Abstract
We study relations of two value dimensions—self-enhancement versus self-transcendence and conservation versus openness to change—with attitudes to cultural diversity. We examine two potential moderators of the value–attitude link, the meaning and the level of cultural diversity. We operationalize the meaning as the perceived consequences of cultural diversity for attaining value-relevant goals. We hypothesize that the perceived consequences depend on a group’s status: majority versus minority and high versus low. We test these hypotheses on representative samples from 25 regions of Russia (N = 1,934). As hypothesized, the self-enhancement/self-transcendence link to attitude, though negative in all groups, was most negative in the low-status group. The conservation/openness link to attitude was negative for majority Russians, as in past research, but it was positive among minorities. Level of cultural diversity had no moderating effect. This study highlights the importance of group differences in the meaning of attitude objects for understanding attitude predictors.
Since 1992, when the Schwartz theory of basic human values was introduced (Schwartz, 1992), hundreds of studies have tested the links between values and various attitudes. Values predict attitudes towards specific objects to the extent that these objects help or hinder the expression of the values (Sagiv & Schwartz, 1995). Findings typically reveal that values are consistent predictors of attitudes across contexts. However, whether an attitude is perceived as an expression of a certain value depends on contextual characteristics. Some recent studies argue that social context moderates relationships of values with attitudes (for review, see Roccas & Sagiv, 2010). This study contributes to the growing literature on moderators of value–attitude links by examining whether level of cultural diversity and its meaning in different subgroups in Russia moderate how self-enhancement (vs. self-transcendence) and conservation (vs. openness to change) values predict attitudes toward cultural diversity.
These two moderators—level of cultural diversity and its meaning—represent two somewhat competing views of variation in value–attitude relations. The first traces variation in the strength of the value–attitude link to variation in the context in which it occurs (i.e., level of diversity in the cultural environment). The second, though not denying the role of context, traces variation in the link to variation in the meaning of the attitude object (the meaning of diversity). We aim to test both of these predictions.
Most past studies drew upon the mechanism of normative expectations to explain variation in value–attitude relations. This mechanism posits that value–attitude relations are stronger when individuals are free to form their own attitudes unconstrained by normative expectations. This mechanism can explain why value–attitude relations are stronger in loose (vs. tight) cultures, in weak (vs. strong) situations, in cultures that encourage (vs. discourage) self-expression (summarized in Roccas & Sagiv, 2010), in countries with loose (vs. strict) legal regulations (Kuntz, Davidov, Schwartz, & Schmidt, 2015), and in less (vs. more) threatening environments (Boer & Fischer, 2013).
Roccas and Sagiv (2010) applied this reasoning to propose that level of cultural diversity in the environment may also moderate relations of values to attitudes and behaviors. They argued that cultural diversity might weaken these relations but also might enhance them. Greater cultural diversity might weaken these relations by impelling people to define themselves more in terms of their group memberships. This would make normative group expectations salient, thereby encouraging more norm-based and less personal choices. Alternatively, greater diversity might enhance value–attitude relations by making the normative environment more ambiguous. This would provide more freedom to express one’s own values. We assess these as yet untested, competing possibilities for the moderating effect of level of cultural diversity on the value–attitude link.
The second moderator we explore is the meaning of the attitude object. The normative expectations mechanism explains variation across contexts in the strength of value–attitude relations. However, it cannot explain variation in their direction. Yet, several studies reveal that the direction of relations between a value and an attitude does sometimes reverse (e.g., Barni, Vieno, & Roccato, 2016; Hanel, Vione, Hahn, & Maio, 2017; Ponizovskiy, Grigoryan, Kühnen, & Boehnke, 2019). Roccas and Sagiv (2010) argued that differences in meaning of the attitude object could drive variation across cultures in value–attitude relations. They refer to such variation as “particularly elusive.” (p. 37) The current study posits that the meaning of the attitude to cultural diversity differs depending on the position of groups in a diverse society. This enables us to explore this elusive phenomenon.
