cultures tend to have higher population density and resource scarcity, which demand greater social order and thus stricter adherence to norms, which facilitate smooth social organization (e.g., Roos, Gelfand, Nau, & Lun,
2015). Leung and Morris’ (2015) situated dynamics model proposed that tightness- looseness is relevant for structuring situations, which then influences whether norms or values are relevant for informing behavior. Norms are thought to be more salient for behavior when social evaluations are salient, and when the right course of action is ambiguous. Social evaluation pressures are greater in tight situations. There was some NORMS ACROSS CULTURES 8 support for this reasoning in a recent country-level analysis by Smith (2017). Smith found that helping behavior was more strongly related to descriptive norms in contexts which, Gelfand et al. (2011) had classified as more tightly regulated. Based on this line of reasoning, norms should emerge as a stronger predictor of both behavioral intentions and actual behavior in tighter cultures compared with loose cultures.
A third cultural variable that might be relevant is monumentalism versus flexibility (e.g., Minkov, Bond, et al., 2018), which developed out of an effort to reinterpret and understand the long-term orientation dimension in Hofstede’s (2001) framework and future orientation as described by the Chinese Culture Connection
(1987). This dimension captures unique variance distinct from other established frameworks (Minkov, Blagoev, & Hofstede, 2013; Minkov & Hofstede, 2011).
Minkov, Bond and colleagues (2018) described ‘Monumentalism as a metaphor for a cultural tendency to encourage people to be like a monolithic monument: proud, stable, and consistent (made of the same substance outside and inside). Flexibility is the opposite cultural tendency, favoring a modest self-regard, duality, and adaptability’ (p. 321); therefore, “flexibility” is a descriptor of a culture that creates flexible and malleable human selves, capable and desirous of positive self- transformation. “Monumentalism” refers to cultures that value a self reminiscent of a proud and perennial monument. There is no strong perception of a need for self- improvement and the idea that one’s self is flexible and malleable is unpopular.’
(Minkov, Dutt, et al., 2018; pp. 224-225). Empirically, they demonstrated that this dimension distinguishes between a societal tendency to either emphasize consistent and unconditional exchange of help and services among strangers versus a more contextualized, situation-dependent conceptualization of the individual as a part of a NORMS ACROSS CULTURES 9 larger interconnected network. Therefore, the flexibility end emphasizes careful consideration of the situation, does not require consistency of behavior across different contexts, and duplicity in behavior (enacted behavior may not correspond to attitudes and beliefs). These considerations suggest some relevance for understanding norm-behavior relationships. If individuals are expected to be flexible in their behavior and to pay attention to the situational demands, we could expect stronger norm-behavior consistency in contexts characterized by high flexibility and low monumentalism.
A final dimension that we consider here is wealth. The post modernization hypothesis (Inglehart, 1997) first proposed that individuals with sufficient economic resources to secure their living are more likely to seek to express their thoughts and desires as they have the means to do so. In contrast, in more economically deprived conditions, individuals are trying to make ends meet and individual desires and thoughts take lower priority compared with economic survival (Fischer, 2008; Welzel,
2013). By implication, this restricts behavioral choices and people are more likely to follow what others are doing (which is norms). Therefore, the lack of means to express individual propensities and inclinations in resource poor environments will make it likely that behavior is more strongly guided by normative pressures. Indirect evidence to date comes from a number of studies showing that values, attitudes, and reports of behavioral traits are less strongly linked in more resource restricted environments (Boer & Fischer, 2013; Fischer, 2017).
Therefore, the effect of norms on behavior is likely to depend on the cultural and economic context, with stronger normative influences on both intentions and behavior in collectivistic, tighter, and more economically-deprived populations. NORMS ACROSS CULTURES 10
Method
We gathered information from five meta-analyses that used subjective norms
(Paquin & Keating, 2017; Plotnikoff, Costigan, Karunamuni, & Lubans, 2013; Riebl,
2015; Scalco, Noventa, Sartori, & Ceschi, 2017; Schüz, Li, Hardinge, McEachan, &
Conner, 2017) and provided the individual effect sizes for norm-intention or norm- behavior associations re-analyses. Our selection was opportunistic and decided by the quality of the available information reported in those meta-analyses identified in our search. From these selected meta-analyses, we used both the norm – intention and norm – behavior correlations. We excluded effect sizes and samples that could not be attributed to a single national population. Our total sample consisted of 122 effect sizes encompassing 54,970 participants for social norm-intention associations and 127 effect sizes encompassing 47,391 participants for social norm – behavior associations based on a total of 203 studies from 31 countries. We report the country composition of our sample in Table 1.
