The Impact of Ethnic Diversity on Participation in European Voluntary Organizations

Abstract

In this study, we test whether Putnam’s general claim of a negative effect of ethnic diversity holds for (active and passive) involvement in three types of voluntary organizations: leisure, interest, and activist organizations. Using data from the European Social Survey (wave 1), we applied multilevel analyses distinguishing individuals, regions, and countries. Only at the regional level, did we find that ethnic diversity reduced involvement in interest organizations. Yet, ethnic diversity induced passive involvement in activist organizations. Subsequently, we included mechanisms derived from conflict and contact theory to disentangle the indirect effects of ethnic diversity. Ethnic threat perceptions influenced participation in all voluntary associations negatively, while intergroup contact turned out to have positive influences. Our findings stress the necessity of distinguishing different types of voluntary organizations and modes of involvement and underline the importance of incorporating indirect effects of ethnic diversity.

Keywords

ethnic diversity formal social capital associational involvement intergroup contact perceived ethnic threat

Introduction

Increasing migration and its expected future rise (Cornelius & Rosenblum, 2005; Hooghe, Trappers, Meuleman, & Reeskens, 2008) made politicians and scholars increasingly interested in the consequences of ethnic diversity for social cohesion (e.g., Cheong, Edwards, Gouldbourne, & Solomos, 2007). Recently, attention for this topic intensified because research findings from the United States indicated that people living in ethnically diverse settings tend to withdraw from social life, or to “hunker down” (Putnam, 2007, p. 149). Putnam claimed this pattern to be visible in attitudes and behavior, covering informal and formal social ties as well as bridging and bonding social capital. In this study, we will focus on the consequences of living in an ethnically diverse environment for the likelihood to be involved in voluntary organizations, often referred to as formal social capital (cf. Pichler & Wallace, 2007). As this dimension reflects less strong ties than informal social capital (i.e., informal social ties with family members, friends, or colleagues), one might expect ethnic diversity to most strongly influence formal social capital.¹

Earlier research found inconsistent effects of ethnic diversity on associational involvement. Whereas some scholars found negative effects (e.g., Alesina & La Ferrara, 2000; Putnam, 2007; Rotolo & Wilson, 2012), others found no effect (e.g., Gijsberts, Van der Meer, & Dagevos, 2012; Letki, 2008) or even a positive relationship (e.g., Gesthuizen, Van der Meer, & Scheepers, 2009; Kesler & Bloemraad, 2010). For a review of studies, see Wilson (2012). The emerging picture appears to be less conclusive than claimed by Putnam. So far, however, the research mainly has focused on rather general indicators of associational involvement (e.g., Kesler & Bloemraad, 2010; Letki, 2008; Putnam, 2007; Tolsma, Van der Meer, & Gesthuizen, 2009), without systematically differentiating modes of involvement (i.e., active and passive involvement) and not considering that organizations differ strongly in the goals they pursue. Our aim is to test the generalizability of Putnam’s claim, distinguishing different types of voluntary organizations and different modes of involvement. We build on Van der Meer and colleagues’ study, which identified three types of voluntary organizations based on their main goals: leisure organizations (offering socializing and recreational activities to their members), interest organizations (protecting the socioeconomic interests of their members), and activist organizations (with activist goals, advocating broader societal interests; Van der Meer, Te Grotenhuis, & Scheepers, 2009). For each type, we will distinguish between active involvement (e.g., active participation or volunteering) and passive involvement (e.g., donating money).

In addition, we will go one step further, by disentangling underlying mechanisms to increase our understanding of how ethnic diversity relates to formal social capital. Previous research (e.g., Gijsberts et al., 2012; Letki, 2008; Putnam, 2007) only addressed the direct relationship between ethnic diversity and involvement in voluntary organizations, leaving unaddressed the indirect effects of ethnic diversity at the macro level on individual involvement in voluntary associations. These indirect effects are theoretically interesting and empirically important. If ethnic diversity has contradictory indirect effects on people’s involvement in voluntary organizations, these effects may cancel each other out, thus producing an overall absent direct effect of ethnic diversity. Putnam (2007) referred to conflict as well as contact theory, which both suggest different mediating mechanisms. He did not, however, elaborate on nor empirically test the proposed indirect relationships. Recently, Savelkoul, Gesthuizen, and Scheepers (2011) showed that both theories are important to explain the relationship between ethnic diversity and informal social capital, as perceptions of ethnic threat and interethnic contact are related to informal social ties. We will use both theories to derive expectations regarding the relationship between ethnic diversity and formal social capital and test the individual-level mechanisms empirically.

We will build on earlier studies by distinguishing ethnic diversity at the country as well as the regional level simultaneously. Previous research on the relationship between ethnic diversity and formal social capital mainly focused on the municipality or neighborhood level (e.g., Letki, 2008; Putnam, 2007; Tolsma et al., 2009). As it is difficult, or even impossible, to find valid data at these levels for all countries in cross-national research, these studies have only been conducted within single countries. Cross-national studies are rather scarce and do not take into account variation within countries (e.g., Gesthuizen et al., 2009; Kesler & Bloemraad, 2010). Although participation in voluntary organizations often pertains to local participation, we propose that the regional level might be of interest as well. Social life will partly take place outside people’s direct neighborhood or even outside people’s municipality. Not only may people work outside their municipality, they may compete against teams from other municipalities in a sporting match or participate in regional interest organizations. Simultaneously, other mechanisms might be at work at the country level: The composition of ethnic minority groups might differ across countries and perceptions of the present out-groups across countries might vary as a result of differences in media coverage on ethnic minorities (Ter Wal, 2002). By simultaneously taking into account the regional and the country level, we are able to test the effect of ethnic diversity more accurately.²

Using data from the European Social Survey (ESS, wave 1, 2002/2003), supplemented with data on both contextual levels, we will address two research questions:

Research Question 1: To what extent does ethnic diversity within (a) European countries or (b) regions across European countries affect involvement in leisure, interest, and activist organizations?

Research Question 2: To what extent can the understanding of the relationship between ethnic diversity and involvement in voluntary organizations be improved by mechanisms derived from conflict and contact theory?

Theories and Hypotheses

We will set out to explore three general theories—constrict, conflict, and contact theories—which propose contradictory effects of ethnic diversity on formal social capital.