To summarize, this study focuses on the relations of values to the attitude to cultural diversity. It goes beyond the existing literature on moderators of value–attitude relations by examining two moderators that have received little prior attention—level of cultural diversity and its meaning.
Theoretical Overview
Basic Human Values
We adopt the refined theory of basic values for this study (Schwartz et al., 2012). The theory defines values as “desirable trans-situational goals, varying in importance, that serve as guiding principles in the life of a person or other social entity” (Schwartz, 1994, p. 21). As goals, values motivate behavior and predict attitudes. The refined theory of basic values specifies 19 values that form a circular motivational continuum (Schwartz et al., 2012).
Figure 1 presents the motivational continuum of 19 values in the inner circle. Values that express compatible goals are near one another in the circle, whereas values that express conflicting goals are distant from one another. The circular band around the values groups them into four higher order values that represent two bipolar dimensions of motivationally incompatible, higher order values. Openness to change values emphasize independence and readiness for new experiences and ideas. They oppose conservation values that emphasize order, stability, and preservation of traditions. Self-transcendence values emphasize acceptance of and concern for the welfare of others. They oppose self-enhancement values that emphasize self-interested success and dominance over other people and resources. These motivationally opposed higher order values represent the poles of two dimensions, desire for stability versus change (the conservation vs. openness to change dimension) and concern for self-interest versus others’ welfare (the self-enhancement vs. self-transcendence dimension). This study, like many others (e.g., Barni et al., 2016; Verkasalo, Lönnqvist, Lipsanen, & Helkama, 2009), examines how the two dimensions predict specific attitudes.

Circular motivational continuum of 19 values.
Values as Predictors of the Attitude Toward Cultural Diversity
Values are more abstract, fundamental, and stable than attitudes (Homer & Kahle, 1988; Schwartz, Caprara, & Vecchione, 2010). Values underlie and affect specific attitudes towards concrete objects and situations (e.g., Rokeach, 1968; Schwartz, 2015). We make the assumption, supported by value–attitude studies that, if the object of an attitude (e.g., cultural diversity) promotes/blocks the expression of a value, that value is likely to positively/negatively predict the attitude (e.g., Sagiv & Schwartz, 1995).
Following Cox (1993, p. 6), we define cultural diversity as the presence “in one social system, of people with distinctly different group affiliations of cultural significance.” For the Russian context, we operationalize this as differences in ethnicity. Accordingly, positive/negative attitudes towards cultural diversity signify positive/negative evaluations of the presence of different ethnic groups in the environment. Attitudes towards diversity largely overlap with attitudes to immigrants and to other ethnic outgroups. This allows us to base some of our hypotheses on the rich literature on intergroup relations. We briefly note a few of the studies that illustrate this relationship in relevant domains.
In representative national samples from 21 European Social Survey (ESS) countries, self-transcendence values correlated positively, and conservation values negatively, with openness to diversity (Vala & Costa-Lopes, 2010). In a student sample from the US, openness to change and self-transcendence values correlated positively with attitudes to diversity (Sawyerr, Strauss, & Yan, 2005). In 15 ESS West European countries, universalism values were the strongest positive correlate of acceptance of immigrants (Schwartz, 2007). In 17 ESS countries, and in a replication in 20 countries at three time points, self-transcendence values correlated positively, and conservation values negatively, with support for immigration (Davidov & Meuleman, 2012; Davidov, Meuleman, Billiet, & Schmidt, 2008). Finally, power and security values correlated positively, and universalism and benevolence values negatively, with prejudice toward Aboriginal Australians (Feather & McKee, 2008).
Studies of attitudes to other minority groups, or generalized group-focused enmity, found similar associations. In 27 ESS countries, openness to change and universalism values correlated positively, and conservation and power values correlated negatively, with approval of homosexuality (Kuntz et al., 2015). In a German sample, conservation values correlated positively, and universalism values correlated negatively, with group-focused enmity (including anti-immigrant attitudes, sexual/gender prejudice, anti-Semitism, and devaluation of homeless people; Beierlein, Kuntz, & Davidov, 2016).