– Insert Table 1 about here –
Moderator Variables
We used a combined score for individualism (see Fischer & Boer, 2011;
Fischer & Van de Vliert, 2011), averaging normalized scores for Inglehart’s (1997) survival versus well-being dimension, Hofstede’s Individualism index, and
Schwartz’s (1994) autonomy versus embeddedness score for teachers and students.
This indicator has been shown to provide a reliable and valid measure of overall individualism-collectivism across nations (Fischer & Boer, 2011).
We did not have data for 9 countries, therefore, the effects of tightness are estimated with a smaller sample of 22 countries. Higher scores indicated greater NORMS ACROSS CULTURES 11 tightness. We also used a distribution-based score of tightness-looseness proposed by
Uz (2015). We used the combined cultural-tightness/looseness score. We did not have data for seven countries (Australia, China, Hong Kong, New Zealand, Norway,
Pakistan, and Taiwan). Higher scores indicate greater looseness.
Monumentalism versus flexibility was measured with scores provided by
Minkov et al. (2018). We were able to match scores from 26 nations in our data-base.
Higher values indicate higher flexibility. We were able to match scores from twenty- five nations in our data-base. Higher values indicate higher flexibility.
Wealth was measured with the 2017 index of Gross Domestic Product per capita, expressed as Purchasing Power Parity from the CIA World Factbook (Central
Intelligence Agency, 2017). We z-transformed this index to provide a more meaningful scale for our analyses. We report the correlation between those variables in Table 2.
– Insert Table 2 about here –
We also coded the behavioral domain of every effect size as either health- related (e.g., food choice etc., k = 71), or physical exercise (exercise, k = 104), or other (diverse, k = 28). Due to unequal sample sizes across domains, we summarized health-domains and other domains into a single domain (‘Other’) and contrasted this in our analysis with the physical activity (‘Exercise’) domain.
Analytical Strategy
We used the metafor package (Viechtbauer, 2010) in R (R Core Team, 2018) to run our analyses. Primary studies that did not report correlation coefficients but only path coefficients were corrected using the procedure introduced by Peterson and
Brown (2005). All effect sizes were r-to-z-transformed and we ran a fixed-effects meta-regression analysis using REML (restricted maximum likelihood) estimation to NORMS ACROSS CULTURES 12 test the effects of our culture-level moderators on the sample-level correlations. We also included behavioral domain (physical exercise vs other) as a moderator variable.
It is important to note that a fixed-effect model only allows generalizations for these previously published studies (see Lipsey & Wilson, 2001). We include the code for two different specifications of a random effect model with REML (random intercept at study level, random intercept at study and nation-level) on the OSF and briefly discuss these random effect results below.
Results
We are interested in the cultural moderator effects on the effect sizes that were previously reported in published meta-analyses. For this reason, we proceeded straight to the moderator analyses. We first ran each moderator separately for the norm- intention and norm-behavior effect sizes. For norm-intention correlations, all moderator effects (except tightness-looseness measured with the index by Uz, 2015) were significant at p < .05. Norm correlations with behavioral intentions were stronger in less individualistic countries, in tighter countries, in countries with higher monumentalism (less flexibility), and in less economically developed countries. For norm-behavior correlations, all but the behavioral tightness-looseness indicator by Uz
(2015) and the flexibility index were significant. As before, norm-behavior correlations were stronger in less individualistic, in tighter countries and in countries with lower income.
We then examined whether behavioral domain (physical exercise vs other) made a difference on the relationship between norms and behavioral intentions. The norm effects were stronger for physical exercise versus other behavioral intentions in tighter societies and in more economically advanced societies. In other words, NORMS ACROSS CULTURES 13 intentions to exercise were more strongly correlated with norms in countries in which there are tight norms and where individuals on average have more income. Exercise behavior might be particularly desirable, important, or normative in these contexts
(e.g., richer and tighter contexts). When focusing on norm-behavior correlations, there were also some interactions with physical exercise versus other types of behaviors.
Interestingly, the effects were reversed from behavioral intentions for tightness- looseness. In looser countries, exercise behavior was more strongly correlated with norms, compared to other behaviors. In more collectivistic countries, exercise behavior was more strongly correlated with norms, compared to other behaviors. In countries that emphasize flexibility, exercise behavior was more strongly correlated with norms, compared to other behaviors. Finally, in more economically advanced countries, exercise behavior was more strongly correlated with norms, compared to other behaviors.