Constrict Theory

Putnam’s (2007) “constrict theory” states that ethnic diversity reduces social capital: Ethnic diversity triggers “[ . . . ] anomie or social isolation,” fostering people to withdraw from social life, or as Putnam formulated it “[ . . . ] pull in like a turtle” (Putnam, 2007, p.149). He proposed that this negative impact of ethnic diversity is rather general. However, his theoretical reasoning behind this effect remains rather implicit. We suspect that the line of thought underlying his statement might be related to the homophily principle (McPherson, Smith-Lovin, & Cook, 2001): People prefer homogeneous environments, with people who are alike, for instance regarding their ethnicity. As proposed by Gesthuizen et al. (2009), in ethnically diverse contexts, there will be less people of one’s own kind with whom one feels familiar with. As a result, people will feel less comfortable with others and therefore will more likely withdraw from social life, including voluntary associations. Based on Putnam’s general claim, we expect the following:

Hypothesis 1: Ethnic diversity within (a) European countries as well as (b) regions within these countries will reduce the likelihood that people will be involved in leisure, interest, and activist organizations.

Associational involvement can take different forms: People can be actively or passively involved. Whereas active involvement refers to volunteering or active participation and, thus, reflects high(er) levels of face-to-face contacts, passive involvement pertains to donating money with no opportunities for face-to-face contact. Putnam’s constrict theory proposes that ethnic diversity will generally reduce people’s likelihood to be involved in voluntary organizations. However, if living in ethnically diverse environments makes people to feel less comfortable with others (Gesthuizen et al., 2009), they might predominantly avoid face-to-face contacts, that is, active involvement, while passive involvement will be less strongly affected.

Simultaneously, previous research shows that different types of organizations vary with respect to the most prevalent mode(s) of involvement. Whereas leisure organizations are characterized by relatively high levels of active involvement with many people having face-to-face contact, activist organizations are characterized by passive modes of involvement, with many people donating money, without having face-to-face contacts (i.e., “checkbook members”; Gesthuizen, Scheepers, Van der Veld, & Völker, 2013; Putnam, 2000; Van der Meer et al., 2009, see also descriptive statistics, Table A1, Appendix).³ Interest organizations can be positioned in between both other types of organizations, displaying lower levels of actively involved members as compared with leisure organizations. Hence, we propose that the negative influence of ethnic diversity on associational involvement is most likely for organizations predominantly characterized by active modes of involvement, that is, leisure organizations. In sum, we expect the following:

Hypothesis 2: (a) Ethnic diversity within European regions will predominantly reduce the likelihood that people will be actively involved in voluntary organizations and (b) stronger for leisure organizations than for interest and activist organizations.⁴

In constrict theory, the focus is on the direct influence of ethnic diversity on associational involvement, without addressing underlying explanations. Although Putnam did not elaborate on these indirect effects, he mentioned two theories that are important in this respect: conflict and contact theory.

Conflict Theory

Conflict theory is based on realistic group conflict theory (Blalock, 1967; Bobo, 1999; Coser, 1956) and ethnic competition theory (Scheepers, Gijsberts, & Coenders, 2002). According to conflict theory, ethnic diversity fosters actual competition between the ethnic majority group and ethnic minority groups over scarce resources or cultural values, thus inducing perceptions of ethnic threat among members of the (majority) in-group. Previous research showed that these perceptions of ethnic threat are an important determinant of out-group derogation (e.g., Scheepers et al., 2002; Schlueter & Wagner, 2008). We propose that perceptions of ethnic threat might explain why people living in ethnically more diverse environments will withdraw from social life. People perceiving ethnic threat will be more anxious to become confronted with unknown others in general, and with ethnic minority members within voluntary organizations in particular. Consequently, they will be less likely actively involved in these organizations. In addition, they might be less inclined to become passively involved in organizations with ethnic out-group members, or in organizations (also) promoting the interests of ethnic minorities. Summarizing, we expect the following:

Hypothesis 3: (a) Ethnic diversity within European regions will increase people’s level of perceived ethnic threat in that region, (b) which in turn will reduce the likelihood that people will be involved in leisure, interest, and activist organizations.

Contact Theory

The second theory to which Putnam (2007) referred, is intergroup contact theory (Allport, 1954/1979; Pettigrew & Tropp, 2011), which boils down to the proposition that interethnic contact effectively reduces out-group derogation. Previous research repeatedly showed that ethnic diversity increases the likelihood of intergroup contacts, which in turn reduce levels of out-group derogation (e.g., Schlueter & Scheepers, 2010; Schlueter & Wagner, 2008). Moreover, ethnic diversity may also increase intergroup contacts through which social networks might become larger and more diverse. According to Wilson (2000), the larger the social network is quantitatively, the more likely people will become (actively or passively) involved in voluntary organizations. As intergroup contact involves different role models and examples of spending (leisure) time, people’s social networks will also become qualitatively different. This nicely links with Pettigrew’s (1998) deprovincialization effect: Intergroup contact reduces provincial views of the social world and contact with ethnic minorities widens one’s perspective, thereby increasing empathy, also for people belonging to other ethnic groups. Hence, intergroup contact might foster people’s awareness of general problems and, consequently, increase the likelihood of becoming (actively or passively) involved in voluntary organizations addressing these problems. Moreover, having intergroup contact might reduce people’s anxiety to become confronted with ethnic minorities in voluntary associations (Pettigrew & Tropp, 2008), which might take away a psychological threshold to become actively involved in such organizations. Therefore, we hypothesize the following:

Hypothesis 4: (a) Ethnic diversity within European regions will increase the likelihood that people living in these regions have intergroup contact. Moreover, we expect that (b) intergroup contact, in turn, will increase the likelihood that people will be involved in leisure, interest, and activist organizations.

Figure 1 shows our theoretical framework. The numbers refer to our hypotheses.

Figure 1.

Theoretical framework relationship between ethnic diversity and formal social capital.

Data and Measurement

Data

We used data derived from the first wave of the ESS (wave 1, 2002/2003; Jowell & The Central Co-ordinating Team, 2003). This dataset contains a fine-grained measurement of associational involvement across a fairly large number of European countries and regions. Samples were drawn randomly for 21 European countries and Israel. The data were collected by face-to-face interviews with people living in private households, aged 15 years and above.

We only selected European countries for which ethnic diversity at the regional level could be measured. We used a country-specific indicator to group respondents into regional units that correspond to the Nomenclature of Statistical Territorial Units classification scheme (NUTS; see Eurostat, 2003). The NUTS-2 level was the smallest regional level available for a large number of countries. This level refers to medium scale regions (ranging from 800,000 to 3 million inhabitants) and is comparable across European countries. The countries in our final dataset include Austria, Czech Republic, Denmark, Finland, Hungary, Ireland, Italy, the Netherlands, Norway, Poland, Portugal, Slovenia, Spain, Sweden, and Switzerland. Four countries (i.e., Czech Republic, Finland, Italy, and Switzerland) will only be included in a subset of our analyses. For Denmark, the NUTS-2 level was introduced in 2007, which coincided with a restructuring of the NUTS-3 levels. We decided to use the 2007 NUTS-2 classification for Denmark for grouping the previous NUTS-3 regions.

