Contextualizing Intergroup Contact

Abstract

This article examines intergroup contact effects in different political contexts. We expand on previous efforts of social psychologists by incorporating the messages of political parties as a contextual trigger of group membership awareness in contact situations. We argue that the focus among political parties on us-them categorizations heightens the awareness of group memberships. This focus in turn enhances the positive intergroup contact effect by stimulating majority members to perceive contacted persons as prototypical outgroup members. A multilevel analysis of 22 countries and almost 37,000 individuals confirms that the ability of intergroup contact to reduce antiforeigner sentiment increases when political parties focus intensively on immigration issues and cultural differences. Specifically, both workplace contact and interethnic friendship become more effective in reducing antiforeigner sentiment when intergroup relations are politicized. These findings demonstrate the need for widening the scope of the intergroup contact theory in order to cover macro-political conditions.

Keywords

intergroup contact antiforeigner sentiment salience political parties

Numerous studies have consistently reported that face-to-face interaction leaves majority members less negative toward ethnic minority members (Davies et al. 2011; McLaren 2003; Pettigrew and Tropp 2006). Yet scholars emphasize that intergroup contact matters most in situations fulfilling key conditions, group membership salience being one such condition (Allport [1954] 1979; Brown, Vivian, and Hewstone 1999; Hewstone and Brown 1986; Pettigrew 1998). Group membership salience refers to individual awareness of group memberships, such as awareness of outgroup participants’ nationality or religious belonging. Previous studies have found that contact effects tend to be strongest when group membership salience is high (Brown et al. 2007; Ensari and Miller 2002; Paolini, Harwood, and Rubin 2010; Voci and Hewstone 2003). According to the so-called categorization model, high group membership salience ensures that positive contact experiences generalize to the whole outgroup.

Despite the accumulation of literature that confirms the significance of group membership salience, its variability remains puzzling (Pettigrew and Tropp 2011). In particular, the contextual causes conducive to group membership salience are unaddressed. Among these causes, political parties deserve special consideration. Some politicians welcome the transition to a multicultural society, while others call for assimilation (Rydgren 2008; Sniderman and Hagendoorn 2007). Thus, if political parties disclose cues that signal group differences, this may heighten group membership salience. However, no one has yet examined whether intergroup contact effects are in fact related to party politics (Pettigrew 1997, 2006; Pettigrew and Tropp 2006, 2008). This is surprising as scholars have emphasized that “the social-psychological dimension of intergroup relations must never be divorced from the political dimensions” (Hewstone 2009:280).

Consequently, this article expands previous research by introducing a context-oriented approach that relates intergroup contact to the macro political level. Building on political agenda-setting research (Green-Pedersen 2007; Hillygus and Shields 2008; Zaller 1992), we argue that ingroup participants become increasingly aware of group memberships when political parties make collective identities a focal point. In such a macro context, ingroup participants most likely perceive outgroup participants as prototypical group members, whereas outgroup participants tend to be perceived as less prototypical of their group when political parties ignore collective identities. When outgroup participants are perceived as prototypical, positive outcomes of contact settings will generalize to noncontacted outgroup members, whereas generalization fails when outgroup participants are perceived as atypical. Accordingly, we hypothesize that the greater the extent to which political parties focus on immigration issues, the more intergroup contact will reduce antiforeigner sentiment. This hypothesis addresses the joint influence of positive and negative political party cues.

As the categorization model is not commonly accepted (Pettigrew and Tropp 2011), we also test a key implication of the so-called “decategorization model,” which is commonly regarded as an alternative approach to understanding the role of group membership salience in intergroup contact settings. This model suggests that contact effects tend to be weak when group membership salience is high (Brewer and Miller 1984). Emphasis on group attributes may cause greater tensions as ingroup members fail to personalize contacted outgroup members. Moreover, scholars emphasize that ordinary citizens are more attentive to negative information than positive information (e.g., Soroka 2006). Indeed, political parties often focus on the negative consequences of immigration. Drawing on insights from the decategorization model and the “negativity bias” assumption, we derive a rival hypothesis that the greater the extent to which political parties focuses negatively on immigration issues, the less intergroup contact will reduce antiforeigner sentiment. This hypothesis addresses the importance of negative political party cues.

This article’s initial sections address group membership salience and how political parties influence salience in intergroup contact. Subsequently, we test the two rival hypotheses using individual-level and country-level data from 22 countries in a multilevel analysis. The individual-level data on contact and antiforeigner sentiment is from the European Social Survey, whereas the country-level data on political party focus on immigration is from the Comparative Manifestos Project.

The Group Membership Salience Puzzle

The intergroup contact theory emanates from a simple but fundamental idea: ingroup members tend to dislike outgroup members they do not know (Allport ([1954] 1979; Pettigrew 1998). When intimate and frequent contact occurs, most ingroup members realize that their negative expectations about “the other” were exaggerated. From this follows the basic hypothesis that intergroup contact reduces prejudice toward outgroup members (Allport [1954] 1979). Subsequent studies have shown that intergroup contact also relates to broader processes like a reduction in antiforeigner sentiment, which our article addresses (McLaren 2003; Pettigrew and Tropp 2011; Schneider 2008). Following previous research (Semyonov, Raijman, and Gorodzeisky 2006), we define antiforeigner sentiment as a blend of antipathy, varieties of perceived threat, and negative stereotypes.

Moreover, frequent contact, such as cross-group friendship or coworker interaction, involves both cognitive learning about outgroup member characteristics and affective ties between the participants (Davies et al. 2011; Pettigrew 1998; Pettigrew and Tropp 2006, 2008). Finally, the nature of contact also matters—commonly referred to as conditions that facilitate individual-to-group generalization (Pettigrew and Tropp 2011). Ideally, contact situations should embody equal status between the participants, common goals, cooperation, institutional support, and acquaintance potential. For instance, a slave-master relationship involves intense interaction but violates the principle of equal status. Likewise, a strong zero-sum–based competition between outgroup and ingroup members may also “poison” their mutual relationship because of too high personal stakes. In contrast, cross-group friendship meets all conditions. Workplace contact is also likely to meet most of these conditions as teamwork has become dominant in most modern workplaces.

Intergroup contact theory, however, confronts a critical issue: do immediate positive contact experiences among ingroup members generalize to the entire outgroup? The contact effect is trivial if ingroup members’ prejudgments of the outgroup are not modified (see Lee, Farrell, and Link 2004; Pettigrew 1998). Some scholars contend that the contact effect is likely to be weak if ingroup members do not reflect on collective outgroup issues. For the contact effect to be generalized, group characteristics, and outgroup characteristics in particular, must be salient (Voci and Hewstone 2003). Group membership salience is therefore defined as an individual’s awareness of group memberships in intergroup encounters (Ensari and Miller 2002). Hewstone and Brown (1986) have clarified the meaning of group membership salience by distinguishing between interpersonal and intergroup contact.

Interpersonal contact involves encounters between individuals who do not perceive each other as group members (Brown et al. 1999). In these circumstances, the effect of contact on reducing antiforeigner sentiment is weak because ingroup members are unaware of the importance of outgroup identity phenomena such as nationality, religious symbols, race, distinct clothing, or social status. Ultimately, the outcome of interpersonal contact is likely to correspond to the infamous “some of my best friends are black, but . . . ” aphorism (Jackman and Crane 1986). Evidently, face-to-face interaction is always interpersonal but need not trigger any broader awareness of group differences.

In contrast, intergroup contact should be seen as a potential extension of interpersonal contact. Intergroup contact occurs when the outgroup individual is consciously seen as a prototypical member of the outgroup rather than a unique individual (Harwood, Raman, and Hewstone 2006). Intergroup contact entails social identities (us-and-them), whereas interpersonal contact is confined to personal identities (you-and-me).

This framework, however, has two weaknesses. The first one relates to the plausibility of the notion of interpersonal encounters. After all, scholars (Pettigrew and Tropp 2006) have clearly established that the contact effect occurs in numerous and highly different contexts—across time and countries. This robustness indicates that contact situations are probably never purely interpersonal since some outgroup member attributes are always observable such as skin color, distinct habits, group jargon, or accents (Miller 2002). Accordingly, it seems appropriate to redefine intergroup contact as a generic term covering a variety of contact situations that differ in terms of group membership salience. That is, group membership salience can be low, implying that ingroup participants irregularly pay attention to group identities. Or group membership salience can be high, implying that ingroup participants regularly perceive outgroup participants as prototypical of their group.