We then ran a combined analysis in which we entered all four indicators together (separately for the two different tightness-looseness indicators). For behavioral intentions, when using the looseness indicator by Uz (2015), we found significant effects for wealth (B =-.044, 95% CI [-.057, -.031], p < .001), flexibility
(B =-.001, 95% CI [-.001, -.001], p < .001) and individualism (B =-.045, 95% CI
[-.087, -.003], p < .05). The effect for looseness was non-significant (B =-.001, 95%
CI [-.001, .001], p > .05). When using the behavioral measure of tightness by Gelfand et al. (2011), we again found significant effects for wealth (B =-.046, 95% CI [-.057,
-.036], p < .001), flexibility (B = -.001, 95% CI [-.001, -.001], p < .001), tightness (B
= .013, 95% CI [.007, .018], p < .001) and individualism (B =-.043, 95% CI [-.065,
-.022], p < .001). Hence, lower average wealth and more emphasis on monumentalism NORMS ACROSS CULTURES 14
(vs flexibility), greater cultural tightness and greater collectivism were associated with strengthened norm-behavioral intention correlations.
For norm-behavior correlations, we found significant effects for looseness, wealth, and flexibility. Specifically, when using the looseness measure by Uz (2015), we found effects for individualism (B =.350, 95% CI [.273, .428], p < .001), looseness (B =-.010, 95% CI [-.012, -.008], p < .001), wealth (B =-.168, 95% CI
[-.191, -.145], p < .001) and flexibility (B =-.002, 95% CI [-.002, -.002], p < .001).
When using the Gelfand et al.’s (2011) measure of tightness, we found effects for wealth (B =-.109, 95% CI [-.128, -.091], p < .001) and flexibility (B =-.001, 95% CI
[-.001, -.001], p < .001). The effects for individualism (B = -.026, 95% CI
[-.058, .005], p > .05) and tightness (B = -.005, 95% CI [-.013, .002], p > .05) were not significant. The effects for wealth and flexibility mirror the effects for behavioral intentions (see Figures 1 to 4); greater wealth and greater emphasis on flexibility (less emphasis on monumentalism) are associated with weakened norm-behavior correlations. The effects for tightness-looseness and individualism are not consistent with the single predictor analyses and the behavioral intention analyses. These effects might be due to suppressor effects because of the small sample size and intercorrelations between the predictor variables.
– Insert Figure 1-4 about here –
– Insert Table 3 about here –
One important caveat is that the current analyses are based on a fixed-effects model. When running REML with either a random intercept at study level or random intercepts at both study and nation-level, the effects reported are not significant anymore. Future research needs to examine the source of variability and whether a fixed vs random-effect model is more appropriate for cross-cultural meta-analyses in NORMS ACROSS CULTURES 15 which weights are unequally distributed across studies and nations (see Kende et al.,
2017 for an illustration and possible solution).
Discussion
Norms have been discussed extensively in the social psychology literature. We reported a secondary meta-analysis to examine the relative strength of norm- behavioral intention and norm-behavior relationships across previously published studies drawn from different national cultures. Overall, we show that norm-behavior relationships vary significantly and consistently across these cross-cultural studies.
The strongest and most consistent unique effects that we found are for wealth and, specifically for norm-intentions correlations, for monumentalism versus flexibility; the effects of norms are stronger in less economically-advanced contexts and in contexts where there is a cultural emphasis on monumentalism (instead of a cultural focus on flexibility). The wealth effects are in line with previous research (Fischer,
2017; Welzel, 2013), however, the monumentalism vs flexibility effects are novel and our pattern will need further validation and elaboration. We also found some effects of tightness–looseness and individualism; in published studies from more collectivistic and tighter societies, the effects of norms on behavior and behavioral intentions are stronger. We also found some effects of individualism and tightness- looseness; in available studies from more collectivistic and tighter societies, norm effects are stronger. Therefore, an important take-home message is that norm effects vary across these samples from different cultural settings and that we can predict where norms might be more important in this dataset. A recent special issue in this journal provided important new insights into the cultural dynamics, but largely remained silent on the differential importance of cultural differences when discussing norms (for a few notable exceptions, see for example, Frese, 2015; Smith, 2015). NORMS ACROSS CULTURES 16
At the same time, the norm effects appear to vary across behavioral domains in our database; we found that physical exercise versus other behaviors often showed slightly different effects. This suggests that norm effects may not be situation and behavior independent. This raises some important questions for both norm psychology research in general as well as cultural models such as tightness-looseness
(Gelfand et al., 2011) and individualism-collectivism (Triandis, 1995).
Rethinking norms across cultures
Work in sociology treats norms as multidimensional constructs (Gibbs, 1965;
Jackson, 1966; Jasso & Opp, 1997; Opp, 1982), varying in intensity, consensus, and tightness. Intensity is the strength of group members’ approval or disapproval of a specific behavior. Consensus is the extent to which norms are shared among members, indicating collective agreement and acceptance of the norm. Tightness refers to the range of acceptable behaviors within a situation. This view on norms dovetails nicely with the conceptualization of cultural tightness-looseness as discussed above. At the same time, the sociological treatment raises a number of questions for norms in a cross-cultural perspective.