We only included respondents who were born in the survey country and who indicated that they had the citizenship of the country and, moreover, whose parents were born in the survey country as well (i.e., the majority group). After listwise deletion of missing values and eliminating influential cases (which was only Vienna), our dataset contains 21,326 respondents living in 125 regions located in 15 European countries.⁵ Due to data limitations, part of our analyses will be conducted using a subsample, which includes 16,408 respondents living in 88 regions in 11 countries.

Dependent Variable: Formal Social Capital

Formal social capital refers to involvement in voluntary associations (cf. Pichler & Wallace, 2007) and was constructed from measures of types of organizations and modes of associational involvement (e.g., membership, active volunteering, or donating money). Respondents were given a list of voluntary organizations and asked whether and how they were involved. Respondents with missing values (i.e., “refusal” or “no answer”) regarding one or more types of organizations were excluded from the analyses.

In line with Van der Meer et al. (2009), we grouped several types of organizations in three categories, according to the goals they pursue. Leisure organizations include “sports,” “culture/hobby,” and “social” organizations. Interest organizations encompass “trade unions,” “professional/business,” and “consumer” organizations. Activist organizations include “environmental/peace/animal rights” and “humanitarian aid/human rights” organizations. For more detailed information on the types of organizations, we refer to the main questionnaire of the ESS.⁶

First, we constructed dichotomous measures of involvement in the three types of voluntary organizations, comparing respondents who were involved in a certain type of organization, with respondents who were not (reference category). We determined for each type separately whether respondents were involved in at least one such organization, irrespective of the modes of associational involvement. For this reason, we considered whether respondents (a) were a member of, (b) participated actively in, (c) volunteered for, and/or (d) donated money to at least one such voluntary association.

Second, for each type of organization, we made a subdivision of those respondents being involved. We distinguished respondents who said to be (a) not involved, (b) actively involved, or (c) passively involved. For each type, respondents who “participated actively in” and/or “volunteered for” at least one such voluntary association were considered to be actively involved. Respondents who were a “member of” and/or “donated money to” at least one such organization, while not simultaneously participating actively or volunteering, were considered passively involved.⁷ For four countries (i.e., Czech Republic, Finland, Italy, and Switzerland), the questions regarding associational involvement were not asked as multiple response questions or were answered differently due to translation, instruction, or interpretation differences (see Van der Meer et al., 2009). We decided to exclude these countries from the analyses based on these categorical measures, as we do not know whether people are involved in only one or in different modes simultaneously.⁸

Mediating Variables: Perceived Ethnic Threat and Intergroup Contact

We measured perceived ethnic threat on a scale of 0 to 10, using six items referring to economic and noneconomic issues related to immigrants, for instance: “Immigrants take jobs away in [country],” “Immigrants are bad for [country’s] economy” or “The [country’s] cultural life is undermined by immigrants.” Higher scores reflect a higher level of perceived ethnic threat. Earlier research showed that perceived ethnic threat can be measured equivalently by these items across all countries in the ESS (Coenders, Lubbers, & Scheepers, 2004). We excluded respondents with missing values on more than two of the six items listwise. Subsequently, we substituted missing values on items with the value on the highest or second highest correlating item. Finally, we calculated the average score on the six items (Cronbach’s α = .82). Note, that our measure of perceived ethnic threat largely approximates realistic group threat as proposed by Stephan, Renfro, and Davis (2002).

We used two items to measure intergroup contact, reflecting the private and occupational domain of intergroup contact (cf. Schlueter & Wagner, 2008): “Do you have any friends who have come to live in [country] from another country?” and “Do you have any colleagues who have come to live in [country] from another country?” For both items, the answer categories are: “no, none at all”, “yes, a few”, and ”yes, several.” With regard to the item referring to immigrant colleagues, respondents could also answer that they were not currently working, which was combined with the category referring to no immigrant colleagues. We excluded respondents with missing values on one or both items listwise. Finally, both items were coded so that higher values reflect more intergroup contact and were used to construct a five-point scale: “0” (both items: “no, none at all”), “1” (one item: “no, none at all”; other item: “yes, a few”), “2” (both items: “yes, a few”; or one item “no, none at all”; other item “yes, several”), “3” (one item “yes, several”; other item “yes, a few”), and “4” (both items: “yes, several”).

Ethnic Diversity at the Contextual Level

We constructed two commonly used measures of ethnic diversity (e.g., Gesthuizen et al., 2009; see also Hooghe, Reeskens, Stolle, & Trappers, 2009), at the country and the regional level: migrant stock and ethnic fractionalization. They are based on figures derived from the 2001 census provided by Eurostat (2010a), containing information on the number of natives and non-natives at both levels. To construct ethnic fractionalization, we distinguished the following nine “ethnic” groups based on their citizenship and country of birth: natives (citizenship in the survey country as well as born in the country), Western countries (i.e., European Union countries, countries of the European Free Trade Association, North America, and Oceania), Africa, Asia, South, and Central America, former communist countries (i.e., Central and Eastern Europe and European Republics [excluding Baltic] of the former USSR), other European countries (i.e., “rest of Europe,” mainly referring to Turks), “other citizenship” (also referring to people with no citizenship, for example, asylum seekers), and finally, naturalized immigrants (i.e., people with the citizenship of the survey country who are born abroad or whose country of birth is unknown). To construct migrant stock, we considered the first two groups (i.e., natives and Western countries) as “Western” and all other groups as “non-Western.” Migrant stock thus refers to the percentage of non-natives with a non-Western citizenship compared with the total population. Ethnic fractionalization, is based on the complement of the Herfindahl index (HI; see, for example, Alesina, De Vleeschauwer, Easterly, Kurlat, & Wacziarg, 2003, p. 159) and indicates the probability that two randomly selected individuals from a population belong to different (ethnic) groups. For our analyses, we centered both measures at their mean. As both measures turned out to be highly correlated at the country and regional level (r > .90), including them simultaneously in our analyses would lead to multicollinearity. Hence, we decided only to include our migrant stock measures and to use the ethnic fractionalization measures in additional sensitivity analyses.

Control Variables at the Contextual Level

We controlled for the level of unemployment at the country and regional level in 2002. The unemployment rate not only reflects competition over scarce resources, thereby determining perceptions of ethnic threat, but is also shown to be inversely related with wealth, which was previously used as control variable for (in)formal social capital (e.g., Gesthuizen et al., 2009; Kesler & Bloemraad, 2010).