All these studies demonstrate that two dimensions of the motivational circle of values drive attitudes to outgroups. Individuals tend to have more negative attitudes toward outgroups if they are higher on the self-enhancement versus self-transcendence dimension, and on the conservation versus openness to change dimension. Critically, however, all of these findings refer to the attitudes of majority-group members toward minority groups. Moreover, none of the studies considered contextual moderators, except for Kuntz et al. (2015), who studied the normative expectations mechanism. The current study innovates by investigating both majority- and minority-group attitudes, and by considering moderators of value–attitude relations.
Moderators of Value–Attitude Relations
The meaning of cultural diversity
Differences in the meaning of attitude objects are reflected in differences in the perceived consequences of these objects for values. For example, Sagiv and Schwartz (1995) reported that differences in the consequences of outgroup contact for the Jewish majority and Arab minorities in Israel resulted in differences in associations between values and readiness for outgroup contact. Specifically, associations of stimulation and self-direction values with the desire for outgroup contact were positive for the Jewish majority, but negative or null for the Muslim Arab minority. This reflected the groups’ different opportunities for self-expression in the usual contact settings. We argue that the consequences of cultural diversity for values may also differ depending on groups’ majority versus minority, and high versus low, status.
Evidence suggests that groups differ in how they interpret cultural diversity. In a comprehensive review of race relations, Richeson and Sommers (2016) describe differences in perceptions of cultural diversity between White Americans and minority groups in the US. For White Americans, the greater the percentage of racial minority members in a context, the more diverse they perceive the context to be. In contrast, for minorities, the greater the percentage of ingroup members in a context, the more diverse they perceive it to be. This suggests that for majority-group members, diversity refers exclusively to outgroup members, whereas for minority-group members, diversity includes the ingroup.
Van Knippenberg and Schippers (2007) use the term “diversity mind-sets” to refer to people’s understanding of diversity and its consequences. They argue that this understanding may moderate social categorization and decision-making processes that take place in diverse environments. In this study, we consider three groups with different standpoints regarding cultural diversity that may lead to different mindsets. The standpoints are based on whether the group is the majority or minority, and whether it has high versus low status in the societal hierarchy. We postulate that differences in how groups interpret diversity (i.e., its meaning for them) lead to differences in how their values predict their attitude to cultural diversity.
Level of cultural diversity
We propose that the level of cultural diversity in a context moderates the relations of values to attitudes toward cultural diversity. We could find no studies that considered level of cultural diversity as a moderator. As Roccas and Sagiv (2010) argued, diversity may either enhance value–attitude relations or weaken them. This depends on whether diversity promotes individual freedom of choice or whether it increases group identification and normative control. These theoretical possibilities provide the rationale for investigating how, if at all, the cultural diversity of a context moderates value–attitude relations.
The Current Study
Russia, the site of the current study, is a particularly interesting context for studying attitudes towards cultural diversity because of the presence of so-called “titular” ethnic groups. In 1917, the Soviet government introduced the term “titular nation.” It granted some of the ethnic groups extended rights and named administrative units where they predominantly lived after them. Titularity was excluded from the Russian constitution in the 1990s, but such administrative units still exist as national republics. They have substantial autonomy, including the right to have their own constitution and legislature, and their own official languages in addition to Russian. The titular groups hold most of the political power in their republic. They have higher status and decision-making power than other indigenous minorities or immigrants. The inequality of status created by titularity gives a unique opportunity to disentangle the effects of being a numerical minority in the country from the effects of having higher political status (Alexseev, 2010). In this study, we investigate three groups that differ along two dimensions relevant to cultural diversity: majority (Russians) versus minorities (other titular and nontitular ethnic groups), and high status (Russians and titulars) versus low status (nontitulars).