Most of the research in psychology, especially in cross-cultural comparative work, treats norms as relatively abstract and content-free concepts. The relative intensity of norms across a broad range of behaviors is often the key concern. The questions of consensus and tightness for specific norms are not typically considered.
For example, the situated dynamics framework (Leung & Morris, 2015) focuses on intensity of norms, and treats tightness as a cultural characteristic that makes all norms more salient and relevant for behavior. Our results suggest that norms might be focused on specific behaviors (e.g., norms for physical exercise, socio-sexual relations, driving). These individual norms may have some communality across NORMS ACROSS CULTURES 17 behaviors within a cultural system, but the extent to which this is the case is an important empirical question to study first (see Gelfand et al., 2011). It would be useful to examine the situation and behavior specificity of norms in greater detail across different cultural contexts.
Our results also suggest that behavioral intentions and reports of specific behavior are not necessarily aligned from a cultural perspective. We found that intentions for physical exercise are more strongly correlated with norms in published studies from tight societies, but correlations of norms with reports of behavior are strengthened in published studies in looser societies. We do not want to overemphasize the current patterns, but it might be worth considering how intentions may capture internalized (perceived) societal expectations, whereas behaviors might be driven more by situational strengths (see Leung & Morris, 2015 on the importance of ambiguity). These issues are linked to previously reported problems, such as the observation that self- and collective perceptions of values and behaviors may not strongly overlap in the cultural domain (Fischer, 2006; House, Hanges, Javidan,
Dorfman, & Gupta, 2004). In social psychological research, at least two major biases have been reported: a) the overestimation of the extent to which one’s own personal opinion is shared among the majority (false consensus bias, Ross, Greene, & House,
1977) and b) overestimating the subjective support for norms by others when everybody appears to behave in line with the perceived norm (pluralistic ignorance,
Prentice & Miller, 1996). We urge researchers to pay more attention to possible divergence of perceptions of behaviors of others, intentions, and actual behavior across cultural contexts.
This last point is also important in relation to the noted distinction between descriptive and injunctive norms. Conceptually, TPB researchers often treat social NORMS ACROSS CULTURES 18 norm perceptions as injunctive, but the operationalization may include descriptive norms (see Ajzen, 1991). From a cultural perspective, we could expect slightly different dynamics for descriptive versus injunctive norms. Injunctive norms may be particularly relevant in ambiguous situations and if the social context is more tightly regulated (see Leung & Morris, 2015). This is an important avenue for future research.
Our results also need to be interpreted with some caution. As can be seen in
Table 1, the majority of studies that were included in these meta-analyses have been conducted in Europe, with very few studies reported in Asia, only two studies from
Africa, and no study conducted in Central or South America. It is also of importance to consider the representativeness and comparability of samples, both in terms of their national culture and in relation to each other. We also need to highlight the uneven sample sizes, which influence the sampling weights that are applied for fixed vs random-random effects models. These are a major limitations that are common to many cross-cultural studies using convenience samples. As shown above, the fixed effect models results were not replicated when using random effect models with random intercepts at either study or study and nation-level. These discrepancies need further attention in future studies with larger data bases.
Conclusion
Our re-analysis of published meta-analytic data demonstrated that norm- intention and norm-behavior correlations systematically vary across published studies.
In published studies from more individualistic, high income, more flexible, and loose societies, norms tend to be less strongly related to both behavioral intentions and behaviors. However, these analyses also demonstrated that behavioral domains make NORMS ACROSS CULTURES 19 a difference. Our aim is to stimulate debate about the role of norms for the study of culture and highlight both the potential and the intricacies of a normative approach for cross-cultural research. Our study demonstrates that one of the reasons why norms have been less strongly related to behavior is that the cultural context has not been considered. We hope that the findings of the meta-analysis help to clarify that norms are relevant and systematically relate to both intentions and behavior reports across different cultural contexts. RUNNING HEAD: NORMS ACROSS CULTURES 20
References
Abrams, D., Ando, K., & Hinkle, S. (1998). Psychological attachment to the group:
Cross-cultural differences in organizational identification and subjective norms
as predictors of workers’ turnover intentions. Personality and Social Psychology
Bulletin, 24(10), 1027–1039. doi:10.1177/01461672982410001
Ajzen, I. (1988). Attitudes, personality, and behavior. Milton Keynes: Open
University Press.
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and
Human Decision Processes, 50(2), 179–211. doi:10.1016/0749-5978(91)90020-
T
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social
behavior. Englewood Cliffs: Prentice-Hall.