Figures on country-level unemployment rates were derived from Eurostat (2010b), except for Switzerland (Organisation for Economic Co-operation and Development [OECD], 2010). For most countries in our dataset, information on regional-level unemployment rates could also be obtained from Eurostat (2010b). Again, Switzerland is an exception (OECD, 2010). For Slovenia, unemployment rates were only obtainable from 2005 onward and for the Danish regions from 2007 onward (when the NUTS-2 classification was introduced). We centered our unemployment measures at their mean.

Control Variables at the Individual Level

Finally, we controlled for several individual-level determinants, in line with previous research on formal social capital, ethnic threat perceptions, and intergroup contact (e.g., Schlueter & Scheepers, 2010; Schneider, 2008; Wilson, 2000; Wilson & Musick, 1997). We used information on the number of years of full-time education to assess educational attainment. For respondents with a missing value, we used information based on the categorical ISCED (International Standard Classification of Education) measure (for all countries except Austria where no ISCED was available). For each country we used the mean years of full-time education corresponding to the particular level of education. For students who were still studying at the time of survey, the study length at the time of the interview was used. Respondents with extreme values on the scale of educational attainment (i.e., more than 20 years; N = 313 for full sample and N = 238 for subsample) were coded as the maximum value of 20 years. Employment situation was measured by asking respondents about their main activity in the past 7 days. Next to a condensed version of the EGP-classification (Erikson, Goldthorpe, & Portocarero, 1979) of social classes for respondents who were in paid employment, we distinguished five categories for respondents who were not in paid employment (see Table A1, Appendix). Religiosity was measured by asking respondents how often they attend religious services (apart from special occasions as weddings and funerals). We condensed the original seven-point measurement to four categories ranging from never to once a week or more and included a fifth category for respondents with a missing value. Marital status and (self-reported) level of urbanization were measured distinguishing four, respectively five categories (see Table A1, Appendix). For those respondents with no information, an additional category was distinguished. Finally, we used straightforward measures of gender (with males as reference category) and age (including a squared term of age). To enable a meaningful interpretation of the intercept, we subtracted the minimum age (i.e., 15) for all respondents. For descriptive statistics, we refer to Table A1, Appendix.

Analyses

As individuals are nested within regions, which in turn are nested within countries, we employed hierarchical random intercept regression analyses, using the Mixed and Genlinmixed procedures in SPSS 20. First, we conducted hierarchical logistic regression analyses, to address the influence of ethnic diversity at the country and regional level on our dichotomous measure of associational involvement. Next, we employed hierarchical multinomial regression analyses, addressing the regional-level effect of ethnic diversity on active and passive involvement. Here, we considered the likelihood to be actively or passively involved as compared with not being involved. Moreover, we addressed the likelihood to be actively involved as compared with passively involved. This reflects people’s choice of how they want to be involved, for those who are involved. To test the effects of ethnic diversity at the regional level on perceived ethnic threat and intergroup contact, we conducted hierarchical linear regression analyses. As we only have cross-sectional data, causal relationships should be interpreted carefully. Yet, prior evidence and theoretical reasoning suggests that the causal order that we assume, is quite plausible. We will come back to this issue in our discussion.

Before we will discuss our results, we first address some methodological issues. First, attention should be paid to the importance of distinguishing between contextual levels: the country level and the regional level. Sensitivity analyses (available on request) revealed that if one does not control for the nesting of regions within countries, the regional-level effect of migrant stock on some dependent variables becomes significant or would be overestimated, whereas for other dependent variables, the effect would be no longer significant after controlling for the nesting of regions within countries. Hence, if we only consider one of both contextual levels, next to the individual level, this would lead to different, erroneous, conclusions. As the number of countries in our hierarchical multinomial regression analyses is limited, care is needed in interpreting the country-level findings. Hence, we only present our findings at the regional level, though include the country-level determinants in our analyses for a more accurate estimation of regional-level effects of ethnic diversity.

Second, our theoretical framework (see Figure 1) calls, ideally, for hierarchical structural equation modeling. However, as these techniques allow us to consider only two hierarchical levels, and our preliminary analyses already stressed the importance of distinguishing two contextual levels next to the individual level, we decided to conduct separate hierarchical regression analyses. First, we used the three types of formal social capital as dependent variables. Next, we considered intergroup contact and perceived ethnic threat as dependent variables. Finally, we included intergroup contact and perceived ethnic threat as predictors of formal social capital. To control for the assumed negative relationship between our mediating variables (cf. Savelkoul, Scheepers, Tolsma, & Hagendoorn, 2011; Schlueter & Scheepers, 2010), we included intergroup contact as determinant of perceived ethnic threat and vice versa.

Third, we conducted several sensitivity analyses to determine the robustness of our findings (results are available on request). First, we used our alternative measure of ethnic diversity (i.e., ethnic fractionalization). Moreover, we considered the effects of migrant stock on involvement in each type of voluntary organization separately. Although our distinction between leisure, interest, and activist organizations provides a clear guideline for grouping several types of voluntary associations (cf. Van der Meer et al., 2009), we are aware that some organizations in one category might vary on some aspects. Overall, the sensitivity analyses revealed substantially similar results, indicating the robustness of our findings.

Results

Based on Putnam’s (rather general) constrict theory, we expected direct negative effects (at the country and regional levels) on involvement in all types of voluntary organizations. As shown in Table 1, we found no significant effect of the national-level migrant stock on involvement in any type of voluntary organization. At the regional level, we only found a direct negative effect of migrant stock on the odds of involvement in interest organizations (b = −.025).⁹ We found a positive effect of migrant stock at the regional level on the odds of involvement in activist organizations (b = .029). People living in European regions with larger proportions of non-Western ethnic minorities are, thus, more likely to be involved in activist organizations, whereas they are less likely to be involved in interest organizations.¹⁰ Based on these findings, we have to refute Hypothesis 1a on the country-level effect of migrant stock for all types of voluntary organizations. Only for involvement in interest organizations, our findings corroborate Hypothesis 1b on the regional-level effect of migrant stock.¹¹

Table 1.

Results Three-Level Logistic Regression Analyses: Formal Social Capital—Direct Effects (Contextual-Level Determinants; N = 21,326; 125 NUTS-2 Regions; 15 Countries).

	M1		M2		M3
	Formal social capital leisure		Formal social capital interest		Formal social capital activist
	b	SE	b	SE	b	SE
Intercept	−1.014	0.198***	−2.846	0.297***	−3.245	0.222***
Regional-level variables
Migrant stock	−0.010	0.012	−0.025	0.014*	0.029	0.012**
Unemployment rate	−0.023	0.009**	−0.004	0.011	−0.036	0.011***
Country-level variables
Migrant stock	0.062	0.049	0.046	0.079	0.017	0.054
Unemployment rate	−0.105	0.046*	−0.134	0.075*	−0.064	0.051

Source. European Social Survey (wave 1, 2002/2003), Eurostat (2010a, 2010b), OECD (2010).

