Re-assessing self-interest theory: Can intergroup contact constrain self-interested reactions?

Abstract

This article expands previous research on self-interested reactions and intergroup contact by showing that (1) intergroup contact constrains the impact of personal economic vulnerability on perceived out-group threat, (2) personal economic vulnerability has no influence on perceived out-group threat when contact is most intense, and (3) the ability of intergroup contact to reduce perceived out-group threat is strongest among the most economically vulnerable. These findings were generated by a fixed-effects regression of 21 countries and almost 34,000 respondents sampled in the European Social Survey (2002–2003). More generally, our findings suggest that self-interested reactions can be overpowered by the cognitive and affective components of intergroup contact. The concluding section expands on the theoretical implications of the results.

Keywords

Ethnic threat self-interest intergroup contact vulnerability affective ties cognitive learning

Introduction

Although perceived out-group threat among in-group members is common in many countries, social scientists still debate why some feel more threatened than others (Sides and Citrin, 2007). According to the influential self-interest theory, perceived out-group threat reflects personal economic vulnerability (Bobo and Hutchings, 1996; Citrin and Green, 1990; Kluegel and Smith, 1983; Sears and Funk, 1991; Sniderman and Hagendoorn, 2007). The most vulnerable in-group members fear redundancies or welfare cuts, whereas the least vulnerable feel less threatened because they have more resources (Raijman et al., 2003). On that account, perceived out-group threat is a socially fixed disposition.

In contrast, intergroup contact theory offers a competing explanation by arguing that in-group members’ negative perceptions of out-group members can be improved (Pettigrew and Tropp, 2008). Social psychologists have long argued that intergroup contact improves cross-group relations because of its ability to stimulate affective involvement and cognitive learning about out-group member characteristics (Pettigrew and Tropp, 2011: 77–84; Stephan et al., 2000; Tausch et al., 2007).

Obviously, self-interest theory privileges egocentric motives, whereas intergroup contact theory emphasizes pro-social motives. However, these competing human motivations do not always operate in isolation from each other (Dixon, 2006). This raises the question of whether personal economic vulnerability nurtures perceived out-group threat uniformly among contacted and non-contacted in-group members. Surprisingly, no one has yet examined this important question. Four previous studies, however, report that larger numbers of immigrants increase perceived out-group threat, but less so among contacted in-group members (McLaren, 2003; Schneider, 2008; also Semyonov et al., 2004; Weber, 2015). These studies have applied the number of immigrants as an indicator of actual interethnic competition. Although undeniably relevant, this approach fails to establish an adequate test of self-interest theory. Self-interest theory is not about the competitive process in itself, but about personal vulnerability – or personal stakes – in this struggle (Kluegel and Smith, 1983; Tolnay and Beck, 1992).

In view of this weakness, we expand previous research on both self-interested reactions and intergroup contact by claiming that intergroup contact is a boundary condition of self-interested reactions in interethnic relations. Specifically, we argue that intergroup contact constrains the impact of personal economic vulnerability on perceived out-group threat because it generates affective ties between the participants and stimulates cognitive learning. The influence of self-interest is weakened when affective ties are established, whereas self-interested reactions toward out-group members are stronger when these members remain remote and unknown.

Accordingly, this article hypothesizes that the greater the extent of intergroup contact, the less personal economic vulnerability fuels perceived out-group threat – ultimately, when contact is most intense, the relationship between personal economic vulnerability and perceived out-group threat disappears. Inspired by Tolnay and Beck (1992), we specify this hypothesis by paying attention to the most economically vulnerable as they are of paradigmatic interest. First, the most vulnerable are a critical test group as their stakes are highest, implying that they should be disproportionally less susceptible to positive contact experiences. Next, if the most vulnerable nonetheless prove highly sensitive toward personal interaction, this necessarily accentuates the powerful influence of intergroup contact.

The constrained self-interest hypothesis is tested using fixed-effects regression on the European Social Survey (ESS) (2002–2003). The empirical analyses confirm that intergroup contact strongly constrains the influence of personal economic vulnerability on perceived out-group threat. In fact, personal economic vulnerability has no influence on perceived out-group threat when contact is most intense. Furthermore, we show that personal contact is particularly influential among the most economically vulnerable.

The simple self-interest theory

Explaining majority members’ reactions by resource indicators calls for self-interest theories, sometimes referred to as realistic group conflict theory (Hayes and Dowds, 2006; Quillian, 1995; Tolnay and Beck, 1992). Realistic group conflict theory includes different variants, but we wish to focus on a particular one – the so-called simple self-interest model as it has been labeled by Bobo and Hutchings (1996).

The simple self-interest theory proposes that negative ethnic attitudes are not irrational responses because they often rest on an objective basis (Bobo and Hutchings, 1996; Sides and Citrin, 2007). Social groups compete for scarce resources, and all groups defend their existing privileges (Coenders and Scheepers, 1998). When a society becomes multiethnic, competition is established between in-group and out-group members. This competition will occur in the educational institutions, the job market, and the housing market (Hernes and Knudsen, 1992).While competition creates disadvantages for all groups, it is clear that the well-off are less vulnerable than the economically unprivileged (Hello et al., 2006; Pichler, 2010). Both the privileged and less-privileged groups pursue their self-interest, but the fact that personal economic resources are unequally distributed explains why these groups cannot have similar attitudes toward out-group members (cf. Semyonov et al., 2008). Tolnay and Beck (1992: 49) summarize the essential point by stating that the in-group is not monolithic in character – because it is internally stratified according to socioeconomic resources.

The simple self-interest model and its personal vulnerability correlate also relate to non-economic issues. Admittedly, this is debated; however, our understanding of threat is based on recent attempts to view symbolic and realistic threat perceptions as complementary and interrelated as emphasized by advocates of the so-called integrated threat theory (Riek et al., 2006). For instance, criminal activity among immigrants is more likely to occur in low-status group neighborhoods rather than in middle-class residential areas. Immigration can also erode traditional social networks in local working-class neighborhoods, and it can make social interaction more difficult because of different languages, life styles, customs, or religions (Bobo and Hutchings, 1996; Stephan et al., 2008).¹ One may argue that such social changes only carry intangible ‘symbolic costs’, but it should not be ignored that they also relate to personal economic resources. Unlike the privileged, the economically unprivileged cannot as easily move to a middle-class neighborhood or send their children to prestigious private schools. Understandably, previous studies have concluded that low-status groups prefer a traditional life style within the confines of the nation-state, whereas high-status groups welcome globalization (Inglehart, 1997; Napier and Jost, 2008).