The other weakness is fundamental as it concerns whether group membership salience is itself in need of explanation. Despite its undeniable relevance, the concept of group membership salience remains puzzling: what factors are conducive to group membership salience? This question takes on special theoretical importance, as scholars tend to refer to psychological mechanisms at the individual level without relating group membership salience to its political context (e.g., Brown and Hewstone 2005; Brown et al. 1999; Pettigrew and Tropp 2011). Obviously, group membership salience is an individual-level psychological phenomenon, but group membership salience variability can hardly be comprehended independently of the political environment. Indeed, previous research has indicated that ordinary citizens are responsive to situational triggers such as the relative salience of national identity or of economic versus cultural threats (e.g., Brader, Valentino, and Suhay 2008; Sniderman, Hagendoorn, and Prior 2004). Yet, the relative salience of issues that relate to group differences is strongly influenced by the agenda-setting priorities of political parties, as emphasized by agenda-setting scholars (e.g., Green-Pedersen and Krogstrup 2008). To justify these claims, the nature of party competition and the role of political parties as sources of information must be clarified.

Political Parties Stimulate Group Membership Salience

Positive and Negative Political Party Cues

Political parties compete in order to attract votes (Downs 1957). To survive in the competitive struggle, political parties must attract public attention and do so by airing messages and cues that clarify their key issue positions. More importantly, political party competition is a relational process that escapes the immediate control of any individual party (Green-Pedersen 2007). A political party can ignore some issues but must address those its competitors consider important. In the United States, the Republicans often challenge the Democrats by calling for harsher immigration policies (Kinder and Kam 2009). In European multi-party systems, extreme right-wing parties typically initiate debates about immigration, which forces the established parties to respond by clarifying their stands on immigration issues. Considering that immigration issues usually trigger major disagreements between left-wing and right-wing parties, mass media coverage is often also extensive. In effect, and as emphasized by Carmines and Stimson (1986), political parties are key actors in the process of bringing issues to the public attention; they mobilize issues through press releases, TV and radio performances, parliamentary debates, party manifestos, and party conferences (Hillygus and Shields 2008).

But do ordinary citizens really care about party messages? The prevailing scholarly view is that most citizens are inattentive to most political issues most of the time (Hutchings 2001; Sniderman, Brody, and Tetlock 1991). Likewise, ordinary citizens’ knowledge of political issues is minimal (Carpini and Keeter 1996), and scholars emphasize that ordinary citizens find it difficult to relate their daily life experiences to political issues.

Yet, citizens are neither totally inattentive nor incompetent. Because of their cognitive limitations, most people instinctively rely on external sources of information in order to make reasoned judgments. Zaller (1992) suggests that ordinary citizens pay attention to the information offered by elites, including top-level politicians. Similarly, Hopkins (2011) emphasizes that political parties are key politicizing agents who frequently encourage people to connect their own experiences and discoveries to national politics. In particular, Hopkins (2011) argues that political parties assist citizens by priming some issues at the expense of others, thus reducing real-world complexity. This line of reasoning suggests that political parties play an important role in determining the objects of mass opinion formation.

Yet, citizens are somewhat attentive to political party messages, but with what effect on intergroup contact? Building on Harwood (2010), we understand intergroup contact as a recurring process in which the individual’s focus on group attributes may be strong or weak. Accordingly, there must necessarily be conditions external to the immediate contact situation that account for this variation. Consider two different political contexts of intergroup contact, for the sake of illustration. The first is a context in which political parties never prime ethnic group issues in public, leaving the impression among ingroup members that outgroup members are harmless and certainly not threatening. The second is a context in which political parties intensively air their positive or negative stands on immigrants and immigration in public. In this particular context, it appears necessary and legitimate to care about outgroup issues. The political parties do not stimulate awareness of immigrants as group members in the first case, whereas they provoke reflections on us-and-them in the second. In particular, when political parties make collective identities a focal point, they may stimulate ingroup members to reflect on the (mis)match between personal experiences with outgroup members and outgroup stereotypes such as threat images.

However, it is important to understand the nature of political party focus on public issues. In democratic societies, citizens are exposed to opposing views rather than one organizing message (Sniderman and Theriault 2004). Specifying the dynamics of party competition, there can be little doubt that extreme right-wing parties—but also the traditional right-wing parties—emphasize the negative consequences of immigration. They focus on “them” (Carter 2005). As a competitive response, left-wing and center-right parties de-emphasize the negative consequences of immigration while calling for a sober second thought among “us” ingroup members (Bale 2008). Consequently, when political parties intensively address the immigration issue, us-and-them categorizations are likely to become dominant in the public debate. Assuming that people pay some attention to party cues, we believe that such a political climate stimulates interacting participants to perceive each other as group members rather than unique individuals. In such a politicized context where group membership is highly salient, contact experiences with individuals that disconfirm negative expectations thus reduces antiforeigner sentiment toward the outgroup as a whole. This brings us to our first hypothesis, which emphasizes the joint influence of all political party cues:

Hypothesis 1. The political enhancement effect: The greater the extent to which political parties focus on immigration issues, the more intergroup contact will reduce antiforeigner sentiment.

Negative Political Party Cues and Threat

Although we have emphasized that the public in democratic societies is exposed to opposing views from political parties about the consequences of immigration, negative threat messages might deserve special attention. Indeed, real party politics is very much about identifying “the bad guys” who cause trouble according to some predefined yardstick. Political parties underline their own democratic importance by producing such causal stories that ultimately make specific groups responsible for “undesirable” consequences (Stone 1989).

For instance, some political parties emphasize that immigration causes cultural disintegration, crime, misuse of social welfare programs, and job redundancies (Alonso and da Fonseca 2011). Such threat messages may have important consequences, according to some public opinion scholars. Specifically, Soroka (2006) has argued that peoples’ responses to positive and negative information are likely to be asymmetric due to cognitive weighting. Extreme and negative information has more influence on the average citizen than positive information. People tend to focus on information that is novel that most frequently happens to be negative—rather than positive news that usually confirms the status quo. This argument is consistent with social psychological research, which concludes that negative information carries more mental weight than positive information in the formation of impressions of others (Lau 1982). Negative information stimulates the human surveillance system whose principal function is to respond to signals of threat. Threat in turn elicits unease and anxiety. The human bias in favor of focusing on negative news (or disruptions) is most probably instinctive as it has presumably been crucial for survival and evolution (Marcus, Neuman, and Mackuen 2000).

Bringing these considerations into the framework of intergroup contact research calls for the so-called “decategorization model” introduced by Brewer and Miller (1984). Their model of interethnic contact is particularly influenced by social identity theory according to which strong collective identities combined with a need for positive self-identity tend to produce loyalty toward one’s own group as well as disloyalty toward most other groups. Accordingly, unlike Hewstone and Brown (1986), they argue that high group membership salience weakens the ability of intergroup contact to reduce antiforeigner sentiment. If ingroup members are very aware of group membership differences, they also tend to neglect their positive experiences with specific outgroup members. Consequently, Brewer and Miller (1984) claim that intergroup contact is most effective when individual identities are emphasized as this feature ultimately prevents undifferentiated negative stereotyping. Put differently, high group membership salience stimulates ingroup member participants to put outgroup member participants into preexisting categories.

Although we cannot test the Brewer and Miller (1984) model in detail, its key insight relates easily to the arena of political parties. Certain types of political party rhetoric may be particularly conducive to extreme category-based social identities. This is so because not only the average citizen but also ingroup member participants are attentive to threat messages from political parties, which most frequently confirm immigrant stereotypes such as “immigrants undermine our society.” Certainly, positive messages do occur, but they may have much less influence on the citizenry. Thus, if the human mind has an inbuilt negativity bias, it follows that positive personal experiences with immigrants may to a large extent be overpowered by an intense stream of negative threat messages. Negative messages stimulate ingroup member participants to treat outgroup member participants as “exceptions to the rule,” and such subtyping necessarily weakens the generalization effect of contact. Combining insights from the decategorization model with the potential power of negativity bias among ingroup members who interact with outgroup members, we get the following hypothesis:

Hypothesis 2. The political reduction effect: The greater the extent to which political parties focus negatively on immigration issues, the less intergroup contact will reduce antiforeigner sentiment.

Data, Measures, and Model Specifications

To test these two hypotheses, we have combined individual-level data on contact and antiforeigner sentiment with country-level data on political party focus in a multilevel approach. We estimate the effect of contact on antiforeigner sentiment at different levels of political party focus on immigrant issues. The present investigation thus treats nations as the contextual unit in which political agenda setting and competition affect group membership salience at the individual level. The choice of the national level rather than counties, regions, or any other contextual unit reflects our attempt to examine the role of political parties. Political parties almost by definition strive to be national agents, as their long-term survival depends on their ability to attract attention from the electorate as a whole. Likewise, immigration is usually established as an issue relating to the nation rather than any particular corner of society.