Significant at p < .05. **Significant at p < .01. ***Significant at p < .001 (one-sided test of significance).

Next, we expected that ethnic diversity would predominantly reduce active modes of involvement, with high levels of face-to-face contacts. Moreover, we proposed that the impact of ethnic diversity would be most strongly negative for leisure organizations (which are largely characterized by active modes of involvement) and less so for interest and activist organizations. As shown in Table 2, regional ethnic diversity does not negatively influence people’s likelihood of being involved in leisure organizations, neither actively, nor passively. This underlines our findings presented in Table 1. We do find, however, a negative effect of ethnic diversity at the regional level on active (b = −.045) and passive (b = −.034) involvement in interest organizations. Again, this pattern corresponds to our results shown in Table 1. Living in ethnically diverse regions, moreover, increases the odds of passive involvement (b = .043), while it did not affect active involvement in activist organizations. The positive effect of ethnic diversity on involvement in activist organizations, which we found earlier (Table 1) is, thus, due to an increased likelihood to be passively (rather than actively) involved.

Table 2.

Results Three-Level Multinomial Regression Analyses: Formal Social Capital—Direct Effects (Contextual-Level Determinants; N = 16,408; 88 NUTS-2-Regions; 11 Countries).

	M1—formal social capital—leisure			M2—formal social capital—interest			M3—formal social capital—activist
	PAS vs. NI	ACT vs. NI	ACT vs. PAS	PAS vs. NI	ACT vs. NI	ACT vs. PAS	PAS vs. NI	ACT vs. NI	ACT vs. PAS
	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)
Intercept	−1.786 (0.261)***	−1.693 (0.263)***	0.091 (0.183)	−3.014 (0.360)***	−4.683 (0.362)***	−1.680 (0.245)***	−3.550 (0.303)***	−4.846 (0.298)***	−1.311 (0.335)***
Regional level
Migrant stock	−0.000 (0.014)	−0.017 (0.015)	−0.014 (0.016)	−0.034 (0.011)**	−0.045 (0.021)*	−0.009 (0.020)	0.043 (0.014)**	0.008 (0.021)	−0.032 (0.021)
Unemployment rate	0.014 (0.018)	0.016 (0.019)	0.002 (0.022)	0.014 (0.018)	0.028 (0.027)	0.020 (0.029)	−0.046 (0.024)*	−0.001 (0.028)	0.046 (0.033)

Source. European Social Survey (wave 1, 2002/2003), Eurostat (2010a, 2010b), OECD (2010).

Note. Controlled for all individual-level background characteristics (education, employment status, religiosity, marital status, gender, age, age squared, and urbanization) as well as migrant stock and unemployment rate at the country level. Not controlled for perceived ethnic threat and intergroup contact. PAS = passively involved; ACT = actively involved; NI = not involved.

Significant at p < .05. **Significant at p < .01. ***Significant at p < .001 (one-sided test of significance).

Overall, we found limited support for our expectations that ethnic diversity would predominantly have a negative effect on active involvement (Hypothesis 2a), or on involvement in leisure organizations that are characterized by active modes of involvement (Hypothesis 2b). We only found a negative effect of the regional-level migrant stock on (active and passive) involvement in interest organizations. We will come back to this in our discussion.

Next, we focused on the indirect effects of migrant stock on formal social capital, via perceived ethnic threat and intergroup contact, to answer our second research question as to what extent the understanding of the relationship between ethnic diversity and involvement in voluntary organizations can be improved by mechanisms derived from conflict and contact theory. To address these indirect effects, we conducted several hierarchical (multinomial) regression analyses separately. First, we considered perceived ethnic threat and intergroup contact as dependent variables (Table 3, Models 1 and 2). Next, we included both variables as predictors for active and passive involvement in leisure, interest, and activist organizations. These results are presented in Models 3, 4, and 5 of Table 3. We expected that migrant stock would increase perceptions of ethnic threat as a result of competition over scarce resources or cultural values. However, as we did not find a significant effect of migrant stock (Table 3, Model 1), we have to refute Hypothesis 3a. Yet, as we expected, perceived ethnic threat does decrease the odds of active (b = −.074) and passive (b = −.030) involvement in leisure organizations, active (b = −.118) and passive (b = −.026) involvement in interest organizations and active (b = −.275) and passive (b = −.174) involvement in activist organizations. Hence, Hypothesis 3b is supported for all dimensions of associational involvement. Obviously, the influence of perceived ethnic threat turns out to be rather general: The more people perceive ethnic threat, the less likely they will be actively and passively involved in all types of voluntary associations. Moreover, people perceiving ethnic threat are less likely to be involved actively as compared with passively in all types of organizations.

Table 3.

Results Three-Level (Multinomial) Regression Analyses: Formal Social Capital, Perceived Ethnic Threat and Intergroup Contact—(In)Direct Effects (Contextual-Level Determinants; N = 16,408; 88 NUTS-2-Regions; 11 Countries).

			M3^b—formal social capital—leisure			M4^b—formal social capital—interest			M5^b—formal social capital—activist
	M1^a perceived ethnic	M2^a intergroup contact	PAS vs. NI	ACT vs. NI	ACT vs. PAS	PAS vs. NI	ACT vs. NI	ACT vs. PAS	PAS vs. NI	ACT vs. NI	ACT vs. PAS
	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)
Intercept	6.468(0.129)***	1.503(0.072)***	−1.704(0.278)***	−1.393(0.274)***	0.309(0.211)	−3.015(0.369)***	−4.150(0.381)***	1.176(0.283)***	−2.646(0.310)***	−3.535(0.330)***	−0.925(0.379)**
Individual level
Perceived ethnic threat		−0.090(0.005)***	−0.030(0.015)*	−0.074(0.014)***	−0.044(0.016)**	−0.026(0.015)*	−0.118(0.022)***	−0.088(0.022)***	−0.174(0.016)***	−0.275(0.028)***	−0.099(0.030)***
Intergroup contact	−0.213(0.012)***		0.113(0.023)***	0.176(0.021)***	0.064(0.023)**	0.167(0.022)***	0.238(0.030)***	0.072(0.029)**	0.165(0.022)***	0.299(0.035)***	0.134 (0.037)***
Regional level
Migrant stock	−0.016 (0.010)	0.043(0.005)***	−0.006(0.014)	−0.027(0.015)*	−0.018(0.016)	−0.042(0.012)***	−0.059(0.021)**	−0.016(0.020)	0.032(0.014)*	−0.012 (0.021)	−0.041(0.021)*
Unemployment rate	−0.018 (0.012)	−0.012 (0.006)*	0.014(0.018)	0.016(0.019)	0.002(0.022)	0.015(0.018)	0.026(0.027)	0.017(0.029)	−0.048(0.024)*	−0.007(0.028)	0.043(0.033)

Source. European Social Survey (wave 1, 2002/2003), Eurostat (2010a, 2010b), OCED (2010).