Intergroup contact theory

While self-interest theory relates to pocketbook logic (Bobo and Hutchings, 1996), the so-called intergroup contact hypothesis builds on social psychological theory that emphasizes the importance of predispositions and emotional reactions (Miller et al., 2004). Naturally, this hypothesis offers a different explanation of intergroup conflict. According to the traditional version, negative reactions toward out-group members reflect exaggerated (negative) prejudgments (Allport, [1954] 1979: 9). Yet, although prejudgments are socially learned, they are not insensitive toward actual, personal experiences. Contact situations most frequently improve intergroup relations because they provide first-hand information about out-group members that often disconfirm in-group members’ prejudgments (Gundelach, 2014).

Subsequent research shows that contact situations not only involve cognitive learning (Miller, 2002). They also establish affective ties between the participants by enhancing intergroup empathy and by reducing anxiety about the presence of out-group members (Page-Gould et al., 2008; Pettigrew and Tropp, 2008; Swart et al., 2010; Vezzali et al., 2010; Vezzali and Giovannini, 2012). Thus, previous studies show that knowledge, reduced anxiety, and enhanced empathy are immediate consequences of contact; it also follows that they identify key components of the average intergroup contact situation. In other words, these characteristics explain why intergroup contact may be capable of influencing the association between personal economic vulnerability and threat-based reactions toward out-group members.

Although meta-analyses show that positive contact effects are robust across various social situations, scholars also emphasize that they are likely to be stronger when certain facilitating conditions are met (Pettigrew and Tropp, 2006, 2011: 61–73). Ideally, contact settings should include equal status between the participants, common goals, cooperation, institutional support, and acquaintance potential (Allport, [1954] 1979: 281; Koschate and Van Dick, 2011). For instance, intense zero-sum-based competition between out-group and in-group members may ‘poison’ their mutual relationship because personal stakes are too high. Rather, intimacy (closeness), cooperation, and functional interdependence are of special importance (Allport, [1954] 1979: 489; Dovidio et al., 2003). In particular, these conditions underline the special role of personal friendship as this type of contact is likely to meet most of the conditions (Pettigrew and Tropp, 2011).

Self-interest constrained by intergroup contact

The simple self-interest theory does not address the ability of intergroup contact to reduce negative reactions toward out-group members. This impersonal approach to ethnic relations possibly originates from the notion of self-interest itself, which leaves little room for misperceptions of external threats, let alone positive feelings toward out-group members. Bobo and Hutchings (1996) note that ‘some specifications of this approach ignore … subjective interpretation and assignment of meaning to “social facts”’ (p. 953). Indeed, self-interest is a non-affective, fixed, and dominant disposition that motivates across various contexts (Sears and Funk, 1991).

Few will deny that self-interest can have a strong motivational influence under favorable circumstances (Monroe, 2001). Yet, frequently the circumstances may be unfavorable as many studies conclude that the influence of self-interest is limited or even non-existent (e.g. Hainmueller and Hiscox, 2010; Sears et al., 1979; Sears and Funk, 1990; Sides and Citrin, 2007). Specifically, Scheve and Slaughter (2001) reported that less-skilled workers in the United States were more likely to oppose immigration, but workers in high-immigration areas were not more opposed than people who lived elsewhere. Likewise, using a more indirect indicator of realistic threat, Dixon (2006) showed that the impact of the perceived number of Blacks in the community on prejudice was conditional on intergroup contact. Greater number of Blacks nurtured prejudice among non-contacted Whites, whereas greater number of Blacks reduced prejudice among contacted Whites (for similar results, see McLaren, 2003; Schneider, 2008; Weber, 2015).

Even so, the number of out-group members at the national or local level is a poor indicator of self-interested reactions in a narrow sense. Consequently, although previous studies deserve credit for identifying intergroup as a boundary condition, a key question remains: To what extent may majority members’ self-interested reactions toward out-group members be constrained by direct contact with them? In answering this question, one needs to realize that the human mind comprises both instrumental and affective reasoning devices as emphasized by Marcus et al. (2000; also Monroe, 2001). Whether the human mind is momentarily biased in favor of one of them depends on the nature of the interactional context. It seems natural to be selfish when shopping, whereas self-interested motivation appears less dominant in contexts of personal interaction with friends or co-workers (see Mackuen et al., 2007). Intimate interaction stimulates empathy, implicating concern about other peoples’ emotions and well-being. Empathy is a pro-social rather than self-interested propensity (Eisenberg et al., 2010). Similarly, Batson et al. (1997: 106) argue that empathy has the potential to constrain egoistic desires. Likewise, reduced anxiety as a consequence of regular contact with out-group members engenders more thorough thinking and readiness to avoid exaggerated negative beliefs about out-group members. And greater knowledge about out-group members is likely to foster some mutual cultural understanding. In combination, these considerations are consistent with psychological research reporting that direct personal experiences can shape perceptions (Fazio et al., 1982). Direct personal experiences with out-group members make evaluation of them easier. The evaluation is likely to become more positive as intergroup contact settings enhance empathy, reduce anxiety, and provide greater knowledge about out-group members. Eventually, positive evaluations (stimulated by contact) are likely to reduce perceived out-group threat (also Pickett et al., 2014).

Our argument predicts that it matters whether or not the impact of self-interest works in isolation from intergroup contact stimuli. From this we can deduce our constrained self-interest hypothesis. More specifically, economically vulnerable persons having no contact with out-group members will feel more threatened by these than economically vulnerable persons who frequently interact with out-group members – provided that most of the facilitative conditions of contact are met. Among the first group, factors capable of constraining self-interested reactions are inactive. Among the other most economically vulnerable group, positive stimuli in terms of cognitive learning and affective ties are active. This means that self-interested reactions can be mentally unconstrained, constrained, or possibly overpowered. By ‘overpowering’ we mean positive outcomes of contact that make the impact of personal economic vulnerability disappear.

Accordingly, we test the following hypothesis:

The constrained self-interest hypothesis: The greater the extent of intergroup contact, the less personal economic vulnerability nurtures perceived out-group threat – ultimately, when contact is most intense, the relationship between personal economic vulnerability and perceived out-group threat disappears.