The analysis uses between-country variation in political party focus as we assume individual-level group membership salience to increase as a consequence of the political parties’ enduring focus on outgroups (we will test this assumption in the following). Accordingly, we combine data from the European Social Survey (ESS) and the Comparative Manifestos Project (CMP), which allows the examination of the individual-level effect of contact across varying levels of political party focus. We have included country-level data on political party focus from the 22 countries included in the first round of the ESS. Presented in Table 1, all of the countries are current members of the Organization for Economic Co-operation and Development (OECD), and Israel is the only non-European country.

Table 1.

Antiforeigner Sentiment (AFS), Contact, and Political Party Focus on Immigration, by Country

		Percentage distribution of contact (none/few/several)		Political party focus
Country	Mean level of AFS	Colleagues	Friends	Total	Negative
Sweden	3.83	57/27/16	33/41/26	.04	.00
Luxembourg	4.65	65/18/18	24/42/34	1.63	.83
Switzerland	4.73	50/29/21	27/46/27	1.37	.37
Spain	4.74	78/18/3	64/30/06	2.71	.00
Norway	4.81	60/24/16	46/35/19	1.75	.29
Finland	4.82	81/16/3	58/35/07	1.74	.08
Denmark	4.82	60/33/7	54/39/07	4.36	2.16
Ireland	4.85	68/24/7	58/35/07	.91	.00
Italy	4.99	81/16/3	65/31/04	.85	.13
France	5.00	62/22/16	36/36/28	.48	.83
Netherlands	5.01	61/24/15	59/31/10	1.02	.37
Austria	5.03	60/23/17	50/30/20	.64	.14
Poland	5.09	94/6/1	77/22/01	.41	.02
Germany	5.18	68/23/9	54/35/11	.61	.38
Belgium	5.28	65/22/13	56/30/14	1.11	.60
Portugal	5.29	73/22/5	58/33/09	.36	.74
United Kingdom	5.33	70/19/10	58/30/12	1.02	.14
Slovenia	5.35	65/30/5	47/46/07	.44	.33
Czech Republic	5.77	78/16/8	70/20/10	.69	.24
Israel	5.87	60/23/17	49/32/19	7.80	.59
Hungary	6.17	76/18/6	70/22/07	.71	.00
Greece	6.70	89/9/3	72/23/05	.08	.04
Total	5.15	69/21/10	54/33/13	1.40	.38

Sources: 2002–2003 European Social Survey and Comparative Manifestos Project.

Note: Sorted by antiforeigner sentiment. Cell entries for antiforeigner sentiment are means from a 15-item scale that runs from zero (no antiforeigner sentiment) to ten (maximum antiforeigner sentiment). See Appendix A for individual-level descriptives on antiforeigner sentiment and contact. The country-level descriptives on antiforeigner sentiment and contact are based on the 36,627 respondents included in the analyses reported in Tables 2 and 3.

The ESS is a biannual, cross-national survey and generally considered a high-quality source of comparable cross-national survey data because of high standards guiding the survey design and data collection process (Norris 2004). We have used the first round collected in 2002–2003 because it contains data on two forms of intergroup contact and antiforeigner sentiment as well as a number of relevant individual-level controls. The respondents were sampled from the residing population (aged 15 and older) in each country, with an average response rate of 61 percent. Further details concerning the sampling procedure and fieldwork can be found in ESS (2011).

The CMP (Volkens et al. 2011) covers the content and focus of political party manifestos and election programs from a wide range of countries (including all 22 ESS countries). The CMP offers longitudinal manifesto data collected each election year from all of the political parties winning legislative seats, and the CMP reports the share of written documents these parties have devoted to each of 56 predefined categories covering seven major policy areas. Details on coding procedure, categories, and the manifesto selection procedure can be found in Werner, Lacewell, and Volkens (2011).

Measures

To measure antiforeigner sentiment, we have chosen 15 items that address reactions toward immigrants and immigration. More specifically they address ingroup members’ negative/positive reactions toward: (1) immigrants of a different race/ethnic group, (2) immigrants from poorer countries in Europe, (3) immigrants from poorer countries outside Europe, (4) long-term unemployment among immigrants, (5) immigrants’ general rights, (6) immigrant criminals, (7) immigrants’ influence on national employment rates, (8) immigrants’ influence on welfare programs, (9) immigrants’ influence on the national economy, (10) immigrants’ influence on the national culture, (11) immigrants’ influence on the country as a whole, (12) immigrants’ influence on crime rates, (13) immigrants as managers, (14) immigrants as close relatives, and (15) cease to immigration. (For complete question wording of all items including descriptive information, see Appendix A).¹ The 15 items tap a variety of negative reactions toward immigrants—without referring to any specific ethnic group (see also Pettigrew 1997). The latter feature ensures that the scale taps attitudes toward the entire outgroup and ultimately the individual-to-group generalization effect. An unrotated principal-component factor analysis showed that the 15 items loaded strongly on a single (first) factor, with loadings between .45 and .77, producing an eigenvalue of 6.32. Two other factors emerged with eigenvalues of 1.34 and 1.21, thus meeting the default Kaiser criterion. Still, the first factor explains 42 percent of the total variance, whereas the other two explain about 8 and 9 percent. In effect, we decided to proceed with all 15 items. To minimize missing observations (i.e., to maintain representativeness), we allowed up to five “don’t know” responses and averaged the remaining genuine responses from each respondent. Removing respondents with more than five “don’t know” answers resulted in a sample loss of less than 3 percent.² The final scale ranges from zero to ten, with higher values indicating greater antiforeigner sentiment (M = 5.16; SD = 1.69; Cronbach’s alpha = .90 for the pooled sample). Within-country reliability analysis also indicated high reliability, with alpha ranging from .85 (Poland) to .92 (France). Table 1 displays the mean level of antiforeigner sentiment in the 22 countries, and it shows considerable cross-national variation. Additionally, this cross-national variation in antiforeigner sentiment is statistically significant (see the following), which indicates the need for multilevel modeling.

Our independent variable—intergroup contact—was identified by two self-reported measures: (1) contact in the workplace and (2) interethnic friendship. The specific wording of the two contact measures addresses the number of immigrant colleagues or friends, with none, few, and several as response categories. Table 1 shows the distribution of intergroup contacts across the 22 countries, whereas Appendix A reports the individual-level distributions.

The two contact measures have different strengths and weaknesses. Workplace contacts are less subject to self-selection mechanisms than friendship contacts because the setting is more involuntary than the voluntary nature of friendships (Mutz 2002). 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 considered an essential condition for the contact effect to occur (Hamberger and Hewstone 1997; Pettigrew and Tropp 2011).

Equally important, intergroup contact was measured by both workplace contact and personal friendship to cover interaction between majority and outgroup members in different social domains. In fact, the correlation between workplace contact and interethnic friendship is moderate (tau-b = .37), indicating that the two measures are far from identical. Thus, applying two distinctively different measures of contact allows for an assessment of the social coverage of political party messages. Ideally, political party cues should penetrate differing types of intergroup contact. Moreover, if the two measures generate similar results, the influence of unique characteristics (i.e., self-selection or measurement error) is more likely to be limited.

To measure our key interacting variable—political party focus on immigration—we used CMP data. In order to identify the mutual response logic of political party competition, we used the summed share of party manifesto sentences devoted to positive and negative aspects of multiculturalism (for similar approaches, see Alonso and de Fonseca 2011; Helbling, Reeskens, and Stolle forthcoming). In the manifesto data codebook, positive aspects of multiculturalism are defined as statements addressing “cultural diversity, communalism, cultural plurality and pillarization; preservation of autonomy of religious, linguistic heritages within the country including special educational provisions, while negative aspects oppose cultural diversity and support cultural integration” (Volkens et al. 2011:13–14).³ Thus, our measure identifies the proportion of statements addressing cultural, linguistic, and religious diversity issues in political party manifestos. In some countries, statements about multiculturalism may involve issues not intimately related to immigrants, but they will most probably relate to differences between majority members and ethnic minority members. Hence, strictly speaking, this variable measures the focus of political parties on immigration and cultural differences/identities. Such a relatively broad measure is required, since recent analyses suggest that symbolic boundaries against immigrants (us-them categorizations) differ considerably across European countries (Bail 2008). Thus, a broad and composite measure of political party focus on immigration is the most adequate.