Note. Controlled for all individual-level background characteristics (education, employment status, religiosity, marital status, gender, age, age squared, and urbanization) as well as migrant stock and unemployment rate at the country level. PAS = passively involved; ACT = actively involved; NI = not involved.

Models 1 and 2: Hierarchical linear regression analyses.

Models 3, 4, and 5: Hierarchical multinomial regression analyses.

Significant at p < .05. **Significant at p < .01. ***Significant at p < .001 (one-sided test of significance).

For intergroup contact, we found support for an indirect effect of migrant stock: The more ethnic minorities of non-Western descent live within European regions, the more ethnic majority members have contact with ethnic minorities (b = .043), which is in line with our expectations (Hypothesis 4a). Intergroup contact, in turn, increases the odds of active (b = .176) and passive (b = .113) involvement in leisure organizations, active (b = .238) and passive (b = .167) involvement in interest organizations and active (b = .299) and passive (b = .165) involvement in activist organizations, corroborating Hypothesis 4b. Once again, the emerging picture appears to be strikingly consistent, showing that the more people have intergroup contact, the more likely they are to be involved actively and passively in all types of voluntary organizations. Moreover, people with more intergroup contact are more likely to be involved actively as compared with passively in all types of organizations.

Including perceived ethnic threat and interethnic contact into our models as micro-level determinants that link ethnic diversity to associational involvement, the direct macro-level effects of ethnic diversity changed considerably. The initial positive effect of migrant stock at the regional level on passive involvement in activist organizations weakened. The absent regional-level effect on (active) involvement in leisure organizations turned into a negative effect. And for interest organizations, we found that the negative effects of migrant stock on passive and active involvement became stronger. Overall, considering our very consistent results regarding the effects of perceived ethnic threat and interethnic contact on associational involvement, and their relationships with ethnic diversity, our understanding of the relationship between ethnic diversity and involvement in voluntary organizations has been greatly improved by including these mechanisms.

Finally, although it is not the focus of our study, we would like to mention that most of the effects of our individual-level control variables on involvement in voluntary organizations, intergroup contact, and perceived ethnic threat (results available on request), are in line with previous research (e.g., Schlueter & Scheepers, 2010; Schneider, 2008; Tolsma et al., 2009; Van der Meer et al., 2009; Wilson & Musick, 1997).

Conclusion and Discussion

Interest in the relationship between ethnic diversity and social capital has increased since Putnam (2007) claimed that people living in ethnically more diverse contexts would be more likely to withdraw from social life and, consequently, display lower levels of social capital. In this contribution, we focused on the influence of ethnic diversity on formal social capital, which encompasses involvement in formally constituted voluntary associations (cf. Pichler & Wallace, 2007). We tested whether Putnam’s general claim holds for different modes of involvement (i.e., active and passive) in three types of voluntary organizations: leisure, interest, and activist organizations (cf. Van der Meer et al., 2009). The effects of ethnic diversity turned out to be less general than expected based on Putnam’s constrict theory. Only at the regional level, we found clear influences of living in ethnically diverse regions on the likelihood to be involved in voluntary organizations. These influences vary, however, across the three types of organizations and the different modes of involvement. Ethnic diversity at the regional level only decreased active and passive involvement in interest organizations, but increased rather than decreased passive involvement in activist organizations, and did not affect involvement in leisure organizations at all.

We expected that ethnic diversity would predominantly decrease the likelihood to be actively involved and in organizations characterized by relatively high levels of face-to-face contact, that is, leisure organizations (Gesthuizen et al., 2013; Van der Meer et al., 2009). We found no support for this expectation. However, we found that living in regions with larger proportions of ethnic minorities decreases the likelihood of involvement in interest (not leisure) organizations, reducing active and passive involvement. We argue that, as involvement in leisure organizations may reflect basic social needs, referring to the importance that people attach to socializing and recreating, people might less easily avoid involvement in leisure organizations, whereas they do withdraw from interest organizations, which promote goals that are more remote from their basic needs.

In this study, we also aimed to improve the understanding of the relationship between ethnic diversity and associational involvement. To address our second research question, we incorporated mechanisms derived from conflict theory (Blalock, 1967; Scheepers et al., 2002) and from contact theory (Allport, 1954/1979; Pettigrew & Tropp, 2011).

Whereas ethnic diversity proved to be unrelated to perceived ethnic threat, we found that the more people perceive ethnic threat, the less likely they are to be involved in leisure, interest, or activist organizations, either actively or passively.¹² As perceptions of ethnic threat might induce people to feel anxious about confronting unknown others in general, and ethnic minorities in particular, such perceptions may induce these people to withdraw from social life, including involvement in voluntary organizations. Recent research showed similar effects for informal social capital (Savelkoul et al., 2011): Perceived ethnic threat turned out to reduce informal social meeting (with friends, relatives, and work colleagues). However, this mechanism did not apply to rather strong ties (like helping). Our findings support this line of thought: On average, involvement in a variety of types of voluntary organizations (particularly passive involvement) pertains to weaker ties.

Moreover, our results showed that ethnic diversity at the regional level increased the likelihood of majority group members to have contact with people belonging to ethnic minority groups. Intergroup contact, in turn, increased (active and passive) involvement in all types of voluntary organizations very consistently. These findings support crucial previous insights. Interethnic contact changes people’s social network quantitatively and qualitatively: Larger and more diverse social networks may reduce provincial views and anxiety (Pettigrew, 1998; Pettigrew & Tropp, 2008) and increase the likelihood to participate in voluntary associations (Wilson, 2000).

We are aware that the causal order between both mediating variables and formal social capital might be disputable to some extent. Future research should, preferably, use panel data to address these causality issues more profoundly. Nevertheless, our results provide initial evidence for the proposed causal order. We found a positive effect of intergroup contact and a negative effect of perceived ethnic threat on passive associational involvement. As these modes of involvement do not involve face-to-face contacts, reversed causality is largely ruled out. Furthermore, additional analyses revealed substantially similar effects for interethnic contact with friends as well as colleagues (results available on request). If associational involvement would increase intergroup contact, one would only expect this effect for interethnic contact with friends.

Another direction that future research could take is to focus on the (direct and indirect) influence of ethnic diversity at lower aggregate levels (e.g., municipalities or neighborhoods) within single countries: Earlier studies pointed at the importance of proximity (e.g., Tolsma et al., 2009) and (particularly active) participation in voluntary organizations is to a large extent a local type of participation.