This interaction hypothesis involves another testable implication capable of indicating the inherent capacity of intergroup contact to modify strong self-centered motivations. Specifically, if the most economically vulnerable prove disproportionately sensitive toward contact experiences with out-group members, this constitutes additional support to the claim that self-interest motives can be effectively constrained by cognitive learning and affective involvement. To provide a fair test of the simple self-interest theory, we believe that the most economically vulnerable need be treated as the critical test case – because their stakes are highest.²

Data, measures, controls, and method

The present investigation utilizes the ESS data set from 2002–2003 because it includes two important measures of intergroup contact.³ The ESS is commonly regarded as a reliable and high-quality data source (Norris, 2004), and further specific details about the sampling procedure and fieldwork can be found in ESS (2011). This data source is particularly attractive because of its cross-national nature – it includes 40,000 respondents from 21 countries.⁴ Compared to single-nation cross-sectional surveys, the ESS obviously allows for a stronger test of theoretical hypotheses. The data set also includes members of ethnic minorities, but these were excluded from our analyses as we examine the intergroup contact effect among majority members (causing a sample loss of about 5 %).⁵ Finally, it may be considered a key strength that data were collected during a period of economic prosperity. Had the data been collected during the world-wide financial crisis (2008–2014), our test would probably have been biased in favor of confirming effects of personal economic hardship. Accordingly, our findings may be considered conservative.

To measure our key variable – personal economic vulnerability – we used the respondent’s subjective assessment of his or her economic well-being. Referring to household income, this variable has four response categories: (1) ‘living comfortably on present income’, (2) ‘coping on present income’, (3) ‘finding it difficult on present income’, and (4) ‘finding it very difficult on present income’ (plus refusals, don’t knows, and no answer, which were all excluded from the analysis). This indicator of personal economic vulnerability is attractive because of its composite nature. An obvious alternative candidate is self-reported personal or household income with predefined brackets (e.g. Kluegel and Smith, 1983), but such measures are unlikely to adequately cover a person’s financial well-being or hardship. The point is that personal economic vulnerability derives from several sources, such as regular wage income, the number of adults in a household, their amount of debt, personal assets, and number of children. When a respondent characterizes her own financial situation as ‘comfortable’ or ‘very difficult’, she most likely observes her household’s total net balance (i.e. rather than her personal income in isolation). This measure was subsequently treated as a metric variable and rescaled to vary from 0 to 1, higher values indicating personal economic vulnerability. Table 1 shows the average of this variable, which indicates considerable cross-national variation in terms of personal economic vulnerability. Unsurprisingly, the number of respondents reporting financial hardship is relatively larger in poorer countries (e.g. Greece, Poland, and Israel).

Table 1.

Descriptives of perceived out-group threat, personal economic vulnerability, and intergroup contact, by country.

Countries (N =21)	Perceived threat (0−10)	α	Economic vulnerability (0–1)	Intergroup contact (0–1)	N
Austria	M: 5.32; SD: 1.77	0.86	M: 0.31; SD: 0.27	M: 0.33; SD: 0.33	1868
Belgium	M: 5.68; SD: 1.52	0.83	M: 0.33; SD: 0.33	M: 0.27; SD: 0.30	1506
Switzerland	M: 5.09; SD: 1.33	0.79	M: 0.19; SD: 0.23	M: 0.45; SD: 0.32	1756
Czech Republic	M: 6.23; SD: 1.55	0.84	M: 0.45; SD: 0.25	M: 0.17; SD: 0.27	1060
Germany	M: 5.56; SD: 1.60	0.84	M: 0.30; SD: 0.24	M: 0.25; SD: 0.28	2518
Denmark	M: 5.18; SD: 1.61	0.85	M: 0.14; SD: 0.22	M: 0.25; SD: 0.24	1336
Spain	M: 5.31; SD: 1.48	0.82	M: 0.32; SD: 0.26	M: 0.17; SD: 0.24	1442
Finland	M: 4.96; SD: 1.44	0.81	M: 0.33; SD: 0.23	M: 0.18; SD: 0.22	1812
United Kingdom	M: 5.74; SD: 1.72	0.88	M: 0.25; SD: 0.24	M: 0.25; SD: 0.29	1827
Greece	M: 6.99; SD: 1.86	0.87	M: 0.52; SD: 0.29	M: 0.14; SD: 0.23	2047
Hungary	M: 6.25; SD: 1.65	0.84	M: 0.49; SD: 0.25	M: 0.16; SD: 0.24	1190
Ireland	M: 5.33; SD: 1.70	0.84	M: 0.27; SD: 0.26	M: 0.23; SD: 0.26	1769
Israel	M: 5.36; SD: 1.95	0.83	M: 0.45; SD: 0.30	M: 0.42; SD: 0.35	1655
Italy	M: 5.27; SD: 1.47	0.73	M: 0.28; SD: 0.25	M: 0.15; SD: 0.22	1090
Luxembourg	M: 4.54; SD: 1.57	0.71	M: 0.19; SD: 0.23	M: 0.45; SD: 0.33	1122
The Netherlands	M: 5.43; SD: 1.38	0.80	M: 0.19; SD: 0.23	M: 0.26; SD: 0.28	2140
Norway	M: 5.19; SD: 1.37	0.80	M: 0.19; SD: 0.23	M: 0.34; SD: 0.31	1684
Poland	M: 5.47; SD: 1.58	0.82	M: 0.49; SD: 0.22	M: 0.08; SD: 0.15	1731
Portugal	M: 5.57; SD: 1.60	0.83	M: 0.46; SD: 0.26	M: 0.22; SD: 0.28	1177
Sweden	M: 4.52; SD: 1.55	0.81	M: 0.18; SD: 0.22	M: 0.39; SD: 0.31	1775
Slovenia	M: 5.73; SD: 1.48	0.80	M: 0.27; SD: 0.25	M: 0.25; SD: 0.25	1279

Total (21)	M: 5.47; SD: 1.69	0.83	M: 0.30; SD: 0.27	M: 0.26; SD: 0.29	33,784

M: mean; SD: standard deviation.

Perceived (out-group) threat comprises six items that were averaged and turned into an index that ranges from 0 to 10, higher values indicating greater perceived out-group threat.

Source: European Social Survey (2002–2003).