To be consistent with Hypothesis 1, we must identify the proclivity of political parties to focus on immigration. Accordingly, we summed the shares of party manifesto sentences for all parties that won parliamentary seats; this total was subsequently divided by the number of parties, yielding a mean score of political party focus for each country in a specific election year. Furthermore, to identify the long-term effect of political party focus, we summed the national averages across all elections during the period from 1990 to 2002 and divided by the number of elections during that period. This relatively long time horizon minimizes the influence of sudden short-term fluctuations in the political agenda, while also acknowledging that it takes time for political parties to penetrate the level of mass opinion. Moreover, focusing on this particular time period, which predates the ESS data collection, helps reduce the simultaneity bias that would occur if the aggregate level of antiforeigner sentiment among the electorate influences the focus of political parties.⁴ The final measure is thus an average across all parties for the whole period and for each country. Formally, the total political party focus measure was calculated as follows:

P o l i t i c a l p a r t y f o c u s o n i m m i g r a t i o n_{j} = \frac{1}{N E_{j}} \sum_{e} \frac{1}{N P_{e j}} \sum_{p} (M P_{p e j} + M N_{p e j}),

where (j) denotes country, (e) election, and (p) political party; NE is the number of elections, NP is the number of political parties, MP is the share of the manifesto used on positive statements about multiculturalism, whereas MN is negative statements about multiculturalism.

This formula yields a measure that varies between .04 and 7.80 (M = 1.40; SD = 1.72), indicating that an average political party in an average election year in an average country in our sample allocates slightly less than 1.5 percent of its manifesto to addressing the positive and negative aspects of immigration and cultural matters. Table 1 shows the level of total political party focus on immigration in the 22 countries.

To measure negative political party focus (to test Hypothesis 2) we also used equation 1 while ignoring MP_pej. This yields a measure of the average amount of political manifestoes addressing negative aspects of immigration and cultural issues across political parties and national elections (M = .38; SD = .48). Table 1, the last column, reports the level of negative political party focus in each country.⁵

Considering that politics is commonly believed to be a pragmatic business, the use of manifesto data to measure the true agenda of political parties may appear problematic. Party manifesto data do not accurately measure political parties’ exact policy positions, but they are likely to indicate the primary goals and major concerns of political parties in the long run. In fact, Helbling and Tresch (2011) have shown that manifesto data correlate strongly with expert judgments on policy issue priorities. By averaging across all parties and manifestos from several years, our measure identifies an enduring focus on immigrant issues among political parties; admittedly, we assume that the written text of party manifestos will influence what politicians openly declare in public about their preferred goals and key worries.

Finally, the association between mean level of antiforeigner sentiment and political party focus deserves initial mentioning because of potential problems of endogeneity. The activities of political parties could be strongly driven by mass-level sentiments. However, the mean level of antiforeigner sentiment and the average level of political party focus inform this important issue. Table 1 shows that some countries combine scores above the grand mean of antiforeigner sentiment with scores below the grand mean of political party focus (e.g., Belgium, Czech Republic, Greece, Hungary, and Portugal). In others, the opposite combination occurs. This implies that the association between antiforeigner sentiment and political party focus is weak, which suggests that political party activities in the area of immigration are not deterministically governed by the electorate.⁶ We return to this relationship in the empirical analysis.

Controls and Specifications

Gender, age, education, relative household income, and political ideology (left-right self-placement) were included as individual-level control variables in order to reduce self-selection bias that would occur if contact is affected by some of the same factors affecting antiforeigner sentiment. Appendix A shows coding procedures, notes, and descriptive statistics. We also included a number of country-level controls: GDP per capita, share of immigrants, and unemployment rate. While controlling for GDP per capita is standard procedure in cross-national analyses, the additional controls were included as they may relate to the type and likelihood of workplace contact, political party focus on immigration issues, and antiforeigner sentiment. Appendix A also shows the descriptive statistics and the sources of these variables.

The country-level and individual-level controls and our main independent interaction variable were included in a random effects (RE) specification using the maximum likelihood estimator. The RE specification is the preferred approach in multilevel modeling, primarily because of its ability to cope with potential autocorrelation originating from the hierarchical structure of the data—individuals nested in countries, in our case. The RE specification also allows for the testing of variables varying only at the higher level (e.g., political party focus). To ensure unbiased estimates, however, any unobserved country-level variable must be uncorrelated with the included individual-level variables in the RE specification. This assumption is critical in multilevel modeling, as failure to meet it generates biased results. Accordingly, we test this assumption by comparing the results generated by the random effects estimation with results from a country fixed effects (FE) estimator. Although the FE estimator cannot provide a test of country-level variables (as they are not identified separately from the country fixed effects), it does control all constant effects of country-level variables, which in turn allows the validation of the RE results (Wooldridge 2001). Tables B3 and B4 in the online Appendix B show the results generated by the FE estimator, and they are almost identical to the RE results reported in the following. The validity of the RE specification is indicated by the actual similarity of the coefficients and most certainly vindicated by the Hausman tests that fail to reject the null hypothesis claiming no specification bias. In sum, our results are unlikely to be biased because of unobserved country-level variables. Accordingly, we henceforth focus on the RE results.

As our theory is confined to natives’ contact with immigrants, all non-native residents were excluded from the analysis, resulting in a total sample size of 36,627 individuals from 22 countries.⁷ The average per country effective sample size is 1,665, with a maximum of 2,598 and a minimum of 970.

Analysis and Results

In Table 2, Model 0 shows the results of the so-called null model with no covariates. This model tests for the occurrence of autocorrelation and therefore whether multilevel modeling is required. The results indicate substantial across-country variance in terms of antiforeigner sentiment, and this is vindicated by the likelihood ratio test ( ${\bar{χ}}^{2}$ = 4697, p < .01). Hence, multilevel modeling is needed to address autocorrelation.

Table 2.

The Average Effect of Contact and Political Party Focus on Antiforeigner Sentiment

Model	0	1a	1b	1c	1d
Contact measure	—	Colleagues	Friends	Colleagues	Friends
Political party focus measure	—	Total	Total	Negative	Negative
Individual-level variables
Immigrant colleagues/friends
None		Reference
Few		−.23** (−11.19)	−.58** (−33.32)	−.23** (−11.19)	−.58** (−33.32)
Several		−.50** (−18.19)	−1.03** (−41.43)	−.50** (−18.19)	−1.03** (−41.43)
Country-level variables
Political party focus		.01 (.12)	−.01 (−.18)	.03 (.24)	−.00 (−.01)
Constant	5.15** (41.95)	5.63** (11.09)	5.84** (12.12)	5.63** (11.09)	5.84** (12.09)
Residual variance
Between countries	.33**	.20**	.18**	.20**	.18**
Within countries	2.49**	2.15**	2.05**	2.15**	2.05**
N (individuals)	36,627
N (countries)	22

Note: Unstandardized coefficients, with z statistics in parentheses. All models (except Model 0) include controls for gender, age, income, education, ideology, country-level wealth, unemployment, and percentage foreign born. Online Appendix B Table B1 reports coefficients for all variables.

p < .05. **p < .01 (two-tailed tests).

Before testing Hypothesis 1 and Hypothesis 2, we report models without interaction terms. These models estimate the average contact effect across countries with varying degrees of party focus—and they also estimate the average effect of political party focus. For the sake of clarity, Table 2 only reports the coefficients for the central variables, while online Appendix B Table B1 also reports the control variable coefficients. Model 1a of Table 2 uses immigrant colleagues as a measure of interethnic contact, whereas total political party focus serves as an indicator of political party cues. Model 1a shows that workplace contact is negatively related to antiforeigner sentiment. The category “no immigrant colleagues” serves as the baseline category. Thus, respondents with few immigrant colleagues have significantly less antiforeigner sentiment than those who have no contact with immigrants in their workplace, and respondents with several immigrant colleagues have even lower antiforeigner sentiment.

Likewise, Model 1b of Table 2 shows that friendship is also inversely related to antiforeigner sentiment; however, the effect of friendship is clearly larger than the effect of workplace contact. This is consistent with previous research emphasizing that the influence of friendship is usually stronger as it meets most facilitating conditions (Pettigrew and Tropp 2011:115–19). Additionally, the relatively larger influence of friendship holds true for both measures of political party focus, distinguishing between the total number of political party cues and purely negative ones (shown in Model 1d). Even more important for our purposes, the estimates of Models 1a–d show that the coefficient for total political party focus is insignificant. Apparently, the public focus of political parties on immigration issues does not correlate with the level of antiforeigner sentiment for an average citizen. This supports the observation made earlier that political party activities do not necessarily reflect all popular demands.

Model 2a and 2b of Table 3 test the political enhancement effect (Hypothesis 1) by including the interaction terms between intergroup contact and total political party focus on immigration. Consistent with our expectations, the interaction terms are statistically significant and inversely related to antiforeigner sentiment. The negative signs of both contact measures indicate that the antiforeigner sentiment differential between contacted and noncontacted ingroup members is largest in political contexts characterized by intense political party focus on immigration as predicted by Hypothesis 1.

Table 3.