Ajzen, I., & Klobas, J. (2013). Fertility intentions: An approach based on the theory of
planned behavior. Demographic Research, 29, 203–232.
doi:10.4054/DemRes.2013.29.8
Armitage, C. J., & Conner, M. (2001). Efficacy of the Theory of Planned Behaviour:
A meta-analytic review. British Journal of Social Psychology, 40(4), 471–499.
doi:10.1348/014466601164939
Boer, D., & Fischer, R. (2013). How and when do personal values guide our attitudes
and sociality? Explaining cross-cultural variability in attitude–value linkages.
Psychological Bulletin, 139(5), 1113–1147. doi:10.1037/a0031347
Bontempo, R., & Rivero, J. C. (1992). Cultural variation in cognition: The role of
self-concept in the attitude-behavior link. In Meeting of the American academy
of management. Las Vegas, Nevada.
Central Intelligence Agency (2017). The World Factbook. Retrieved November 28, NORMS ACROSS CULTURES 21
2018, from https://www.cia.gov/library/publications/resources/the-world-
factbook/
Chiu, C.-Y., Gelfand, M. J., Yamagishi, T., Shteynberg, G., & Wan, C. (2010).
Intersubjective Culture. Perspectives on Psychological Science, 5(4), 482–493.
doi:10.1177/1745691610375562
Cialdini, R. B., Demaine, L. J., Sagarin, B. J., Barrett, D. W., Rhoads, K., & Winter,
P. L. (2006). Managing social norms for persuasive impact. Social Influence,
1(1), 3–15. doi:10.1080/15534510500181459
Cialdini, R. B., Reno, R. R., & Kallgren, C. A. (1990). A focus theory of normative
conduct: Recycling the concept of norms to reduce littering in public places.
Journal of Personality and Social Psychology, 58(6), 1015–1026.
doi:10.1037/0022-3514.58.6.1015
Cialdini, R. B., Wosinska, W., Barrett, D. W., Butner, J., & Gornik-Durose, M.
(1999). Compliance with a request in two cultures: The differential influence of
social proof and commitment/consistency on collectivists and individualists.
Personality and Social Psychology Bulletin, 25(10), 1242–1253.
doi:10.1177/0146167299258006
Chinese Culture Connection (1987). Chinese values and the search for culture-free
dimensions of culture. Journal of Cross-Cultural Psychology, 18(2), 143–164.
doi:10.1177/0022002187018002002
Fischer, R. (2006). Congruence and functions of personal and cultural values: Do my
values reflect my culture’s values? Personality and Social Psychology Bulletin,
32(11), 1419–1431. doi:10.1177/0146167206291425
Fischer, R. (2008). Rewarding seniority: Exploring cultural and organizational
predictors of seniority allocations. The Journal of Social Psychology, 148(2), NORMS ACROSS CULTURES 22
167–186. doi:10.3200/SOCP.148.2.167-186
Fischer, R. (2017). Personality, values, culture: An evolutionary perspective.
Cambridge University Press. doi:10.1017/9781316091944
Fischer, R., & Boer, D. (2011). What is more important for national well-being:
Money or autonomy? A meta-analysis of well-being, burnout, and anxiety across
63 societies. Journal of Personality and Social Psychology, 101(1), 164–184.
doi:10.1037/a0023663
Fischer, R., Ferreira, M. C., Assmar, E., Redford, P., Harb, C., Glazer, S., … Achoui,
M. (2009). Individualism-collectivism as descriptive norms. Journal of Cross-
Cultural Psychology, 40(2), 187–213. doi:10.1177/0022022109332738
Fischer, R., & Mansell, A. (2009). Commitment across cultures: A meta-analytical
approach. Journal of International Business Studies, 40(8), 1339–1358.
doi:10.1057/jibs.2009.14
Fischer, R., & Van de Vliert, E. (2011). Does climate undermine subjective well-
being? A 58-nation study. Personality and Social Psychology Bulletin, 37(8),
1031–1041. doi:10.1177/0146167211407075
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An
introduction to theory and research. Reading, MA: Addison-Wesley.
Frese, M. (2015). Cultural practices, norms, and values. Journal of Cross-Cultural
Psychology, 46(10), 1327–1330. doi:10.1177/0022022115600267
Gelfand, M. J., & Harrington, J. R. (2015). The motivational force of descriptive
norms. Journal of Cross-Cultural Psychology, 46(10), 1273–1278.
doi:10.1177/0022022115600796
Gelfand, M. J., Raver, J. L., Nishii, L., Leslie, L. M., Lun, J., Lim, B. C., …
Yamaguchi, S. (2011). Differences between tight and loose cultures: A 33-nation NORMS ACROSS CULTURES 23
study. Science, 332(6033), 1100–1104. doi:10.1126/science.1197754
Gibbs, J. P. (1965). Norms: The problem of definition and classification. American
Journal of Sociology, 70(5), 586–594.
Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors,
institutions, and organizations across nations (2nd ed.). Thousand Oaks, CA.:
Sage Publications.
House, R. J., Hanges, P. J., Javidan, M., Dorfman, P. W., & Gupta, V. (2004).
Culture, leadership, and organizations : the GLOBE study of 62 societies. Sage
Publications.
Inglehart, R. (1997). Modernization and postmodernization : cultural, economic, and
political change in 43 societies. Princeton University Press.
Jackson, J. (1966). A conceptual and measurement model for norms and roles. The
Pacific Sociological Review, 9(1), 35–47. doi:10.2307/1388305
Jasso, G., & Opp, K.-D. (1997). Probing the character of norms: A factorial survey
analysis of the norms of political action. American Sociological Review, 62(6),
947–964.
Kende J., Phalet K., Van Den Noortgate W., Kara A., Fischer R. (2018). Equality
revisited: A cultural meta-analysis of the contact hypothesis. Social
Psychological and Personality Science, 9, 887–895.
Leung, K., & Morris, M. W. (2015). Values, schemas, and norms in the culture–
behavior nexus: A situated dynamics framework. Journal of International
Business Studies, 46(9), 1028–1050. doi:10.1057/jibs.2014.66
Lipsey, M. W., & Wilson, D. B. 2001. Practical meta-analysis. London: Sage.
Matsumoto, D., Seung Hee Yoo, S. H., & Fontaine, J. (2008). Mapping expressive
differences around the world. Journal of Cross-Cultural Psychology, 39(1), 55– NORMS ACROSS CULTURES 24
74. doi:10.1177/0022022107311854
McDermott, M. S., Oliver, M., Svenson, A., Simnadis, T., Beck, E. J., Coltman, T., …
Sharma, R. (2015). The theory of planned behaviour and discrete food choices: a
systematic review and meta-analysis. International Journal of Behavioral
Nutrition and Physical Activity, 12(1), 162. doi:10.1186/s12966-015-0324-z
Minkov, M., Blagoev, V., & Hofstede, G. (2013). The boundaries of culture. Journal
of Cross-Cultural Psychology, 44(7), 1094–1106.
doi:10.1177/0022022112466942
Minkov, M., Bond, M. H., Dutt, P., Schachner, M., Morales, O., Sanchez, C., …
Mudd, B. (2018). A reconsideration of Hofstede’s fifth dimension: New
flexibility versus monumentalism data from 54 countries. Cross-Cultural
Research, 52(3), 309–333. doi:10.1177/1069397117727488
Minkov, M., Dutt, P., Schachner, M., Jandosova, J., Khassenbekov, Y., Morales, O.,
Sanchez, C.J., & Mudd, B. (2018). What values and traits do parents teach to
their children? New data from 54 countries. Comparative Sociology, 17, 221-
252.
Minkov, M., & Hofstede, G. (2011). The evolution of Hofstede’s doctrine. Cross
Cultural Management: An International Journal, 18(1), 10–20.
doi:10.1108/13527601111104269
Opp, K.-D. (1982). The evolutionary emergence of norms. British Journal of Social
Psychology, 21(2), 139–149. doi:10.1111/j.2044-8309.1982.tb00522.x
Paquin, R. S., & Keating, D. M. (2017). Fitting identity in the reasoned action
framework: A meta-analysis and model comparison. The Journal of Social
Psychology, 157(1), 47–63. doi:10.1080/00224545.2016.1152217
Peterson, R. A., & Brown, S. P. (2005). On the use of beta coefficients in meta- NORMS ACROSS CULTURES 25
analysis. Journal of Applied Psychology, 90(1), 175–181. doi:10.1037/0021-
9010.90.1.175
Plotnikoff, R. C., Costigan, S. A., Karunamuni, N., & Lubans, D. R. (2013). Social
cognitive theories used to explain physical activity behavior in adolescents: A
systematic review and meta-analysis. Preventive Medicine, 56(5), 245–253.
doi:10.1016/J.YPMED.2013.01.013
Prentice, D. A., & Miller, D. T. (1996). Pluralistic ignorance and the perpetuation of
social norms by unwitting actors. Advances in Experimental Social Psychology,
28, 161–209. doi:10.1016/S0065-2601(08)60238-5
R Core Team. (2018). R: A language and environment for statistical computing.
Vienna, Austria. Retrieved from https://www.r-project.org/
Riebl, S. K. (2015). Understanding adolescents’ sugary beverage consumption: A
review and application of the Theory of Planned Behavior. Virginia Polytechnic
Institute and State University.