We postulate that the meaning of cultural diversity differs across the three groups along these two dimensions. Majority versus minority status affects the perception of whether the ingroup is part of that diversity or not (Richeson & Sommers, 2016). Minorities see themselves as part of diversity, whereas the Russian majority does not. Status differences affect the perception of threat in intergroup contexts. Studies of intergroup relations consistently find that lower group status is associated with higher perceived threat (e.g., Stephan et al., 2002). Figure 2 summarizes the factors likely to affect the meaning of cultural diversity in the three groups we studied.

Meaning of cultural diversity as a function of majority/minority identity and high/low status.
Hypotheses
Consistent findings in previous studies of the relations between values and attitudes to diversity and to outgroups suggest two predictions. The importance of self-enhancement versus self-transcendence values and of conservation versus openness to change values should negatively predict attitudes towards cultural diversity. Previous studies included only majority-group members. However, based on the moderating effect of group membership that we discuss in what follows, we expect variations from the usual pattern of correlations in minority groups.
Meaning of cultural diversity as a moderator of value–attitude relations
We develop hypotheses for each group by considering the meaning of cultural diversity, that is, its consequences for group values. The relationship of attitude toward diversity with the conservation/openness value dimension should reflect the meaning of cultural diversity as either including the ingroup or excluding it. This value dimension is the primary driver of ingroup attitudes, with higher conservation relating to stronger ingroup loyalty (Boer & Fischer, 2013). The Russian ethnic majority likely understands attitude toward diversity as attitude toward outgroups. Outgroup attitudes typically relate negatively to conservation values. Hence, we expect value–attitude relations to replicate those previously found for majority groups in other contexts. That is, we expect attitude to diversity to relate negatively to the conservation/openness value dimension.
In contrast, the two ethnic minority groups may understand cultural diversity as including their own group along with others. Hence, for them, attitude to diversity should positively relate to the conservation/openness value dimension. We therefore hypothesize that attitude to cultural diversity correlates negatively with the conservation/openness value dimension for the Russian majority, but positively for the two minority groups (H1). This implies that there will be no association between attitude toward diversity with the conservation/openness dimension in the total study sample.
Despite the minorities’ perception of cultural diversity as including the ingroup, diversity poses threats to the status quo for all. Social status differences between groups may intensify or mitigate the threat that is imposed by the presence of dissimilar others in a diverse society. For low-status ethnic minorities, the presence of dissimilar others intensifies the threat because it highlights the social inferiority of their position that makes it difficult for them to escape or resist domination by others. For the Russian majority and the high-status titular ethnic groups, their socially superior position enables them to dominate others in their environment and may minimize economic or cultural threats that the presence of different “others” may bring. We therefore hypothesize that the negative relationship between attitude to cultural diversity and the self-enhancement/self-transcendence value dimension is stronger for low-status ethnic minorities than for the Russian majority and high-status ethnic minorities (H2).
Cultural diversity in the region as a moderator of value–attitude relations
Level of cultural diversity may moderate value–attitude relations in contradictory ways. The presence of multiple ethnic groups with diverse norms and practices might strengthen value–attitude relations because it may loosen the normative demands and expectations prevalent in the environment. This would allow individuals more freedom to express their own values. Conversely, the presence of multiple ethnic groups might weaken value–attitude relations by making intergroup differences more salient. This may increase identification with one’s ingroup and lead to greater adherence to ingroup norms rather than to own values (Roccas & Sagiv, 2010). No empirical research has yet assessed these possibilities. We therefore address this question by testing whether associations between values and attitude differ depending on the level of cultural diversity in the region.
Controls
We controlled for age, gender, level of education, perceived financial strain, and level of religiosity. These variables have been associated both with values and with attitudes to outgroups in the literature (e.g., Billiet, 1995; Davidov et al., 2008; Schwartz, 2007; Semyonov, Raijman, & Gorodzeisky, 2006).