The Moderating Effect of Political Party Focus on the Contact–Antiforeigner Sentiment Relationship

Model	2a	2b	2c	2d
Contact measure	Colleagues	Friends	Colleagues	Friends
Political party focus measure	Total	Total	Negative	Negative
Individual-level variables
Immigrant colleagues/friends
None	Reference
Few	−.18** (−6.74)	−.51** (−22.81)	−.20** (−8.92)	−.54** (−28.05)
Several	−.40** (−11.36)	−.98** (−30.93)	−.43** (−13.82)	−1.00** (−35.08)
Country-level variables
Political party focus	.02 (.34)	.01 (.14)	.06 (.44)	.04 (.31)
Interaction (focus × colleagues/friends)
None	Reference
Few	−.03** (−3.02)	−.05** (−4.77)	−.05* (−2.24)	−.09** (−4.20)
Several	−.07** (−4.66)	−.04** (−2.70)	−.13** (−3.91)	−.07* (−2.36)
Constant	5.61** (10.97)	5.81** (11.97)	5.61** (11.03)	5.82** (12.01)
Residual variance
Between countries	.20**	.18**	.20**	.18**
Within countries	2.14**	2.05**	2.15**	2.05**
N (individuals)	36,627
N (countries)	22

Note: Unstandardized coefficients, with z statistics in parentheses. All models include controls for gender, age, income, education, ideology, country-level wealth, unemployment, and percentage foreign born. Online Appendix B Table B2 reports coefficients for all variables.

p < .05. **p < .01 (two-tailed tests).

Focusing specifically on workplace contact, Model 2a of Table 3 reveals that a 1 percentage point increase in total political party focus enlarges the difference in antiforeigner sentiment between a person with no immigrant colleagues and a person with several immigrant colleagues by .07 on the antiforeigner sentiment scale. The effect is remarkably strong. Total political party focus ranges from .04 to 7.80 (shown in Table 1), implying that the estimated difference in antiforeigner sentiment between a person with no workplace contact and a person with several immigrant colleagues in the least “politicized” country is .40, whereas it is .95 in the most “politicized.” Thus, the contact effect more than doubles across the range of observed intensity of total political party focus. Confining the comparison to the interquartile range of focus (.48 to 1.63) yields differences of .43 compared to .51.

To fully elucidate our main result concerning workplace contact, Figure 1 plots the predicted marginal differences in antiforeigner sentiment (including 95 percent confidence intervals) between persons with no contact and persons with few and several immigrant colleagues, respectively, across the observed range of total political party focus. Figure 1 clearly shows that the total political party focus on immigration substantively moderates the relationship between workplace contact and antiforeigner sentiment. Moreover, since both confidence intervals do not include zero on the y-axis across the entire observed range of total political party focus, contacted ingroup members have significantly less antiforeigner sentiment no matter the level of total political party focus on immigration. This finding is consistent with our expectation that contact between majority members and immigrants is most probably never purely interpersonal; even in a non-politicized context, contact effects generalize to the entire outgroup. Even so, the enhancement effect triggered by political parties is unmistakably clear.

Figure 1.

Differences in Antiforeigner Sentiment (AFS) at Varying Degrees of Total Political Party Focus

Returning to Table 3 (Model 2b), it becomes clear that the moderated effect of friendship is essentially similar to the moderated effect of workplace contact (shown in Model 2a); however, it seems that political party focus on immigration issues reduces antiforeigner sentiment among ingroup members having few and several friends to the same extent (online Appendix B Figure B1 graphs this particular interaction effect). This means that the specified interaction terms between contact, antiforeigner sentiment, and political party focus are insensitive to the specific measure of intergroup contact. The results are robust, but these findings also suggest that political party cues about immigration influence group membership salience in different social domains involving both strong and less strong intimacy.

So far the analyses have supported the political enhancement effect (Hypothesis 1). But do the empirical data support the Hypothesis 2 rival, claiming that negative political party cues reduce the contact effect? Table 3 addresses this question in Model 2c and 2d in which the coefficients from Model 2a and 2b are re-estimated on the basis of the negative political party focus measure. According to Hypothesis 2, the interaction terms between contact and negative political party focus should be positive, implying that the contact effect should be strongest in countries with low levels of negative political party focus on immigrants. Model 2c and 2d of Table 3 do not, however, lend support to Hypothesis 2. Indeed, the interaction terms involving workplace contact and friendship are statistically significant—but, surprisingly, negative in sign. Put differently, negative political party focus enhances the ability of both workplace contact and friendship to reduce antiforeigner sentiment. These findings are certainly inconsistent with one of the key implications of the decategorization model claiming that categorization-based identities are harmful to the outcomes of intergroup contact.

As our hypotheses are competing, it follows that the rejection of Hypothesis 2 implies additional support for Hypothesis 1. Moreover, as the results of two different measures of political party focus are similar, we are confident that the results supporting Hypothesis 1 are robust. Taken together, these findings suggest that political party focus on immigrant issues enhances the positive effects of intergroup contact irrespective of the specific tone of political party rhetoric.

Probing the Psychological Mechanism of Group Membership Salience

So far, our results indicate that the focus of political parties on immigration heightens group membership salience, which in turn enhances the contact effect. However, the empirical tests have not applied any direct measure of group membership salience at the individual level. Hence, additional empirical evidence to support our claim that political parties actually do stimulate group membership salience is required.

Unfortunately, the European Social Survey (ESS) does not offer ideal measures of group membership salience, but previous studies have argued that issue salience variability at the mass level can be measured by “don’t know” responses (Shapiro and Mahajan 1986). According to this perspective, a “don’t know” response means no opinion on an issue (Krosnick 1991). Building on this assumption, Page and Shapiro (1983) suggest that “don’t know” responses reflect lack of personal interest and attention. Yet, it is likely that political elites and their activities can cause ordinary citizens to perceive politics as central to their lives (Nie and Andersen 1974). For instance, Shapiro and Mahajan (1986) have shown that from the 1960s and onward, women became less inclined to select the “don’t know” option.

Following this elite-centered approach, we created a proxy for group salience by counting the number of “don’t know” responses to 29 immigrant items addressing knowledge about and attitudes toward immigration (and integration) policies. This proxy variable enables us to test whether the total immigration focus of political parties in fact increases group membership salience. If the positive and negative messages of political parties are capable of increasing group membership salience, the number of “don’t know” responses should be smallest among respondents in the countries with intense political focus on immigration. This will indicate that enduring exposure to political debate about immigration and immigrants stimulates ingroup members to perceive contacted immigrants as prototypical outgroup members.

We have tested this claim in a multilevel model with the number of “don’t know” responses as the dependent variable (measured at the individual level) and total political party focus (measured at the country level as in Table 2 and 3) as the independent variable. We included the same country-level and individual-level controls as in Table 2 and 3 as well as both measures of intergroup contact. In addition to these controls, we included a supposedly very effective control variable, referring to the number of “don’t know” responses to all questions in the survey that do not relate to immigrant issues. This variable, measured at the individual level, is likely to identify any general propensity to select the “don’t know” response including the differential need for cognition—a personality trait considered important for understanding respondent motivation in survey research (Krosnick 1991).⁸

Table 4 shows this test, although we only report the total political party focus coefficient and the coefficient related to the number of “don’t know” responses to other questions (Table B7 in online Appendix B reports all coefficients). Table 4 shows that the effect of total political party focus on the number of “don’t know” responses is negative and statistically significant (p < .01), suggesting that political parties are certainly capable of making immigration issues salient among the citizenry. This result strongly corroborates the political enhancement effect hypothesis (Hypothesis 1).

Table 4.

The Effect of Total Political Party Focus on “Don’t Know” Responses

Other “don’t know” responses	.07** (96.12)
Total political party focus	−.03** (−2.96)
Constant	−.88** (−9.23)
N (individuals)	37,525
N (countries)	22

Note: Unstandardized coefficients (obtained with the random effects negative binomial estimator) with z statistics in parentheses. The model includes the same controls as the models in Table 2 and 3 as well as both measures of intergroup contact; but the coefficients for these controls are omitted from the table. Online Appendix B Table B7 reports all coefficients.

p < .05. **p < .01 (two-tailed tests).

Conclusions and Implications

In their impressive meta-analytic test of intergroup contact theory based on 515 studies and over a quarter of a million individuals, Pettigrew and Tropp (2006) reported that greater intergroup contact is generally associated with less resentment toward outgroup members. Even more recently, Pettigrew and Tropp (2011:209) concluded that “our meta-analysis uncovered . . . no multi-level studies.” Multilevel analyses of intergroup contact do exist, but they are certainly rare (McLaren 2003; Quillian 1995; Schneider 2008). None of these multilevel-based studies, however, have examined the political context of intergroup contact.