Roos, P., Gelfand, M., Nau, D., & Lun, J. (2015). Societal threat and cultural variation
in the strength of social norms: An evolutionary basis. Organizational Behavior
and Human Decision Processes, 129, 14–23. doi:10.1016/J.OBHDP.2015.01.003
Ross, L., Greene, D., & House, P. (1977). The “false consensus effect”: An egocentric
bias in social perception and attribution processes. Journal of Experimental
Social Psychology, 13(3), 279–301. doi:10.1016/0022-1031(77)90049-X
Scalco, A., Noventa, S., Sartori, R., & Ceschi, A. (2017). Predicting organic food
consumption: A meta-analytic structural equation model based on the theory of
planned behavior. Appetite, 112, 235–248. doi:10.1016/J.APPET.2017.02.007
Schüz, B., Li, A. S.-W., Hardinge, A., McEachan, R. R. C., & Conner, M. (2017).
Socioeconomic status as a moderator between social cognitions and physical NORMS ACROSS CULTURES 26
activity: Systematic review and meta-analysis based on the Theory of Planned
Behavior. Psychology of Sport and Exercise, 30, 186–195.
doi:10.1016/J.PSYCHSPORT.2017.03.004
Schwartz, S. H. (1994). Are there universal aspects in the structure and sontents of
human values? Journal of Social Issues, 50(4), 19–45. doi:10.1111/j.1540-
4560.1994.tb01196.x
Smith, P. B. (2015). Yes, subjective norms are important, but let’s not lose sight of
cultural differences. Journal of Cross-Cultural Psychology, 46(10), 1310–1313.
doi:10.1177/0022022115599444
Smith, P.B. (2017). Cultural values versus cultural norms as predictors of differences
in helping behaviors and in emotion regulation: A preliminary nation-level test
related to the Leung-Morris model. Management and Organization Review,
13(4), 739-766. doi: 10.1017/mor.2017.51.
Triandis, H. C. (1995). Individualism & collectivism. Taylor & Francis.
Uz, I. (2015). The index of cultural tightness and looseness among 68 countries.
Journal of Cross-Cultural Psychology, 46(3), 319–335.
doi:10.1177/0022022114563611
Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package.
Journal of Statistical Software, 36(3), 1–48.
Wan, C., Chiu, C., Tam, K.-P., Lee, S., Lau, I. Y., & Peng, S. (2007). Perceived
cultural importance and actual self-importance of values in cultural
identification. Journal of Personality and Social Psychology, 92(2), 337–354.
doi:10.1037/0022-3514.92.2.337
Welzel, C. (2013). Freedom rising : human empowerment and the quest for
emancipation. NORMS ACROSS CULTURES 27
White, K. M., Smith, J. R., Terry, D. J., Greenslade, J. H., & McKimmie, B. M.
(2009). Social influence in the theory of planned behaviour: The role of
descriptive, injunctive, and in-group norms. British Journal of Social
Psychology, 48(1), 135–158. doi:10.1348/014466608X295207
Zhang, L., Zhu, J., & Liu, Q. (2012). A meta-analysis of mobile commerce adoption
and the moderating effect of culture. Computers in Human Behavior, 28(5),
1902–1911. doi:10.1016/J.CHB.2012.05.008
Figure 1: Meta-Analytic Scatterplot for Flexibility as a single predictor for norm-intention correlations.
Note: Higher scores represent greater flexibility (lower stability and consistency). Point size represents combined total sample size per
country. Correlation coefficients represent Fisher z-transformed values. NORMS ACROSS CULTURES 28
Figure 2: Meta-Analytic Scatterplot for Wealth (expressed as purchasing power parity per capita) as a single predictor for norm-
intention correlations.
Point size represents combined total sample size per country. Correlation coefficients represent Fisher z-transformed values. NORMS ACROSS CULTURES 29
Figure 3: Meta-Analytic Scatterplot for Flexibility as a single predictor for norm-behavior correlations.
Note: Higher scores represent greater flexibility (lower stability and consistency). Point size represents combined total sample size per
country. Correlation coefficients represent Fisher z-transformed values. NORMS ACROSS CULTURES 30
Figure 4: Meta-Analytic Scatterplot for Wealth (expressed as purchasing power parity per capita) as a single predictor for norm-
behavior correlations
Note: Point size represents combined total sample size per country. Correlation coefficients represent Fisher z-transformed values.