Method
Sample
Power considerations
We conducted an a priori power analysis in G*Power to determine the sample size needed for the regression analyses. We used an effect size of f2 = 0.009, an average effect size for interaction effects in psychology (Aguinis, Beaty, Boik, & Pierce, 2005). The analysis indicated that to achieve a power of 0.95 with α = .05, a total sample of N = 1,446 was required.
We could not calculate a precise estimate of power to find an interaction effect in the multilevel model because of the large number of unknown parameters in the population (e.g., ICC, slope variance). We therefore used the common criterion for satisfactory power in multilevel models, a minimum of 20 second-level units (Kreft & De Leeuw, 1998). The recommended sample for examining a cross-level interaction is 30 groups with 30 observations within each group (Kreft, 1996 as cited in Hox, 2010). We had 25 regions at the second level and an average of 77 participants per region. Hence, our data should provide sufficient power to find an interaction effect.
Participants
A stratified random sample, representative of the Central (N = 1,004) and North Caucasus (N = 930) federal districts of Russia, responded to a face-to-face survey in their homes during 2012. Sampling and data collection were carried out by the Institute for Comparative Social Research (CESSI, www.cessi.ru). The sample covered 25 randomly selected regions, including six national republics. It included 31 ethnic groups plus a category of “other.” Respondents were aged 18 to 60 years, and 56% were female. All interviews were conducted in Russian. 1 The measures reported in this study are part of a larger survey; other measures obtained in this survey are reported elsewhere (see Khaptsova & Schwartz, 2016; Tatarko, Mironova, & van de Vijver, 2017; Tatarko & Schmidt, 2016). 2
Table 1 presents sociodemographic characteristics of the three groups, and Supplement A in the online supplemental material presents characteristics of the 25 regions. We classified the titular ethnic groups in each national republic as “high-status ethnic minority” (groups marked with an asterisk in Supplement A). We classified all other non-Russian ethnic groups as “low-status ethnic minorities.” We excluded 54 Russians who lived in one of the national republics where they constitute a minority.
Sample characteristics of three groups.
Measures
Values
We employed the revised version of the Portrait Value Questionnaire (PVQ-R) from Schwartz et al. (2012). This PVQ version includes 57 value items (three for each of the 19 values), with a 6-point response scale (1 = not like me at all, 6 = very much like me). Only those items that clearly fall on one of the two value dimensions were included in the analysis. Hence, we excluded hedonism, face, and humility value items. Reliability scores for higher order values in the three groups varied from .71 to .75 for openness to change (nine items), from .86 to .88 for conservation (15 items), from .85 to .87 for self-transcendence (15 items), and from .79 to .83 for self-enhancement values (nine items). Mean reliability scores across the 25 regions varied from .67 for openness to change to .80 for conservation values. We measured the self-enhancement/self-transcendence dimension by subtracting the self-transcendence score from the self-enhancement score, and the conservation/openness dimension by subtracting the openness to change score from the conservation score. Supplement C presents scale properties for values in each of the 25 regions, and invariance tests for the three groups.
Attitude to cultural diversity
We measured attitude to diversity by averaging responses to two items from Onyx and Bullen (2000): “If people of different ethnicities live near each other it makes life better” and “I don’t mind living among people of different ethnicities.” Respondents answered on a 5-point scale (1 = strongly disagree, 5 = strongly agree). Cronbach’s alphas for the attitude measure ranged from .70 to .98, with a mean of .84 across the 25 regions, and from .86 to .89 across the three groups. Supplement B presents scale properties for each of the 25 regions, and invariance test of the scale for the three groups.
Age
Computed from year of birth.
Gender
Coded by interviewers: male = 1, female = 2.
Level of education
Response to “What is the highest level of education you have completed?” (1 = basic secondary education, 11 = academic degree Stage II-PhD; European Social Survey [ESS], 2018).
Perceived financial strain
Response to “Which of the descriptions on this card comes closest to how you feel about your household’s income nowadays?” (1 = living comfortably on present income, 2 = coping on present income, 3 = difficult on present income, 4 = very difficult on present income; ESS, 2018).