Consistent with previous research (Schneider, 2008), we have chosen six items to measure perceived out-group threat.⁶ These items were also chosen because of their ability to stimulate rational reactions. Their question wordings are as follows: (1) ‘Most people who come to live here work and pay taxes. They also use health and welfare services. On balance, do you think people who come here take out more than they put in or put in more than they take out?’ (2) ‘Would you say it is generally bad or good for [country]’s economy that people come to live here from other countries?’ (3) ‘Would you say that people who come to live here generally take jobs away from workers in [country], or generally help to create new jobs?’ (4) ‘Would you say that [country]’s cultural life is generally undermined or enriched by people coming to live here from other countries?’ (5) ‘Is [country] made a worse or a better place to live by people coming to live here from other countries?’ and (6) ‘Are [country]’s crime problems made worse or better by people coming to live here from other countries?’⁷ The six items were averaged and turned into an index that ranges from 0 to 10, higher values indicating greater perceived out-group threat. Respondents with more than three ‘don’t know’ responses were excluded from the analyses resulting in a sample loss of about 1 percent. The reliability of the index is acceptable (Cronbach’s α = 0.83 for the pooled sample). More specifically, Table 1 shows that Cronbach’s α ranges from 0.71 to 0.88 across the included countries. In all likelihood, our construct seems to identify a fairly similar phenomenon in otherwise different countries. Moreover, Table 1 reveals considerable cross-national differences in terms of perceived out-group threat. In view of the grand average (= 5.47), citizens of Greece, Hungary, and the Czech Republic feel disproportionally more threatened by out-group members than citizens of Sweden, Luxembourg, and Finland.

To measure another key variable – intergroup contact – we have chosen workplace contact and interethnic friendship which are both included in the ESS data set. The specific wording of the two contact measures addresses the number of immigrant colleagues or friends, with none, few, and several as response categories.⁸ Responses to these two items were summated into an index of intergroup contact, which was subsequently rescaled to vary from 0 to 1, higher values indicating greater contact intensity. Descriptive details about this measure are reported in Table 1.

We combined these two measures because they have different strengths and weaknesses. Workplace contacts are less subjective to self-selection than friendship contacts because the setting is more involuntary than the voluntary nature of friendships (Huckfeldt et al., 2013; Mondak and Mutz, 2006; Mutz, 2002; Pettigrew, 1998). Workplace contact, however, is not a perfect setting as having immigrant colleagues may occasionally involve superficial contact (i.e. encounters without intimacy). In contrast, friendship almost by definition ensures strong intimacy, which is commonly regarded a key facilitative condition for the contact effect (Pettigrew and Tropp, 2011). Likewise, building on previous studies and meta-analyses of the mediators of intergroup contact effects, we see our composite measure of contact as an indicator of cognitive learning and affective ties between the participants (Pagotto et al., 2010; Pettigrew and Tropp, 2008, 2011).⁹ Finally, combining workplace contact and friendship in a composite measure increases social coverage as the two individual measures are not that strongly related (tau beta = 0.36). Recall that our dependent variable relates to both economic and non-economic threat reactions – this feature makes it important to tap a variety of face-to-face interactions on different social sites.

We included a number of relevant controls at the individual level: gender (with males as reference category), age (in years, ranging from 18–89), education (treated as metric and subsequently recoded to range from 0 to 1, higher values indicating higher education), occupation (as dummies), and self-reported ideology (treated as metric and recoded to vary from 0 to 1, higher values indicating right-wing orientation).¹⁰ Personal income was excluded because this measure in the ESS data set has a massive number of missing cases (13,353), which makes it highly unreliable.

The empirical analyses include no controls at the national level as we apply a fixed-effects regression analysis specification. This choice of statistical modeling reflects the fact that the specified hypothesis relates exclusively to the individual level of social interaction, which makes it all the more urgent to ensure that macro variables have no influence on the estimated individual-level relationships.¹¹ The fixed-effects regression specification explicitly deals with omitted variable bias. In our case, this means that macro-level influences are eliminated by controlling for country dummies, thus leaving controlled estimates at the individual level. In other words, the country dummies explain all variance at the country level, implying that there is no variance left to be explained by additional country-level variables (Allison, 2009). Consistent with conventional standards, the country dummy coefficients are not reported.

Table 2 reports the correlations between all our predictors. It shows that the predictors are not strongly correlated, except for the fact that education and occupation are strongly related. This is no surprise, and these variables only serve as control variables in our study. On the whole, the predictors identify distinctive characteristics as they are not overlapping to any greater extent. Moreover, and most importantly, the first column in Table 2 shows that personal economic vulnerability is only weakly associated with other indicators of socioeconomic status (i.e. education and occupation). This result supports the claim that our measure of personal economic vulnerability taps distinctive information that remains unidentified by other socioeconomic indicators, such as education or occupation in particular. In other words, the higher educated (or middle-class members) are not necessarily very wealthy, and the lower educated (or working-class members) are not necessarily very poor. Thus, asking people about the extent to which they live comfortably/uncomfortably on their current income is probably a more accurate measure of personal economic vulnerability than education and occupation.

Table 2.

Correlations between predictors.

	PEV	Contact	Education	Left–right	Age	Gender	Occupation
PEV	-
Contact	−0.131	-
Education	−0.253	0.259	-
Left–right	−0.017	−0.053	−0.009	-
Age	0.034	−0.246	−0.237	0.047	-
Gender	0.074	−0.071	−0.049	−0.025	0.019	-
Occupation	0.239	−0.192	−0.481	−0.018	0.014	0.019	-

PEV: personal economic vulnerability.

All coefficients are statistically significant at p < 0.05.

Empirical results

Turning to the test of our constrained self-interest hypothesis, Table 3 reports the main findings organized into two columns. Model 1 shows the effects of intergroup contact and personal economic vulnerability after various extraneous variables have been controlled. Model 2 shows the relevant interaction term meant to test our hypothesis. Focusing on Model 1, it appears that the effects of intergroup contact and personal economic vulnerability are diametrically opposed. Consistent with previous research (Pettigrew et al., 2007), intergroup contact is negatively related to perceived out-group threat, implying that those who are in contact with out-group members perceive them as less threatening than those who have no such contact.

Table 3.

The relationship between personal economic vulnerability and perceived out-group threat, including the interacting effect of intergroup contact.