This state of affairs makes the conclusion of this investigation all the more intriguing. Indeed, the major finding of this investigation is that political parties are able to strengthen the relationship between intergroup contact and antiforeigner sentiment: the more political parties make immigrants and collective identities a public concern, the greater is the effectiveness of intergroup contact in reducing antiforeigner sentiment. This result is based on a multilevel analysis of 22 different countries—including several control variables that cover both individual-level and country-level characteristics. Moreover, political party messages can stimulate group membership salience in both intimate and less intimate types of contact. This suggests that most types of meaningful intergroup contact are potentially receptive to the stimuli of national party politics.

A central component of our model of politicized intergroup contact effects has been Hewstone and Brown’s (1986) original framework, which proposes that salient group membership differences are likely to enhance the ability of intergroup contact to reduce antiforeigner sentiment. Hewstone and Brown’s (1986) framework contrasts with the so-called decategorization model, according to which the contact effect will be strongest when collective identities are nonsalient (Brewer and Miller 1984).

We have not tested the full implications of the decategorization model. The results of our incomplete test, however, do lend firm support to Hewstone and Brown’s (1986) model rather than the decategorization model because purely negative political party cues also enhance the intergroup contact effect. The very fact that different measures of political party focus produce similar results suggests that the specific substance of political rhetoric is rather unimportant. What matters for the influence of intergroup contact is the extent to which political parties actually make immigration issues a national concern.

Yet, although Hewstone and Brown (1986) are correct in arguing that group membership salience is ultimately an individual-level phenomenon, their approach needs further refinement. In answering what stimulates the individual awareness of group memberships and group differences, political parties are obvious candidates. First, because of their privileged access to the media, political parties are very powerful agents in the process of defining what the public should care about. Second, identity politics is necessarily focused on us-them categorizations, even among the political parties striving to dismantle them.

In terms of theoretical advancement, we have provided a more comprehensive understanding of the meaning and functioning of group membership salience. Group membership salience is a mix of psychological mechanisms and macro-political stimuli that refer to party messages. More fundamentally, the present investigation supports Oakes and Haslam’s (2001) meta-theoretical proposition that categorization does not necessarily produce bigotry or hostility. In particular, they seem correct when arguing that the conditions under which categorization occurs are crucial. Among those who do not have contact experiences with outgroup members, highly salient categorizations are likely to sustain their negative assessments (also Paolini et al. 2010); as we have shown, however, when direct personal contact experiences intersect with salient and enduring categorizations, the ingroup participants have less antiforeigner sentiment toward outgroup participants. We interpret the enhancement effect as indicating the capacity of political parties to mobilize interest in sensitive matters related to group identities and distinctively different ways of life in modern multiethnic societies.

Naturally, this investigation also has its limitations and raises questions for further research. To fully understand the specific mental processes stimulated by political party focus, further research and better data are required. In particular, we have been unable to show the extent to which ingroup members become aware of outgroup memberships as a consequence of increased awareness of the significance attached to their own reference group. After all, social identity formation is a relational process possibly also involving self-critical evaluations among contacted ingroup members. Pettigrew (1997:174) has called this phenomenon “deprovincialization,” implying that contacted ingroup members become less inclined to believe in the unquestionable superiority of their own norms and habits. Evidently, some political parties directly stimulate reflections on the limitations of one’s own culture and habits, whereas others emphasize the inferiority of outgroup cultural attributes. In both cases, however, the political parties put daily interactions in a larger multicultural perspective. Unfortunately, we have been unable to measure precisely how ingroup members reason on the basis of different types of political messages about themselves and the outgroup members with whom they interact. Yet, experiments examining how contacted ingroup members (compared to noncontacted) react toward different political messages are obviously needed.

Our investigation has transcended the traditional division of labor between a subdiscipline of social psychology and political science. This division is unfortunate in view of our results, which emphasize the need for treating insights from these two disciplines as complementary. Social psychology offers the fundamental insight that what matters in any society is the extent to which majority and minority members interact relatively peacefully and thus learn to respect various intergroup differences (within reasonable limits). Political science offers a broader top-down perspective emphasizing the role of political elites—and it seems as though politicians can actually stimulate intergroup relations positively. In substantial terms, the present investigation thus provides part of a solution to the fundamental puzzle identified by Gordon W. Allport ([1954] 1979:276) about 60 years ago when emphasizing that occupational contact suffers from an inherent limitation, because “it may not affect the individual’s customary perceptions and habits.” Allport was right, but we want to add that this issue is much less puzzling if contact situations are theoretically linked to the political context from the very beginning. More generally, the support for our refinement of the categorization hypothesis suggests that the intergroup contact theory needs to pay attention to the dynamics of the political system. After all, intergroup contact always occurs in a political context.

Footnotes

Appendix

Appendix A.

Variable Details

Concept	Descriptive statistics and question wording	Mean/SD or distribution (percentage)	Source
Individual-level variables
Antiforeigner sentiment	Mean score on the following items (all rescaled from 0 to 10). Respondents who failed to answer, or responded “don’t know” on five or more items were treated as missing.	5.16/1.68	ESS
Item 1	“To what extent do you think [country] should allow people of a different race or ethnic group as most [country] people to come and live here?”	5.12/2.75	ESS: imdfetn
Item 2	[same intro and ending as previous] “from the poorer countries in Europe.”	4.83/2.73	ESS: eimpcnt
Item 3	[same intro and ending as previous] “from the poorer countries outside Europe.”	5.06/2.75	ESS: impcntr
Item 4	“If people who have come to live and work here are unemployed for a long period, they should be made to leave.”	5.49/2.93	ESS: imunplv
Item 5	“People who have come to live here should be given the same rights as everyone else.”	3.51/2.62	ESS: imsmrgt
Item 6	“If people who have come to live here commit a serious crime, they should be made to leave.”	7.89/2.54	ESS: imscrlv
Item 7	“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?”	5.59/2.22	ESS: imtcjob
Item 8	“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?”	5.90/2.21	ESS: imbleco
Item 9	“Would you say it is generally bad or good for [country]’s economy that people come to live here from other countries?”	5.14/2.35	ESS: imbgeco
Item 10	“Would you say that [country]’s cultural life is generally undermined or enriched by people coming to live here from other countries?”	4.32/2.45	ESS: imueclt
Item 11	“Is [country] made a worse or better place to live by people coming to live here from other countries?”	5.31/2.17	ESS: imwbcnt
Item 12	“Are [country]’s crime problems made worse or better by people coming to live here from other countries?”	6.94/2.03	ESS: imwbcrm
Item 13	“And now thinking of people who have come to live in [country] from another country who are of a different race or ethnic group from most [country] people. How much would you mind or not mind if someone like this was appointed as your boss?”	3.28/3.28	ESS: imdetbs
Item 14	[same intro as previous item] . . . “married a close relative of yours?”	3.60/3.46	ESS: imdetmr
Item 15	“If a country wants to reduce tensions it should stop immigration.”	5.37/2.78	ESS: stimrdt
Immigrant colleagues	Categorical variable with three categories: (1) none, (2) few, and (3) several, based on the question: “Do you have any colleagues at work who have come to live in [country] from another country?”^a	(1) 69, (2) 21, (3) 10	ESS: imgclg
Immigrant friends	Categorical variable with three categories: (1) none, (2) few, and (3) several, based on the question: “Do you have any friends who have come to live in [country] from another country?”	(1) 55, (2) 33, (3) 13	ESS: imgfrnd
Gender (male)	Dummy variable; 1 = male	48	ESS: gndr
Age	Calculated as year of interview – year of birth^b	46.06/18.10	ESS: inwyr, yrbrn
Education	Categorical variable with six categories: (1) less than lower secondary education, (2) lower secondary education completed, (3) upper secondary education completed, (4) postsecondary non-tertiary education completed, (5) tertiary education completed, and (6) not revealed/other.^c	(1) 15, (2) 24, (3) 38, (4) 2, (5) 21, (6) 1	ESS: edulvla
Relative household income	Categorical variable with five categories: (1) below the 10th percentile, (2) between the 10th percentile and the median, (3) between the median and the 90th percentile, (4) above the 90th percentile, and (5) not revealed.^c Relative income refers to the respondent’s income compared to that of his or her countrymen.^d	(1) 4, (2) 28, (3) 34, (4) 14, (5) 19	ESS: hinctnt, hinctnfr, hinctnhu, hinctnie^e
Political ideology	“In politics people sometimes talk of ‘left’ and ‘right’. Where would you place yourself on a scale from 0 to 10, where 0 means the left and 10 means the right?” “Don’t know” has been coded as 5 (n = 3,538). Refusals have been excluded.^c	5.06/2.08	ESS: lrscale
Country-level variables
Total political party focus	See Equation 1 in text. The measure for United Kingdom is based on manifesto data from Great Britain.	1.40/1.72	Volkens et al. (2011): per607, per608
Negative political party focus	Based on Equation 1, but including only negative statements about multiculturalism ( $M N_{p e j}$ ). The measure for United Kingdom is based on manifesto data from Great Britain.	.38/.48	Volkens et al. (2011): per608
Foreign-born	Share of population born abroad in 2000 (percentage).	10.39/8.82	World Bank (2011)
Unemployment rate	Unemployed as share of total labor force in 2002 (percentage).	7.02/3.92	World Bank (2011)
GDP/capita	Gross domestic product per 1,000 capita converted to constant 2005 international dollars utilizing purchasing power parity rates in 2002.	29,147/10,728	World Bank (2011)

Notes: Descriptive information about the individual-level variables is based on the 36,626 respondents included in the analyses reported in Table 2 and 3. ESS = European Social Survey.