Religiosity
Response to “How religious would you say you are?” (0 = not at all religious, 10 = very religious; ESS, 2018).
Cultural diversity index (CDI)
We measured level of cultural diversity at the regional level with the ethno-linguistic fractionalization index (Mauro, 1995; Taylor & Hudson, 1972): CDI =
Analyses
We tested H1–H2 with regression analyses with robust maximum likelihood estimators and Wald chi-square test for parameter estimates. To make the results comparable with previous studies on value–attitude relations, we conducted a hierarchical regression analysis. 3 In the first step, we included only the value dimensions as predictors. We then added group membership to the model, with Russians as the reference group, and the interaction term. In the final step, we included all the control variables to check the robustness of the effects.
We tested the moderating effect of level of cultural diversity on value–attitude relations with multilevel regression. We used restricted maximum likelihood (REML) estimation for second-level sample sizes smaller than 50, as recommended by Hox (2010). We controlled for age, gender, education, religiosity, and perceived financial strain in all analyses.
Results
Group membership as a moderator of value–attitude relations
Table 2 presents the results of the regression analysis. Main value effects were consistent with expectations. The self-enhancement/self-transcendence value dimension negatively predicted attitudes, whereas the conservation/openness value dimension had no significant effect. Group membership moderated the link between values and attitude, as predicted. Figure 3 presents the interactions.
Hierarchical regression models predicting attitude to cultural diversity from values (unstandardized regression coefficients).
Note. C/O = conservation versus openness to change, SE/ST = self-enhancement versus self-transcendence.

Relations between value dimensions and attitude to cultural diversity as a function of group membership.
Consistent with H1, conservation/openness related negatively to attitude to diversity among Russians, but positively among ethnic minorities (see Figure 3a). Self-enhancement/self-transcendence related negatively to attitude to diversity in all three groups. However, consistent with H2, the relationship was significantly more negative among low-status minorities (see Figure 3b). Controlling for possible confounds did not change the interaction effects. Of the control variables, only religiosity and perceived financial strain predicted attitudes. Religious people held more positive attitudes toward diversity, whereas people who perceived more financial strain held more negative attitudes toward diversity.
To determine whether one or both poles of the value dimensions are driving the associations between value dimensions and attitude, we additionally tested the model with four higher order values separately (see Table E in the supplemental material). On the conservation/openness dimension, there was a significant interaction between group membership and conservation values, χ2(2) = 6.28, p = .043, but there was no interaction with openness values, χ2(2) = 0.80, p = .671. The conservation–attitude link was more positive among high-status ethnic minorities than among Russians (b = 0.27, SE = 0.12, p = .020). On the self-enhancement/self-transcendence dimension, there was a significant interaction between group membership and self-enhancement values, χ2(2) = 8.03, p = .018, but there was no interaction with self-transcendence values, χ2(2) = 5.76, p = .056. The negative self-enhancement–attitude link was stronger among low-status ethnic minorities than among Russians (b = −0.16, SE = 0.08, p = .042).
These results support our theorizing regarding group differences in the way cultural diversity is construed. Minority groups perceive cultural diversity as including the ingroup; therefore, diversity does not hinder the attainment of conservation goals for them, as it does for the Russian majority. Threats to status and power posed by cultural diversity, on the other hand, are more strongly perceived among those with lower standing in the societal hierarchy. Hence, we observe a stronger negative relationship between self-enhancement values and attitudes to diversity among low-status minorities.
Cultural diversity as a moderator of value–attitude relations
Intraclass correlation indicated that 78% of the variance in attitude toward diversity was within regions and 22% between regions. We therefore used multilevel analysis to take the nesting of individuals within regions into account. We centered all individual-level variables around their group means, and contextual variables around the grand mean (Enders & Tofighi, 2007). Table 3 presents the results of a sequential series of nested models.
Multilevel regression models predicting attitude to cultural diversity (unstandardized regression coefficients).