Table 1. Descriptives of the overall Sample NORMS ACROSS CULTURES 31
Number of Percentage of the Total Combined N of Samples Combined N of Samples Country Samples Number of Samples Behavior Intention Australia 26 12.81 3055 4786 Austria 1 .49 215 215 Belgium 1 .49 0 456 Canada 25 12.32 6371 4315 China 1 .49 0 211 Czech Republic 1 .49 1054 1054 Denmark 2 .99 0 652 Estonia 1 .49 0 432 Finland 3 1.48 0 639 France 3 1.48 125 3578 Germany 3 1.48 538 470 Greece 5 2.46 0 1389 Hong Kong 3 1.48 136 641 Hungary 1 .49 0 150 India 1 .49 0 220 Iran (Islamic Republic of) 2 .99 521 910 Italy 5 2.46 3037 3539 Netherlands 8 3.94 7570 6840 New Zealand 2 .99 110 565 Norway 9 4.43 808 2757 Pakistan 1 .49 0 184 Portugal 1 .49 156 177 Republic of Korea 3 1.48 767 836 Romania 1 .49 35 0 Singapore 1 .49 0 235 Spain 2 .99 0 853 NORMS ACROSS CULTURES 32
Taiwan 2 .99 614 300 Tanzania 1 .49 0 312 Uganda 2 .99 372 1518 United Kingdom 48 23.65 8970 5913 United States of America 38 18.72 12937 10823
Table 2. Correlations and Descriptive Statistics of Country-Level Moderator Variables. Mean SD Min Max 1 2 3 4 Individualism 0.33 0.84 -1.18 1.51 Uz (2015): Looseness 64.30 24.40 20.10 119.80 .78** Gelfand et al. (2011): Tightness 6.66 2.72 2.60 12.30 -.51* -.21 Flexibility 49.08 64.43 -64.00 199.00 -.35 -.05 .24 Wealth (GDP per capita Purchasing Power 38587.10 19800.28 2400.00 90500.00 .41* .45* -.08 .31 Parity) Note. GDP = gross domestic product. * p <.05, ** p < .01. NORMS ACROSS CULTURES 33
Table 3. Results of the Meta-Regressions on Norm-Intention and Norm-Behaviour Links. Intention Behaviour Block 1 Block 2 Block 1 Block 2 Individualism Intercept .42[.41, .44]*** .48[.46, .49]*** .29[.27, .31]*** .33[.30, .36]*** Individualism -.02[-.03, -.00]** -.02[-.03, -.01]** -.08[-.09, -.06]*** -.04[-.06, -.01]** Domain -.16[-.20, -.12]*** -.04[-.07, -.00]* Domain * Individualism -.04[-.09, .01] -.10[-.13, -.06]*** Tightness - looseness (using data by Uz, 2015; lower scores indicate tighter societies) Intercept .40[.37, .44]*** .43[.39, .48]*** .19[.14, .23]*** .31[.24, .39] *** Tightness - looseness -.00[-.00, .00] .00[-.00, .00] .00[-.00, .00] -.00[-.00, .00] Domain .11[.02, .21]* -.20[-.29, -.10] *** Domain * Tightness - looseness -.00[-.01, -.00]*** .00[-.00, .00] Tightness - Looseness (using data by Gelfand, 2011; higher scores indicate tighter societies) Intercept .31[.28, .34]*** .39[.36, .42]*** .12[.08, .16]*** .11[.06, .15]*** Tightness - looseness .02[.02, .03]*** .01[.01, .02]*** .02[.01, .03]*** .04[.03, .05]*** Domain -.27[-.36, -.18]*** .08[.01, .16]* Domain * Tightness - looseness .03[.01, .04]*** -.04[-.05, -.03]*** Wealth Intercept .41[.40, .42]*** .46[.45, .47]*** .23[.22,.24]*** .29[.28, .31]*** GDP -.01[-.02, -.00]* -.01[-.02, -.00]* -.05[-.06, -.04]*** -.04[-.06, -.03]*** Domain -.22[-.24, -.20]*** -.10[-.12, -.08]*** Domain * GDP .11[.09, .14]*** .03[.00, .05]* Monumentalism – Flexibility Intercept .46[.45, .47]*** .54[.53, .56]*** .23[.22, .24]*** .39[.37, .42]*** Flexibility -.00[-.00, -.00]*** -.00[-.00, -.00]*** -.00[-.00, .00] -.00[-.00, -.00]*** Domain -.22[-.26, -.19]*** -.27[-.29, -.24]*** Domain * Flexibility .00[-.00, .00] .00[.00, .00]*** NORMS ACROSS CULTURES 34
Notes. GDP = gross domestic product. Domain was coded 0 for ‘Other’ and 1 for ‘Exercise’. Parameter estimates are unstandardized regression coefficients. Numbers in brackets display upper and lower bound of 95% confidence intervals. * p <.05, ** p <.01, *** p < .001.