	Model 1 (without interaction)	Model 2 (interaction)
Economic vulnerability	0.587*** (0.034)	0.811*** (0.043)
Intergroup contact	−1.146***(0.031)	−0.891*** (0.043)
Economic vulnerability × contact		−0.883*** (0.103)
Occupational class (F-test)	***	***
Upper middle class	−0.158***(0.038)	−0.159*** (0.038)
Lower middle class	0.043(0.033)	0.049 (0.033)
Working class	0.255*** (0.035)	0.258*** (0.035)
Miscellaneous	0.038 (0.044)	0.035 (0.044)
Education	−0.736*** (0.033)	−0.734*** (0.033)
Age	−0.001* (0.001)	−0.001* (0.001)
Gender	0.021 (0.017)	0.020 (0.017)
Left–right self-placement	0.697*** (0.040)	0.700*** (0.040)
Intercept	5.549*** (0.055)	5.473*** (0.055)
N (respondents)	33,784	33,784
N (countries)	21	21

Unstandardized coefficients, with standard errors in parentheses (fixed-effects model). Reference categories for categorical variables are: occupational class (= upper class) and gender (= male).

p < 0.05;***p < 0.001 (two-tailed t-tests).

In contrast, and consistent with previous studies (e.g. Brader and Valentino, 2007), personal economic vulnerability is positively related to perceived out-group threat. As personal financial hardship increases, perceived out-group threat intensifies. Apparently, personal economic vulnerability (i.e. self-interest) is important when other relevant characteristics are controlled.¹² Some of the control-variable effects support this conclusion on the importance of personal economic vulnerability. In particular, (1) working-class members perceive out-group member as more threatening than upper-class members, and (2) the higher educated perceive out-group members as less threatening than the lower educated. Both class position and education are commonly regarded as indicators of economic privileges (e.g. Coenders and Scheepers, 2003; Bobo and Hutchings, 1996; McLaren, 2003; Schneider, 2008). Yet, as previously argued, these indicators are not by themselves sufficient to tap real personal vulnerability; this is also vindicated by the fact that personal economic vulnerability has an independent effect when other socioeconomic indicators have been controlled (see Table 3).

Although self-interest theory has been supported, this is evidently a ceteris paribus conclusion – assuming that the effect of personal economic vulnerability is additive as well as mentally unconstrained. In contrast, we wish to examine the reactions of self-interested vulnerable persons when they interact with out-group members who they perceive as threatening competitors. Model 2 in Table 3 shows the relevant findings by presenting the estimated effect of the multiplicative term (personal economic vulnerability × intergroup contact). Apart from being statistically significant, the sign of the interaction coefficient brings the first important piece of information. In our case, the interaction coefficient is negative in sign (−0.883), and this needs to be related to the positive sign of personal economic vulnerability (0.811). Substantially, this means that the ability of personal economic vulnerability to increase perceived out-group threat is weakened when intergroup contact occurs. That is, when personally vulnerable in-group members interact with out-group members, they perceive them as less threatening. In contrast, when vulnerable in-group members do not interact with out-group members, they consider them as more threatening.

Consistent with conventional guidelines (Brambor et al., 2006), Figure 1 visualizes the substantive meaning of the multiplicative term in Table 3. The straight line in Figure 1 thus plots the estimated marginal conditional effect of personal economic vulnerability on perceived out-group threat as a function of intergroup contact intensity, and the 95 percent confidence intervals reveal the uncertainty associated with these estimated effects. Accordingly, Figure 1 also shows whether the marginal effects of personal economic vulnerability are significant at all values of intergroup contact intensity (shown on the x axis). The estimated coefficient on the interaction term from Table 3 gives our estimate of the slope of the marginal effect line (−0.883), indicating that the marginal effect of personal economic vulnerability on perceived out-group member threat is estimated to decrease at a rate of about 0.90 for each one-unit increase in intergroup contact intensity (i.e. when contact shifts from 0 to 1). More substantially, Figure 1 shows that personal economic vulnerability has a strong positive impact on perceived out-group member threat when intergroup contact does not occur. This result makes perfect sense as these particular in-group members are exclusively motivated by their self-interest – because personal contact stimuli are inactive. Thus, when contact is absent, the impact of self-interest is mentally unconstrained.

Figure 1.

The marginal effect of personal economic vulnerability (PEV) on perceived out-group threat, conditional on intergroup contact intensity.

Yet, Figure 1 also shows that the impact of personal economic vulnerability on perceived out-group threat is certainly not additive. As intergroup contact intensifies, the impact of personal economic vulnerability becomes smaller and smaller. Indeed, and most importantly, the impact of personal economic vulnerability on perceived out-group threat is statistically indistinguishable from zero when intergroup contact intensity approaches its maximum value (= 1). Put differently, when the contact stimulus is strong, the impact of self-interest is overpowered. Thus, the marginal effect analysis in Figure 1 supports our hypothesis.

Recall that our constrained self-interest hypothesis involves an additional implication: Is the contact effect strongest among the most economically vulnerable in-group members? Figure 2 examines this issue by showing the marginal effect of intergroup contact on perceived out-group threat, conditional on personal economic vulnerability. The dotted lines show the 95 percent confidence intervals around the marginal estimates, and they indicate that intergroup contact reduces perceived out-group threat irrespective of the extent of personal economic vulnerability. Most importantly for our purposes, Figure 2 shows that the impact of intergroup contact strengthens considerably as personal economic vulnerability shifts from its minimum to its maximum value. In fact, the ability of contact to reduce out-group threat is clearly strongest among the most economically vulnerable, and the interaction coefficient from Table 3 informs that the contact effect among the most vulnerable is also significantly different from the least vulnerable in statistical terms (given that personal economic vulnerability has been rescaled to vary from 0 to 1).

Figure 2.

The marginal effect of intergroup contact intensity on perceived out-group threat, conditional on personal economic vulnerability.