Note. SE/ST = self-enhancement versus self-transcendence; C/O = conservation versus openness to change; CD Index = cultural diversity index. All individual-level predictors are group mean centered and group-level predictors are grand mean centered.
Reduction in variance compared with the empty model (residual variance = 0.73***; intercept variance = 0.20**). bImprovement in model fit compared to empty model (−2 log likelihood = 4911.78, df = 2).
p < .001. **p < .01. *p < .05.
Model 1 included all the individual-level variables: gender, age, education, religiosity, perceived financial strain, group membership, and individual values. These variables explained 37.2% of the between-group variance and 2.9% of the individual-level variance in attitudes. The effects of the control variables were consistent with the previous analysis. Both self-enhancement/self-transcendence and conservation/openness dimensions negatively predicted attitude to diversity. Model 2 added the cultural diversity index as a group-level variable. This explained 11.7% of additional variance in attitudes at the group level. The greater the cultural diversity in a region, the more positive the attitude to cultural diversity.
In Models 1 and 2, all slopes were fixed and only the intercept was random. Model 3 tested whether the slopes of value–attitude relations varied across the 25 regions. None of these slopes varied significantly across regions. The 95% CIs for the variance in slopes were [0.004, 0.06] for the self-enhancement/self-transcendence dimension (Wald Z = 1.46, p = .143), and [0.01, 0.13] for the conservation/openness dimension (Wald Z = 1.27, p = .205). Model 4 tested the cross-level interactions of the two value dimensions with level of cultural diversity in the regions. None of the interactions was significant. The 95% CIs for the interactions were [−0.39, 0.39] with the self-enhancement/self-transcendence dimension, and [−0.76, 0.40] with the conservation/openness dimension.
We additionally tested Model 4 replacing the value dimensions with the four higher order values. The results remained unchanged. None of the value–attitude slopes showed significant variation across regions (self-transcendence: Wald Z = 0.75, p = .451; self-enhancement: Wald Z = 1.30, p = .193; openness: Wald Z = 1.12, p = .262; conservation: Wald Z = 1.96, p = .050). Similarly, no interactions were found between the four values and the cultural diversity index (self-transcendence: b = 0.07, SE = 0.33, p = .834; self-enhancement: b = 0.04, SE = 0.20, p = .826; openness: b = 0.04, SE = 0.29, p = .886; conservation: b = −0.41, SE = 0.51, p = .438). Value–attitude associations were similar across regions and did not vary by level of cultural diversity.
Discussion
Research on value–attitude and value–behavior relations is shifting from identifying trends across contexts to searching for moderators of these relations. Moderator studies have considered two broad types of contextual factors. One type refers to factors that influence the meaning of attitudes or values (Ponizovskiy et al., 2019; Sagiv & Schwartz, 1995; Schwartz & Sagiv, 1995). The other refers to environmental conditions that support or suppress expression of values (Bardi & Schwartz, 2003; Boer & Fischer, 2013; Kuntz et al., 2015). We investigated one example of each set of moderators.
The first part of our study examined relationships between individual values and attitude to diversity in three groups whose standpoints might lead them to attribute somewhat different meanings to cultural diversity. We theorized that the two minority groups would understand cultural diversity as referring not only to outgroups but to their own ingroup as well. We also theorized that low-status minorities would experience cultural diversity as more threatening than higher status minorities and majority Russians would.
We further reasoned that the relationships between value dimensions and attitude would reflect these differences in meaning of cultural diversity for the three groups. We hypothesized that the relationship between self-enhancement (vs. self-transcendence) values and attitude to diversity would be most negative among low-status ethnic minorities because their weaker position in the societal hierarchy makes it difficult for them to attain the goals of self-enhancement values. This hypothesis received full support. The negative relationship between self-enhancement values and attitude was twice as strong for low-status minorities as compared to ethnic majority Russians.