In sum, our findings show differential effects of ethnic diversity on associational involvement for different types of organizations or modes of involvement. As different types of voluntary organizations vary in their goals and most prevalent modes of involvement (see Van der Meer et al., 2009) and only some modes of involvement are linked with face-to-face contact, while others are not, it is quite understandable that the influence of ethnic diversity is far more mixed than supposed by Putnam. As such, our findings imply that generalizations suggested by Putnam (2007) are not warranted and future research should preferably use more fine-grained rather than general measurements of formal social capital to disentangle these differentiated effects. Our results stress, moreover, the need to incorporate individual-level mechanisms, that is, ethnic threat perceptions and intergroup contact, which both show very consistent effects on active and passive involvement in all types of voluntary organizations As such, we believe that we have reached a higher level of understanding as to how ethnic diversity translates into more or less associational involvement in European regions.

Footnotes

Appendix

Table A1 (Appendix).

Descriptive Statistics Individual-Level and Contextual-Level Variables.

Variable	Full sample (15 countries; 125 regions; N = 21,326)			Subsample (11 countries; 88 regions; N = 16,408)
Variable	Range	M/percentage	SD	Range	M/percentage	SD
Individual level
Dependent variables—dichotomous
Involvement—leisure organizations	0/1	47.29%		0/1	47.11%
Involvement—interest organizations	0/1	40.80%		0/1	40.62%
Involvement—activist organizations	0/1	23.44%		0/1	24.34%
Dependent variables—categorical	Not applicable
Leisure organizations—not involved				0/1	52.89%
Leisure organizations—actively involved				0/1	27.47%
Leisure organizations—passively involved				0/1	19.64%
Interest organizations—not involved				0/1	59.38%
Interest organizations—actively involved				0/1	8.84%
Interest organizations—passively involved				0/1	31.78%
Activist organizations—not involved				0/1	75.66%
Activist organizations—actively involved				0/1	5.06%
Activist organizations—passively involved				0/1	19.28%
Mediating variables
Perceived ethnic threat	0-10	5.39	1.56	0-10	5.40	1.58
Intergroup contact	0-4	0.91	1.07	0-4	0.93	1.08
Control variables individual level
Age (15 = 0)	0-88	31.16	17.79	0-87	30.87	17.79
Age squared	0-7,744	1,287.62	1,220.50	0-7,569	1,269.60	1,216.02
Educational attainment (years)	0-20	11.72	3.92	0-20	11.80	3.97
Religiosity
Church attendance never (ref.)	0/1	29.45%		0/1	29.67%
Church attendance rarely	0/1	40.77%		0/1	37.98%
Church attendance once a month	0/1	10.15%		0/1	10.59%
Church attendance once a week or more	0/1	19.32%		0/1	21.40%
Church attendance missing	0/1	0.31%		0/1	0.36%
Marital status
Not married/never been married (ref.)	0/1	29.02%		0/1	29.42%
Married	0/1	55.20%		0/1	55.59%
Divorced/living separated	0/1	7.41%		0/1	6.68%
Widow(er)	0/1	8.09%		0/1	8.05%
Marital status missing	0/1	0.28%		0/1	0.26%
Employment status
Service class (ref.)	0/1	18.30%		0/1	18.24%
Routine non manuals	0/1	11.20%		0/1	11.18%
Self-employed	0/1	3.99%		0/1	3.79%
Manual workers	0/1	16.52%		0/1	16.41%
Occupational status missing (employed)	0/1	1.58%		0/1	1.69%
Unemployed	0/1	3.96%		0/1	3.97%
Student	0/1	8.93%		0/1	8.90%
Housekeeping	0/1	11.61%		0/1	12.52%
Retired	0/1	19.16%		0/1	18.20%
Other employment situation	0/1	4.75%		0/1	5.10%
Urbanization
Big city	0/1	14.27%		0/1	14.84%
Suburbs or outskirts of big city	0/1	13.92%		0/1	14.68%
Town or small city (ref.)	0/1	29.85%		0/1	28.68%
Country village	0/1	32.31%		0/1	32.50%
Farm or home in the countryside	0/1	9.43%		0/1	9.02%
Urbanization missing	0/1	0.22%		0/1	0.29%
Gender
Male (ref.)	0/1	48.53%		0/1	48.56%
Female	0/1	51.47%		0/1	51.44%
Regional level
Migrant stock	1.22-20.25	6.09	4.12	1.45-16.15	5.88	3.40
Ethnic fractionalization	0.03-0.58	0.14	0.10	0.03-0.34	0.13	0.07
Unemployment rate	2.00-26.30	8.21	6.39	2.00-26.30	8.27	6.63
Country level
Migrant stock	2.48-18.30	7.14	4.17	2.82-12.42	6.91	2.92
Ethnic fractionalization	0.06-0.48	0.16	0.11	0.06-0.25	0.15	0.06
Unemployment rate	2.80-19.90	6.77	4.39	2.80-19.90	6.65	4.94

Source. European Social Survey (wave 1, 2002/2003), Eurostat (2010a, 2010b), Organisation for Economic Co-operation and Development (2010).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study is part of a project supported by the Netherlands Organization for Scientific Research (Grant 432-008-005).

Notes

Author Biographies

Michael Savelkoul is a PhD candidate and lecturer in the Department of Sociology at Radboud University Nijmegen, the Netherlands. His research is focused on ethnic diversity and social cohesion. His work has appeared in European Sociological Review, Quality & Quantity, and Social Science Research.

Maurice Gesthuizen is assistant professor in the Department of Sociology at Radboud University Nijmegen, the Netherlands. His main research interests are in the fields of educational inequality, economic vulnerability, social capital, and their interrelationships.

Peer Scheepers is professor of comparative methodology in the Department of Sociology at Radboud University Nijmegen, the Netherlands. He is interested in cross-national research on social cohesion issues.

References

Alesina

De Vleeschauwer

Easterly

Kurlat

Wacziarg

(2003). Fractionalization. Journal of Economic Growth, 8, 155-194.

Alesina

La Ferrara

(2000). Participation in heterogeneous communities. The Quarterly Journal of Economics, 115, 847-904.

Allport

G. W.

(1979). The nature of prejudice. Boston, MA: Beacon Press. (Original work published 1954)

Blalock

H. M.

(1967). Toward a theory of minority group relations. New York, NY: John Wiley.

Bobo

L. D.

(1999). Prejudice as group position: Microfoundations of a sociological approach to racism and race relations. Journal of Social Issues, 55, 445-472.

Cheong

P. H.

Edwards

Gouldbourne

Solomos

(2007). Immigration, social cohesion and social capital: A critical review. Critical Social Policy, 27, 24-49.