In most countries, the respondents could also respond “not currently working,” which we have merged with “none” because both answers imply that the respondent has no immigrant colleagues. Furthermore, in some countries respondents were not offered the possibility of answering “not currently working.” The results reported concerning the interaction between cross-group contact and political party focus remain intact when separating “none” and “not currently working” and also when omitting respondents who were not working when surveyed.

Year of interview is missing for a few respondents. In these cases, we assumed year of interview to be 2002.

The results regarding the interaction terms in Table 3 are insensitive to our handling of missing values as the results are substantially similar if we exclude nonresponders on the variables of education, occupation, income, or political ideology.

The household income categories of France, Hungary, and Ireland differ from the remaining countries. To cope with this issue, we have produced a relative income measure instead, where each respondent’s income was calculated as their relative income compared to their countrymen’s. Combined with either a measure of country-level wealth or country fixed effect, we are controlling for both relative and absolute income.

hinctnfr, hinctnhu, and hinctnie are from the separate country files covering France, Hungary, and Ireland.

Acknowledgements

The authors would like to thank Sidsel Vive Jensen, Paul Sniderman, Eric Uslaner, Cara Wong, as well as SPQ editor Karen Hegtvedt, two anonymous reviewers, and several colleagues at the Department of Political Science and Government, Aarhus University, for helpful comments and suggestions.

1

The survey contains additional questions addressing immigrant issues, but they were excluded on grounds of substantive overlap because they relate to policy attitudes or because they tap factual knowledge about immigrants rather than dispositions. Still, we used most of these items in our “don’t know” response analysis, which is reported in .

2

Removing respondents with missing information on the remaining variables further reduced the sample by 2.9 percent.

3

Pillarization refers to the politico-denominational segregation of a society. These societies are vertically divided into several segments or “pillars” according to different religions or political ideologies.

4

Plausible empirical and theoretical arguments exist for using the relatively long time horizon when measuring political party focus. Admittedly, the moderating effect of political party focus on group membership salience could be even stronger if the time horizon is shortened. After all, ordinary citizens are rather inattentive to politics. Accordingly, Models 2a–d were re-estimated on the basis of manifesto data only covering the last election prior to 2002. The results (shown in online Appendix B Table B5) are substantially similar to those obtained with the longer time horizon. The only difference is that the interaction terms in the models addressing negative political party focus are insignificant.

5

As an alternative measure of negative political party focus, we have used the ratio of negative messages to positive messages in order to identify a bias toward negative statements about the dangers of immigrants and immigration (a high ratio indicates strong bias toward negative statements). These results are reported in online Appendix B Table B6 and yields slightly smaller and less accurate estimates than the purely negative measure.

6

Dahlström and Esaiasson’s (2011) study supports this point by showing the massive gap between mass-level opinions and the preferences of parliament members (MP) in the case of Sweden. They report: “During the period 1990–2006, the proportion of respondents supporting restrictive immigration policy has never fallen below 43 percent . . . MP support is usually below 10 percent” (:349).

7

We included respondents who reported being born in their country of residence and respondents whose parents were both born in the same country of residence as the respondent.

8

The dependent variable is the number of “don’t know” responses to 29 European Social Survey (ESS) items listed in Table B8, in online Appendix B. The control variable other “don’t know” responses comprises all items in the survey except for those related to immigration. Because the dependent variable is a count variable showing considerable dispersion (M = 1.24; SD = 3.16), a random effects negative binomial estimator was applied. All “non-natives” were excluded from the analysis.

Bios

Kim Mannemar Sønderskov is an associate professor in the Department of Political Science and Government at Aarhus University in Denmark. His main research interests are social trust and interethnic attitudes. His current research focuses on the effects of neighborhoods and other social and political contexts on political attitudes and behavior. His most recent publications have appeared in European Sociological Review and Journal of Ethnic and Migration Studies.

Jens Peter Frølund Thomsen is an associate professor of comparative politics in the Department of Political Science and Government at Aarhus University in Denmark. His research concerns relations of power, state theory, political socialization, political parties, intergroup contact among both majority and minority members, and majority as well as minority responses to values of conflict. His current research project is the psychology of class distinctions.

References

Allport

Gordon W.

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

Alonso

Sonia

da Fonseca

Saro Claro

. 2011. “Immigration, Left and Right.” Party Politics 18:865–84.

Bail

Christopher A.

2008. “The Configuration of Symbolic Boundaries against Immigrants in Europe.” American Sociological Review 73:37–59.

Bale

Tim

. 2008. “Turning Round the Telescope. Centre-Right Parties and Immigration and Integration Policy in Europe.” Journal of European Public Policy 15:315–30.

Brader

Ted

Valentino

Nicholas A.

Suhay

Elizabeth

. 2008. “What Triggers Public Opposition to Immigration? Anxiety, Group Cues, and Immigration Threat.” American Journal of Political Science 52:959–78.

Brewer

Marilynn B.

Miller

Normann

. 1984. “Beyond the Contact Hypothesis: Theoretical Perspectives on Desegregation.” Pp. 281–302 in Groups in Contact: The Psychology of Desegregation, edited by Miller

Brewer

M. B.

New York: Halstead Press.

Brown

Rupert

Eller

Anja

Leeds

Sarah

Stace

Kim

. 2007. “Intergroup Contact and Intergroup Attitudes: A Longitudinal Study.” European Journal of Social Psychology 37:692–703.

Brown

Rupert

Hewstone

Miles

. 2005. “An Integrative Theory of Intergroup Contact.” Pp. 255–342 in Advances in Experimental Social Psychology. Vol. 37, edited by Zanna

M. P.

San Diego, CA: Elsevier.

Brown

Rupert

Vivian

James

Hewstone

Miles

. 1999. “Changing Attitudes through Intergroup Contact: The Effects of Group Membership Salience.” European Journal of Social Psychology 29:741–64.

10.

Carmines

Edward G.

Stimson

James A.

1986. “On the Structure and Sequence of Issue Evolution.” American Political Science Review 80:901–20.

11.

Carpini

Michael D.

Keeter

Scott

. 1996. What Americans Know About Politics and Why It Matters. New Haven, CT: Yale University Press.

12.

Carter

Elisabeth L.

2005. The Extreme Right in Western Europe. Manchester: Manchester University Press.

13.

Dahlström

Carl

Esaiasson

Peter

. 2011. The Immigration Issue and Anti-Immigrant Party Success in Sweden 1970–2006: A Deviant Case Analysis.” Party Politics 19:343–64.

14.

Davies

Kristin

Tropp

Linda R.

Aron

Arthur

Pettigrew

Thomas F.

Wright

Stephen C.

2011. “Cross-Group Friendships and Intergroup Attitudes: A Meta-Analytic Review.” Personality and Social Psychology Review 15:332–51.

15.

Downs

Anthony

. 1957. An Economic Theory of Democracy. Boston, MA: Addison Wesley.

16.

Ensari

Nurcan

Miller

Norman

. 2002. “The Out-Group Must Not Be So Bad After All: The Effects of Disclosure, Typicality, and Salience on Intergroup Bias.” Journal of Personality and Social Psychology 83:313–29.

17.

ESS. 2011. ESS1—2002 Documentation Report, Edition 6.2. The ESS Data Archive.

18.

Green-Pedersen

Christoffer

. 2007. “The Growing Importance of Issue Competition: The Changing Nature of Party Competition in Western Europe.” Political Studies 55:608–28.

19.

Green-Pedersen

Christoffer

Krogstrup

Jesper

. 2008. “Immigration as a Political Issue in Denmark and Sweden.” European Journal of Political Research 47:610–34.

20.

Hamberger

Jürgen

Hewstone

Miles

. 1997. “Inter-Ethnic Contact as a Predictor of Blatant and Subtle Prejudice: Tests of a Model in Four West European Nations.” British Journal of Social Psychology 36:173–90.

21.