The overall negative effect of the self-enhancement (vs. self-transcendence) value dimension replicates previous studies that investigated relations of values to outgroup attitudes. They found that motivational conflict between power values and universalism values was the most powerful predictor of outgroup attitudes (Feather & McKee, 2008; Kuntz et al., 2015; Schwartz, 2007). The current study innovates by identifying group status as a moderator of this effect. This study illustrates how contextual factors that facilitate or interfere with attaining goals relevant in intergroup contact can affect value–attitude relations.
In past studies, conservation values correlated consistently negatively with attitudes to outgroups (e.g., Beierlein et al., 2016; Davidov et al., 2008; Vala & Costa-Lopes, 2010). We expected the negative association of conservation (vs. openness to change) values with attitude toward diversity to replicate only for the Russian majority but not for the two minority groups. Our findings were consistent with these expectations: the link was negative for Russians but positive for the minorities, yielding a nonsignificant overall effect and an interaction.
Although the associations of conservation values followed the hypothesized pattern, the effects were very weak in all three groups. These weak findings in all groups suggest that a Russia country-level characteristic might be implicated. The high disease stress level in Russia relative to other countries in which this association has been studied 4 may be the critical characteristic implicated (cf. Thornhill & Fincher, 2014). Value–attitude associations in the domain of conservation-motivated attitudes were weaker in countries with high disease stress in a large meta-analysis (Boer & Fischer, 2013). Parasite stress theory argues that societies cope with high disease stress by emphasizing conservation values and practices (Thornhill, Fincher, & Aran, 2009). In such societies, conservation values are a social imperative rather than an individual personal choice. This leads to weaker and less consistent relations between conservation values and various conservative attitudes (Boer & Fischer, 2013).
Level of cultural diversity in the region was the second moderator we examined. Cultural diversity had a positive main effect on attitudes to cultural diversity. This is consistent with the intergroup contact hypothesis (Pettigrew & Tropp, 2006), because greater cultural diversity increases the probability of intergroup contact. More interesting were the conflicting predictions for an interaction between values and level of cultural diversity in their effect on attitudes. These predictions built on the assumption that greater cultural diversity may either weaken or strengthen the normative demands of the environment, thereby either weakening or strengthening value–attitude relations (Roccas & Sagiv, 2010).
We found no significant interaction effect; neither prediction received support. Of course, absence of evidence is not evidence of absence (Lakens, 2017). We therefore do not view this finding as conclusive evidence against moderation of values–attitude associations by level of cultural diversity. A more definitive test of this hypothesis might require more groups at the second level of analysis that provide more variance in value–attitude relations. Furthermore, we cannot exclude the possibility that each of the conflicting effects (positive and negative) of level of cultural diversity occurred in some of the over 30 ethnic groups studied. If the direction of the effect of cultural diversity is conditional upon group membership, the effect in the overall sample might have been cancelled out. Uncovering this possibility will require theorizing about conditions that might promote one or the other effect.
Most past studies of moderators of value–attitude links have focused on differences in the environment. The environment difference we studied (level of cultural diversity in the regions) had no moderating effect. However, the current research found that postulated differences in the meaning of the attitude object, based on the majority versus minority and the high versus low status of the groups, explained variation and even a reversal in value–attitude links. This demonstration that group differences in the meaning of attitude objects matter for value–attitude links identifies a fruitful direction for future research.
Supplemental Material
Online_supplements – Supplemental material for Values and attitudes towards cultural diversity: Exploring alternative moderators of the value–attitude link
Supplemental material, Online_supplements for Values and attitudes towards cultural diversity: Exploring alternative moderators of the value–attitude link by Lusine Grigoryan and Shalom H. Schwartz in Group Processes & Intergroup Relations
Footnotes
Authors’ note
At the time of data collection, the authors were associated with the National Research University Higher School of Economics (HSE), Moscow, Russia. This work is an output of a research project implemented as part of the Basic Research Program at the National Research University HSE.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Data collection for this study was supported within the framework of a subsidy by the Russian Academic Excellence Project ‘5-100’.
Supplemental material
Supplemental material for this article is available online.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