Coenders

Lubbers

Scheepers

. (2004). Majorities’ attitudes towards minorities in Western and Eastern European societies: Results from the European Social Survey 2002-2003 (Report 4 for the European Monitoring Centre on Racism and Xenophobia). Vienna, Austria: European Monitoring Centre on Racism and Xenophobia.

Cornelius

W. A.

Rosenblum

M. R.

(2005). Immigration and politics. Annual Review of Political Science, 8, 99-119.

Coser

L. A.

(1956). The function of social conflict. Glencoe, IL: Free Press.

10.

Erikson

Goldthorpe

J. H.

Portocarero

(1979). Intergenerational class mobility in three Western European societies: England, France and Sweden. British Journal of Sociology, 30, 415-441.

11.

Eurostat. (2003). European regional statistics reference guide. Luxembourg: Office for Official Publication of the European Communities.

12.

Eurostat. (2010a). Demographic structure of the population at the regional level. Retrieved from http://ec.europa.eu/eurostat

13.

Eurostat. (2010b). Unemployment rates at the NUTS-1, NUTS-2 and NUTS-3 Level. Retrieved from http://ec.europa.eu/eurostat

14.

Gesthuizen

Scheepers

Van der Veld

Völker

(2013). Structural aspects of social capital: Tests for cross-national equivalence in Europe. Quality & Quantity, 47, 909-922.

15.

Gesthuizen

Van der Meer

Scheepers

(2009). Ethnic diversity and social capital in Europe: Tests of Putnam’s thesis in European countries. Scandinavian Political Studies, 32, 121-142.

16.

Gijsberts

Van der Meer

Dagevos

(2012). “Hunkering down” in multi-ethnic neighbourhoods? The effects of ethnic diversity on dimensions of social cohesion. European Sociological Review, 28, 527-537.

17.

Hooghe

Reeskens

Stolle

Trappers

(2009). Ethnic diversity and generalized trust in Europe: A cross-national multilevel study. Comparative Political Studies, 42, 198-223.

18.

Hooghe

Trappers

Meuleman

Reeskens

(2008). Migration to European countries: A structural explanation of patterns, 1980-2004. International Migration Review, 42, 476-504.

19.

Jowell

, & The Central Co-ordinating Team. (2003). European Social Survey 2002/2003: Technical report. London, England: Centre for Comparative Social Surveys, City University.

20.

Kesler

Bloemraad

(2010). Does immigration erode social capital? The conditional effects of immigration-generated diversity on trust, membership and participation across 19 countries, 1981-2000. Canadian Journal of Political Science, 43, 319-347.

21.

Letki

(2008). Does diversity erode social cohesion? Social capital and race in British neighbourhoods. Political Studies, 56, 99-126.

22.

McPherson

Smith-Lovin

Cook

J. M.

(2001). Birds of a feather: Homophily in social networks. Annual Review of Sociology, 27, 415-444.

23.

Organisation for Economic Co-operation and Development. (2010). Unemployment rates regional labour market. Available from http://stats.oecd.org

24.

Pettigrew

T. F.

(1998). Intergroup contact theory. Annual Review of Psychology, 49, 65-85.

25.

Pettigrew

T. F.

Tropp

L. R.

(2008). How does intergroup contact reduce prejudice? Meta-analytical tests of three mediators. European Journal of Social Psychology, 38, 922-934.

26.

Pettigrew

T. F.

Tropp

L. R.

(2011). When groups meet: The dynamics of intergroup contact. New York, NY: Psychology Press.

27.

Pichler

Wallace

(2007). Patterns of formal and informal social capital. European Sociological Review, 23, 423-435.

28.

Putnam

R. D.

(2000). Bowling alone: The collapse and revival of American community. New York, NY: Simon & Schuster.

29.

Putnam

R. D.

(2007). E pluribus unum: Diversity and community in the twenty-first century. The 2006 Johan Skytte Prize lecture. Scandinavian Political Studies, 30, 137-174.

30.

Rotolo

Wilson

(2012). State-level differences in volunteerism in the United States: Research based on demographic, institutional, and cultural macrolevel theories. Nonprofit and Voluntary Sector Quarterly, 41, 452-473.

31.

Savelkoul

Gesthuizen

Scheepers

(2011). Explaining relationships between ethnic diversity and informal social capital across European countries and regions: Tests of constrict, conflict and contact theory. Social Science Research, 40, 1091-1107.

32.

Savelkoul

Scheepers

Tolsma

Hagendoorn

(2011). Anti-Muslim attitudes in the Netherlands: Tests of contradictory hypotheses derived from ethnic competition theory and intergroup contact theory. European Sociological Review, 27, 741-758.

33.

Scheepers

Gijsberts

Coenders

(2002). Ethnic exclusionism in European countries: Public opposition to civil rights for legal migrants as a response to perceived ethnic threat. European Sociological Review, 18, 17-34.

34.

Schlueter

Scheepers

(2010). The relationship between outgroup size and anti-outgroup attitudes: A theoretical synthesis and empirical test of group threat- and intergroup contact theory. Social Science Research, 39, 285-295.

35.

Schlueter

Wagner

(2008). Regional differences matter. Examining the dual influence of the regional size of the immigrant population on derogation of immigrants in Europe. International Journal of Comparative Sociology, 49, 153-173.

36.

Schneider

S. L.

(2008). Anti-immigrant attitudes in Europe: Outgroup size and perceived ethnic threat. European Sociological Review, 24, 53-67.

37.

Stephan

W. G.

Renfro

C. L.

Davis

M. D.

(2002). The role of threat in intergroup relations. In Wagner

Tropp

Finchilescu

Tredoux

(Eds.), Improving intergroup relations: Building on the legacy of Thomas F. Pettigrew (pp. 55-72). Oxford, UK: Blackwell Publishing.

38.

Ter Wal

. (Ed.). (2002). Racism and cultural diversity in the mass media: An overview of research and examples of good practice in the EU member states, 1995-2000. Vienna, Austria: European Monitoring Centre on Racism and Xenophobia.

39.

Tolsma

Van der Meer

Gesthuizen

(2009). The impact of neighbourhood and municipality characteristics on social cohesion in the Netherlands. Acta Politica, 44, 286-313.

40.

Van der Meer

Te Grotenhuis

Scheepers

. (2009). Three types of voluntary associations in comparative perspective: The importance of studying associational involvement through a typology of associations in 21 European countries. Journal of Civil Society, 5, 227-241.

41.

Wilson

(2000). Volunteering. Annual Review of Sociology, 26, 215-240.

42.

Wilson

(2012). Volunteerism research: A review essay. Nonprofit and Voluntary Sector Quarterly, 41, 176-212.

43.

Wilson

Musick

(1997). Who cares? Toward an integrated theory of volunteer work. American Sociological Review, 62, 694-713.