Harwood

Jake

. 2010. “The Contact Space: A Novel Framework for Intergroup Contact Research.” Journal of Language and Social Psychology 29:147–77.

22.

Harwood

Jake

Raman

Priya

Hewstone

Miles

. 2006. “The Family and Communication Dynamics of Group Salience.” The Journal of Family Communication 6:181–200.

23.

Helbling

Marc

Reeskens

Tim

Stolle

Dietlind

. Forthcoming. “Bringing Political Parties Back In: Cultural Diversity, Social Cohesion, and Political Mobilization.” Political Studies.

24.

Helbling

Marc

Tresch

Anke

. 2011. “Measuring Party Positions and Issue Salience from Media Coverage: Discussing and Cross-Validating New Indicators.” Electoral Studies 30:174–83.

25.

Hewstone

Miles

. 2009. “Living Apart, Living Together? The Role of Intergroup Contact in Social Integration.” Proceedings of the British Academy 162:243–300.

26.

Hewstone

Miles

Brown

Rupert

. 1986. “Contact Is Not Enough: An Intergroup Perspective on the ‘Contact Hypothesis.’” Pp. 1–44 in Contact and Conflict in Intergroup Encounters, edited by Hewstone

Brown

Oxford: Basil Blackwell.

27.

Hillygus

D. Sunshine

Shields

Todd G.

2008. The Persuadable Voter: Wedge Issues in Presidential Campaigns. Princeton: Princeton University Press.

28.

Hopkins

Daniel J.

2011. “National Debates, Local Responses: The Origins of Local Concern about Immigration in Britain and the United States.” British Journal of Political Science 41:499–524.

29.

Hutchings

Vincent L.

2001. “Political Context, Issue Salience, and Selective Attentiveness: Constituent Knowledge of the Clarence Thomas Confirmation Vote.” Journal of Politics 63:846–68.

30.

Jackman

Mary R.

Crane

Marie

. 1986. “‘Some of my Best Friends are Black . . . ’ Interracial Friendship and Whites’ Racial Attitudes.” Public Opinion Quarterly 50:459–86.

31.

Kinder

Donald R.

Kam

Cindy D.

2009. US Against Them. Chicago: The University of Chicago Press.

32.

Krosnick

Jon A.

1991. “Response Strategies for Coping with the Cognitive Demands of Attitude Measures in Surveys.” Applied Cognitive Psychology 5:213–36.

33.

Lau

Richard R.

1982. “Negativity in Political Perception.” Political Behavior 4:353–77.

34.

Lee

Barrett A.

Farrell

Chad R.

Link

Bruce G.

2004. “Revisiting the Contact Hypothesis: The Case of Public Exposure to Homelessness.” American Sociological Review 69:40–63.

35.

Marcus

George E.

Neuman

W. Russell

Mackuen

Michael

. 2000. Affective Intelligence and Political Judgment. Chicago: The University of Chicago Press.

36.

McLaren

Lauren M.

2003. “Anti-Immigrant Prejudice in Europe: Contact, Threat Perception, and Preferences for the Exclusion of Migrants.” Social Forces 81:909–36.

37.

Miller

Norman

. 2002. “Personalization and the Promise of Contact Theory.” Journal of Social Issues 58:387–410.

38.

Mutz

Diana C.

2002. “Cross-Cutting Social Networks: Testing Democratic Theory in Practice.” American Political Science Review 96:111–26.

39.

Nie

Norman H.

Andersen

Kristi

. 1974. “Mass Belief Systems Revisited: Political Change and Attitude Structure.” Journal of Politics 36:540–91.

40.

Norris

Pippa

. 2004. “From the Civic Culture to the Afrobarometer.” Newsletter of the American Political Science Association Organized Section in Comparative Politics 15:6–11.

41.

Oakes

Penelope J.

Haslam

S. Alexander

. 2001. “Distortion v. Meaning: Categorization on the Trial for Inciting Intergroup Hatred.” Pp. 179–94 in Understanding Prejudice, Racism, and Social Conflict, edited by Augoustinos

Reynolds

K. J.

London: Sage Publications.

42.

Page

Benjamin I.

Shapiro

Robert Y.

1983. “Effects of Public Opinion on Policy.” American Political Science Review 77:175–90.

43.

Paolini

Stefania

Harwood

Jake

Rubin

Mark

. 2010. “Negative Intergroup Contact Makes Group Memberships Salient: Explaining Why Intergroup Conflict Endures.” Personality and Social Psychology Bulletin 38:1723–38.

44.

Pettigrew

Thomas F.

1997. “Generalized Intergroup Contact Effects on Prejudice.” Personality and Social Psychology Bulletin 23:173–85.

45.

Pettigrew

Thomas F.

1998. “Intergroup Contact Theory.” Annual Review of Psychology 49:65–85.

46.

Pettigrew

Thomas F.

2006. “The Advantages of Multilevel Approaches.” Journal of Social Issues 62:615–20.

47.

Pettigrew

Thomas F.

Tropp

Linda R.

2006. “A Meta-Analytic Test of Intergroup Contact Theory.” Journal of Personality and Social Psychology 90:751–83.

48.

Pettigrew

Thomas F.

Tropp

Linda R.

2008. “How Does Intergroup Contact Reduce Prejudice? Meta-Analytic Tests of Three Mediators.” European Journal of Social Psychology 28:922–34.

49.

Pettigrew

Thomas F.

Tropp

Linda R.

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

50.

Quillian

Lincoln

. 1995. “Prejudice as a Response to Perceived Group Threat: Population Composition and Anti-Immigrant and Racial Prejudice in Europe.” American Sociological Review 60:586–611.

51.

Rydgren

Jens

. 2008. “Immigration Sceptics, Xenophobes or Racists? Radical Right-Wing Voting in Six West European Countries.” European Journal of Political Research 47:737–65.

52.

Schneider

Silke

. 2008. “Anti-Immigrant Attitudes in Europe: Outgroup Size and Perceived Ethnic Threat.” European Sociological Review 24:53–67.

53.

Semyonov

Moshe

Raijman

Rebecca

Gorodzeisky

Anastasia

. 2006. “The Rise of Antiforeigner Sentiment in European Societies.” American Sociological Review 71:426–49.

54.

Shapiro

Robert Y.

Mahajan

Harpreet

. 1986. “Gender Differences in Policy Preferences: A Summary of Trends from the 1960s to the 1980s.” Public Opinion Quarterly 50:42–61.

55.

Sniderman

Paul M.

Brody

Richard A.

Tetlock

Philip E.

1991. Reasoning and Choice. Explorations in Political Psychology. Cambridge, UK: Cambridge University Press.

56.

Sniderman

Paul M.

Hagendoorn

Louk

. 2007. When Ways of Life Collide. Princeton: Princeton University Press.

57.

Sniderman

Paul M.

Hagendoorn

Louk

Prior

Markus

. 2004. “Predisposing Factors and Situational Triggers: Exclusionary Reactions to Immigrant Minorities.” American Political Science Review 98:35–49.

58.

Sniderman

Paul M.

Theriault

Sean M.

2004. “The Structure of Political Argument and the Logic of Issue Framing.” Pp. 133–65 in Studies in Public Opinion, edited by Saris

W. E.

Sniderman

P. M.

Princeton: Princeton University Press.

59.

Soroka

Stuart N.

2006. “Good News and Bad News: Asymmetric Responses to Economic Information.” Journal of Politics 68:372–85.

60.

Stone

Deborah

. 1989. “Causal Theories and the Formation of Policy Agendas.” Political Science Quarterly 4:281–300.

61.

Voci

Alberto

Hewstone

Miles

. 2003. “Intergroup Contact and Prejudice toward Immigrants in Italy: The Mediational Role of Anxiety and the Moderational Role of Group Salience.” Group Processes and Intergroup Relations 6:37–54.

62.

Volkens

Andrea

Lacewell

Onama

Lehmann

Pola

Regel

Sven

Schultze

Henrike

Werner

Annika

. 2011. The Manifesto Data Collection. Manifesto Project (MRG/CMP/MARPOR). Berlin: Wissenschaftszentrum Berlin für Sozialforschung.

63.

Werner

Annika

Lacewell

Onama

Volkens

Andrea

. 2011. Manifesto Coding Instructions. 4th ed. Retrieved October 10, 2011 (http://manifestoproject.wzb.eu/documentation/Handbookv4.pdf).

64.

Wooldridge

Jeffrey M.

2001. Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: The MIT Press.

65.

World Bank. 2011. World Development Indicators. Retrieved May 30, 2011 (http://databank.worldbank.org/).

66.

Zaller

John R.

1992. The Nature and Origins of Mass Opinion. Cambridge, UK: Cambridge University Press.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.11 MB