Is this also a strong effect in substantial terms? Additional calculations show that the difference in terms of perceived out-group threat between non-contacted and contacted in-group members among the least vulnerable corresponds to 0.53 standard deviations in the perceived out-group threat index; the difference, however, between non-contacted and contacted among the most vulnerable corresponds to 1.05 standard deviations in the perceived out-group member threat index. Although there is no objective answer to what constitutes a large effect, we follow the guidelines suggested by Cohen (1988) stating that an effect size of 0.20 (standard deviations) is a small effect, an effect size of 0.50 is a medium effect, and an effect size of 0.80 is a large effect. Based on these guidelines, our effect size among the most vulnerable of more than 1 standard deviation certainly qualifies as a large effect. These results also support our hypothesis.¹³

In sum, it appears that intergroup contact has the ability to prevent the impact of personal economic vulnerability on perceived out-group threat.¹⁴ Likewise, the contact effect is strongest among the most economically vulnerable. Building on previous meta-analyses, these results suggest that self-interest can be effectively overpowered by cognitive learning and affective ties, referring to the main components of intergroup contact situations. In combination, Model 2 in Table 3 and Figures 1 and 2 support the constrained self-interest hypothesis.¹⁵

Conclusion and discussion

Two previous studies reported that the influence of the share of immigrants on perceived out-group threat is weakened by intergroup contact (McLaren, 2003; Schneider, 2008). Both studies primarily examined the so-called group competition theory dating back to the work of Blumer (1958). Our article expands this type of research by applying the narrower notion of self-interest measured as personal economic vulnerability among in-group members. We prefer the narrower version in order to keep focusing on personal advantages and disadvantages. Moreover, our data indicate that aggregate notions, such as ‘the interests of the majority’, are somewhat problematic. Indeed, as occupation, education, and personal economic vulnerability all relate to perceived out-group threat, there must be different interests within the majority.

The main finding is that the impact of personal economic vulnerability on perceived out-group threat is constrained by personal interaction. This means that personal economic vulnerability is conducive to perceived out-group threat as long as no contact occurs between in-group and out-group members; however, when intense contact is established, the impact of personal economic vulnerability on perceived out-group threat disappears. Substantially, we believe that this finding makes a significant contribution to intergroup contact research. With a few notable exceptions (Dixon, 2006; McLaren, 2003; Schneider, 2008), this field has neglected intergroup contact as a boundary condition of various associations between social background, threatening contexts, and attitudes (see, for example, Kenworthy et al., 2005; Pettigrew and Tropp, 2005). This situation is puzzling since actual intergroup contact identifies specific interactional contexts that encourage social cognition and stimulate affective involvement between group members including ‘personalization’ of out-group members. In the absence of contact, out-group members remain remote and ‘depersonalized’ – characteristics that most likely nurture perceived out-group threat among those who are, in fact, most vulnerable (see also Kinder and Mendelberg, 1995). In this manner, intergroup contact appears as a significant constraint on the attitudinal consequences of individual-level social characteristics. Adding this perspective to intergroup contact theory expands its social relevance considerably.

We have tried to go even deeper into the mental capacity of intergroup contact compared to the influence of self-interest and its personal economic vulnerability correlate. Indeed, had the contact effect been weakest among the most vulnerable, this would have been testimony to the fact that the mental capacity of contact is rather limited. However, analyses showed that the impact of intergroup contact on perceived out-group threat is actually strongest among the most vulnerable. Thus, the mental capacity of contact is remarkable, and this specific finding is obviously supportive of recent calls for demonstrating that intergroup contact also works among those who are most likely to reject out-group members (e.g. Hodson, 2011).

Where does this leave the simple self-interest theory? Our results suggest that the concept of self-interest in studies of interethnic relations is underspecified as self-interested reactions can be overpowered by cognitive learning and affective ties. This is not to argue that self-interest theory is fundamentally misguided. Rather, the impact of self-interested reactions is easily overestimated because the theory ignores some essential mental constraints on the human mind. Intergroup contact taps some of these constraints, and they need to be built into self-interest theory.

All studies have weaknesses, including ours. As suggested by Sears and Funk (1990), demographic measures and subjective measures of personal economic hardship are relatively poor indicators of self-interest. According to their view, indicators of self-interest should measure tangible personal consequences of specific attitude objects (e.g. state policies). Otherwise it becomes difficult to distinguish between prejudice and rational judgments when specific reactions and associations are examined. Surely, as an indicator of self-interest, our measure of personal economic vulnerability is not ideal. Ultimately, the observed association between personal economic vulnerability and perceived out-group threat might suggest that the most vulnerable are simply more prejudiced than the least vulnerable. In other words, is personal economic vulnerability just another indicator of prejudice – rather than an indicator of objective personal circumstances? The empirical analyses, however, indicate that personal economic vulnerability is not just a measure of prejudice toward out-group members. Noticeably, the effect of personal economic vulnerability remains sizeable after age, education, and occupation have been controlled (see Model 1 in Table 3). Certainly, any of these controls could have made the effect of personal economic vulnerability spurious, most notably education, which is commonly considered the strongest predictor of prejudice (e.g. Case et al., 1989; Schuman et al., 1997; Sniderman and Piazza, 1993; Virtanen and Huddy, 1998; Wagner and Zick, 1995). This means that personal financial hardship has a genuine effect across demographic categories. This resistance toward numerous demographic controls suggests that the personal economic vulnerability measure taps rational judgments rather than dispositional reactions.

Having defended the validity of our key measure, we conclude that intergroup contact has an important role in examining the consequences of self-interest on perceived out-group threat. Our results also suggest that perceived out-group threat has more than an economic foundation. When contact is established, empathy, anxiety, and knowledge most probably influence the extent to which the most vulnerable perceive out-group members as threatening. Accordingly, the self-interest theory needs to recognize that considerations about personal advantages or disadvantages are not necessarily dominant; in fact, they can be overpowered by the cognitive and affective components of intergroup contact. Finally, we do not intend to create an artificial fence between the present investigation and previous studies which have used other measures in their attempt to examine intergroup contact as a boundary condition. Indeed, the overarching conclusion is that intergroup contact appears to have broad importance for blunting the tendency of unfavorable circumstances to nurture racial and ethnic animosity.

Footnotes

Acknowledgements

A previous version of this paper was presented at the Annual Meeting of the International Society of Political Psychology, Rome, 2014. The authors wish to thank the participants for comments. We would also like to thank the reviewers and the Editor for many useful suggestions.

Declaration of Conflicting Interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Allison

(2009) Fixed Effects Regression Models. London: SAGE.

Allport

([1954] 1979) The Nature of Prejudice. Cambridge, MA: Perseus Books.

Batson

Polycarpou

Harmon

. (1997) Empathy and attitudes: Can feeling for a member of a stigmatized group improve feelings toward the group? Journal of Personality and Social Psychology 72(1): 105–118.

Berry

Golder

Milton

(2012) Improving tests of theories positing interaction. Journal of Politics 74(3): 653–671.

Blumer

(1958) Race prejudice as a sense of group position. Pacific Sociological Review 1: 3–7.

Bobo

Hutchings

(1996) Perceptions of racial group competition: Extending Blumer’s theory of group position to a multiracial social context. American Sociological Review 61(6): 951–972.

Brader

Valentino

(2007) Identities, interests, and emotions: Symbolic versus material wellsprings of fear, anger, and enthusiasm. In: Neuman

Marcus

Crigler

. (eds) The Affect Effect. Chicago, IL: The University of Chicago Press, pp. 180–201.

Brambor

Clark

Golder

(2006) Understanding interaction models: Improving empirical analyses. Political Analysis 14(1): 63–82.

Case

Greeley

Fuchs

(1989) Social determinants of racial prejudice. Sociological Perspectives 32(4): 469–483.

10.

Chong

Citrin

Conley

(2001) When self-interest matters. Political Psychology 22(3): 541–570.

11.

Citrin

Green

(1990) The self-interest motive in American public opinion. Research in Micropolitics 3: 1–28.

12.

Coenders

Scheepers

(1998) Support for ethnic discrimination in the Netherlands 1979–1993: Effects of period, cohort, and individual characteristics. European Sociological Review 14(4): 405–422.

13.

Coenders

Scheepers

(2003) The effect of education on nationalism and ethnic exclusionism: An international comparison. Political Psychology 24(2): 313–343.

14.

Cohen

(1988) Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: Lawrence Erlbaum Associates.

15.

Dixon

(2006) The ties that bind and those that don’t: Toward reconciling group threat and contact theories of prejudice. Social Forces 84(4): 2179–2204.

16.

Dovidio

Gartner

Kawakami

(2003) Intergroup contact: The past, present, and the future. Group Processes & Intergroup Relations 6(6): 5–21.

17.

Eisenberg

Eggum

Giunta

(2010) Empathy-related responding: Associations with prosocial behavior, aggression, and intergroup relations. Social Issues and Policy Review 4(1): 143–180.

18.

European Social Survey (ESS) (2011) ESS-2002 Documentation Report, Edition 6.2. The ESS Data Archive.

19.

Fazio

Chen

McDonel

. (1982) Attitude accessibility, attitude-behavior consistency and the strength of the object-evaluation association. Journal of Experimental Social Psychology 18(4): 339–357.

20.

Gundelach

(2014) In diversity we trust: The positive effect of ethnic diversity on outgroup trust. Political Behavior 36(1): 125–142.

21.

Hainmueller

Hiscox

(2010) Attitudes toward highly skilled and low-skilled immigration: Evidence from a survey experiment. American Political Science Review 104(1): 61–84.

22.

Hayes

Dowds

(2006) Social contact, cultural marginality or economic self-interest? Attitudes towards immigrants in Northern Ireland. Journal of Ethnic and Migration Studies 32(3): 455–476.

23.

Hello

Scheepers

Sleegers

(2006) Why the more educated are less inclined to keep ethnic distance: An empirical test of four explanations. Ethnic and Racial Studies 29(5): 959–985.

24.

Hernes

Knudsen

(1992) Norwegians’ attitudes toward new immigrants. Acta Sociologica 35(2): 123–139.

25.

Hodson

(2011) Do ideologically intolerant people benefit from intergroup contact? Current Directions in Psychological Science 20(3): 154–159.

26.

Huckfeldt

Mondak

Hayes

. (2013) Networks, interdependence, and social influence in politics. In: Huddy

Sears

Levy

(eds) The Oxford Handbook of Political Psychology. Oxford: Oxford University Press, pp. 662–698.

27.

Inglehart

(1997) Modernization and Postmodernization. Princeton, NJ: Princeton University Press.

28.

Kenworthy

Turner

Hewstone

. (2005) Intergroup contact: When does it work and why? In: Dovidio

Glick

Rudman

(eds) On the Nature of Prejudice: Fifty Years After Allport. Malden, MA: Blackwell, pp. 278–292.

29.

Kinder

Mendelberg

(1995) Cracks in American apartheid: The political impact of prejudice among desegregated whites. Journal of Politics 57(2): 402–424.

30.

Kluegel

Smith

(1983) Affirmative action attitudes: Effects of self-interest, racial affect, and stratification beliefs on whites’ views. Social Forces 61(3): 797–824.

31.

Koschate

Van Dick

(2011) A multilevel test of Allport’s contact conditions. Group Process & Intergroup Relations 14(6): 769–787.

32.

Mackuen

Marcus

Neuman

. (2007) The third way: The theory of affective intelligence and American democracy. In: Neuman

Marcus

Crigler

. (eds) The Affect Effect. Chicago, IL: Chicago University Press, pp. 124–151.

33.

McLaren

(2003) Anti-immigrant prejudice in Europe: Contact, threat perception, and preferences for the expulsion of migrants. Social Forces 81(3): 909–936.

34.

Marcus

Neuman

Mackuen

(2000) Affective Intelligence and Political Judgment. Chicago, IL: The University of Chicago Press.

35.

Miller

Smith

Mackie

(2004) Effects of intergroup contact and political predispositions on prejudice: Role of intergroup emotions. Group Processes & Intergroup Relations 7(3): 221–237.

36.

Miller

(2002) Personalization and the promise of contact theory. Journal of Social Issues 58(2): 387–410.

37.

Mondak

Mutz

(2006) The workplace as a context for cross-cutting political discourse. Journal of Politics 68(1): 140–155.

38.

Monroe

(2001) Paradigm shift: From rational choice to perspective. International Political Science Review 22(1): 151–172.

39.

Mutz

(2002) Cross-cutting social networks: Testing democratic theory in practice. American Political Science Review 96(1): 111–126.

40.

Nagayoshi

Hjerm

(2015) Anti-immigration attitudes in different welfare states: Do types of labor market policies matter? International Journal of Comparative Sociology. Epub ahead of print 26 June 2015. DOI: 10.1177/0020715215591379.

41.

Napier

Jost

(2008) The ‘antidemocratic personality’ revisited: A cross-national investigation of working-class authoritarianism. Journal of Social Issues 64(3): 595–617.

42.

Norris

(2004) From the civic culture to the Afrobarometer. Newsletter of the American Political Science Association Organized Section in Comparative Politics. Available at: http://www.hks.harvard.edu/fs/pnorris/Acrobat/APSA-CP.pdf

43.

Page-Gould

Mendoza-Denton

Tropp

(2008) With a little help from my cross-group friend: Reducing anxiety in intergroup contexts through cross-group friendship. Journal of Personality and Social Psychology 95(5): 1080–1094.

44.

Pagotto

Voci

Maculan

(2010) The effectiveness of intergroup contact at work: Mediators and moderators of hospital workers’ prejudice toward immigrants. Journal of Community & Applied Social Psychology 20(4): 317–330.

45.

Pettigrew

(1998) Intergroup contact theory. Annual Review of Psychology 49: 65–85.

46.

Pettigrew

Tropp

(2005) Allport’s intergroup contact hypothesis: Its history and influence. In: Dovidio

Glick

Rudman

(eds) On the Nature of Prejudice: Fifty Years after Allport. Malden, MA: Blackwell, pp. 262–277.

47.

Pettigrew

Tropp

(2006) A meta-analytic test of intergroup contact theory. Journal of Personality and Social Psychology 90(5): 751–783.

48.

Pettigrew

Tropp

(2008) How does intergroup contact reduce prejudice? Meta-analytic test of three mediators. European Journal of Social Psychology 38(6): 922–934.

49.

Pettigrew

Tropp

(2011) When Groups Meet: The Dynamics of Intergroup Contact. New York: Psychology Press.

50.

Pettigrew

Christ

Wagner

. (2007) Direct and indirect intergroup contact effects on prejudice: A normative interpretation. International Journal of Intercultural Relations 31(4): 411–425.

51.

Pichler

(2010) Foundations of anti-immigrant sentiment: The variable nature of perceived group threat across changing European societies 2002–2006. International Journal of Comparative Sociology 51(6): 445–469.

52.

Pickett

Baker

Metcalfe

. (2014) Contact and compromise: Explaining support for conciliatory measures in the context of violent intergroup conflict. Journal of Research in Crime and Delinquency 51(5): 585–619.

53.

Quillian

(1995) Prejudice as a response to perceived group threat: Population composition and anti-immigrant and racial prejudice in Europe. American Sociological Review 60(4): 586–611.

54.

Raijman

Semyonov

Schmidt

(2003) Do foreigners deserve rights? Determinants of public views toward foreigners in Germany and Israel. European Sociological Review 19(4): 379–392.

55.

Riek

Mania

Gaertner

(2006) Intergroup threat and outgroup attitudes: A meta-analytic review. Personality and Social Psychology Review 10(4): 336–353.

56.

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(1): 17–34.

57.

Scheve

Slaughter

(2001) Globalization and the Perceptions of American Workers. Washington, DC: Institute for International Economics.

58.

Schneider

(2008) Anti-immigrant attitudes in Europe: Outgroup size and perceived ethnic threat. European Sociological Review 24(1): 53–67.

59.

Schuman

Steeh

Bobo

. (1997) Racial Attitudes in America: Trends and Interpretations (rev. edn). Cambridge, MA: Harvard University Press.

60.

Sears

Funk

(1990) The limited effect of economic self-interest on the political attitudes of the mass public. Journal of Behavioral Economics 19(3): 247–271.

61.

Sears

Funk

(1991) The role of self interest in social and political attitudes. Advances in Experimental Social Psychology 24: 1–91.

62.

Sears

Hensler

Speer

(1979) Whites’ opposition to ‘busing’: Self-Interest or symbolic politics? American Political Science Review 73(2): 669–384.

63.

Semyonov

Raijman

Gorodzeisky

(2008) Foreigners’ impact on European societies. Public views and perceptions in a cross-national comparative perspective. International Journal of Comparative Sociology 49(1): 5–29.

64.

Semyonov

Raijman

Tov

. (2004) Population size, perceived threat, and exclusion: A multiple-indicators analysis of attitudes toward foreigners in Germany. Social Science Research 33(4): 681–701.

65.

Sides

Citrin

(2007) European opinion about immigration: The role of identities, interests and information. British Journal of Political Science 37(3): 477–504.

66.

Sniderman

Hagendoorn

(2007) When Ways of Life Collide. Princeton, NJ: Princeton University Press.

67.

Sniderman

Piazza

(1993) The Scar of Race. Cambridge, MA: The Belknap Press of Harvard University Press.

68.

Stephan

Diaz-Loving

Duran

(2000) Integrated threat theory and intercultural attitudes: Mexico and the United States. Journal of Cross-Cultural Psychology 31(2): 240–249.

69.

Stephan

Renfro

Davis

(2008) The role of threat in intergroup relations. In: Wagner

Tropp

Finchilescu

. (eds) Improving Intergroup Relations: Building on the Legacy of Thomas F. Pettigrew. Malden, MA: Blackwell, pp. 55–72.

70.

Swart

Hewstone

Christ

. (2010) The impact of crossgroup friendships in South Africa: Affective mediators and multigroup comparisons. Journal of Social Issues 66(2): 309–333.

71.

Tausch

Tam

Hewstone

. (2007) Individual-level and group-level mediators of contact effects in Northern Ireland: The moderating role of social identification. British Journal of Social Psychology 46(3): 541–556.

72.

Tolnay

Beck

(1992) Toward a threat model of southern black lynchings. In: Liska

(ed.) Social Threat and Social Control. Albany, NY: State University of New York Press, pp. 33–52.

73.

Vezzali

Giovannini

(2012) Secondary transfer of intergroup contact: The role of intergroup attitudes, intergroup anxiety and perspective taking. Journal of Community & Applied Social Psychology 22(2): 125–144.

74.

Vezzali

Giovannini

Capozza

(2010) Longitudinal effects of contact on intergroup relations: The role of majority and minority group membership and intergroup emotions. Journal of Community & Applied Social Psychology 20(6): 462–479.

75.

Virtanen

Huddy

(1998) Old-fashioned racism and new forms of racial prejudice. Journal of Politics 60(2): 311–332.

76.

Wagner

Zick

(1995) The relation of formal education to ethnic prejudice: Its reliability, validity and explanation. European Journal of Social Psychology 25(1): 41–56.

77.

Weber

(2015) National and regional proportion of immigrants and perceived threat of immigration: A three-level analysis in Western Europe. International Journal of Comparative Sociology. Epub ahead of publication 10 February 2015. DOI: 10.1177/0020715215571950.