Do Pure Liberals Both Endorse Immigration and Reject Muslims? Examining the Inconsistency Hypothesis

Abstract

Some people have consistently positive attitudes toward ethnic minority members in every respect. Sniderman and Hagendoorn (2007) characterize these individuals as “pure liberals.” Because of limited empirical evidence on the characteristics of this category of majority members, the consistency thesis, claiming that pure liberals have no negative reactions toward Muslims, remains contentious. Accordingly, we introduce a new idea – the inconsistency hypothesis – claiming that pure liberals’ commitment to greater ethnic diversity may not convert into unprejudiced reactions toward Muslim minority members. To test our inconsistency hypothesis, we fielded a survey in 2018 that also included a so-called list experiment among 716 Danish university students. The list experiment was designed to measure prejudice toward Muslims. Our analyses of the survey’s observational data revealed that pure liberals supported the general idea of greater diversity and remained positive toward ethnic minorities. However, and unlike previous research, the list experiment showed that around 34 percent of pure liberals in the survey were prejudiced toward Muslims. When we combine the observational and experimental evidence, it appears that many pure liberals had inconsistent attitudes: around a third of pure liberals combined support for the general idea of a multiethnic society with prejudice toward Muslims. Our inconsistency hypothesis contributes to literatures on liberal orientations and prejudice toward Muslims: Many liberals are not liberals, and prejudice is surprisingly powerful. The conclusion emphasizes the role of self-interest and distinct values as drivers of prejudice among the strongest advocates of a multiethnic society.

Keywords

prejudice muslims pure liberals list experiment students “nimbyism”

Introduction

The inflow of Muslims from the Middle East to European nations is provoking varying popular reactions among majority members (Gusciute, Mühlau and Layte 2021). In Western democracies, some segments of the majority group oppose Muslim immigration, while others support it (Bohman 2018). But do the supporters of Muslim immigration accept Muslims without hesitation? It is important to indicate the scope of prejudice toward Muslim ethnic minority members (henceforth: Muslims) in Europe for scientific purposes, but prejudice also poses acute challenges for policymakers, as this “predisposition” among majority members is indicative of how difficult it is for Muslims to achieve a respected position in their new society (Zick, Küpper and Höverman 2011). Simply put, prejudice works against inclusion.

Despite its fundamental importance, prejudice remains notoriously difficult to study (Kuppens and Spears 2014; Creighton and Jamal 2015; Linos, Jakli and Carlson 2021). Indeed, previous research on prejudice suggests that higher-educated individuals in particular may not reveal their true attitudes toward ethnic minorities in mass surveys (Kuppens and Spears 2014). Jackman and Muha (1984) argued that such individuals’ commitment to ethnic tolerance is merely superficial, since they often oppose initiatives meant to reduce racial inequality. Others have carried out rigorous experiments showing how prejudiced people are even more negative than they appear to be (Kuklinski, Cobb and Gilens 1997; Sniderman and Hagendoorn 2007). Likewise, in surveys applying direct questioning, Janus (2010) concluded that both higher-educated individuals and political liberals are more critical of greater diversity than they are willing to admit. Finally, Creighton and Jamal (2020) have revealed how the Brexit campaign increased sensitivity bias toward Muslims among ordinary citizens in the UK.

Even more, we know surprisingly little about the depth of goodwill toward ethnic minorities among their strongest supporters, referred to as “pure liberals” by Sniderman and Hagendoorn (2007). According to Sniderman and Hagendoorn’s (2007, 35) study, pure liberals have consistent attitudes inasmuch as they (a) remain unprejudiced toward ethnic minority members and (b) have no objections to norms or traditions among ethnic minorities. By implication, pure liberals are the proper devotees of a multiethnic society—or are they? We are aware of only one study that concludes that pure liberals are unprejudiced (Sniderman and Hagendoorn 2007, 35). In this context, prejudice means antipathy, which, in turn, relates intimately to social distancing (Allport 1979, 9, 14; Sniderman and Hagendoorn 2007). Following Careja (2016), prejudice is also directed at a specific group (i.e., Muslims) in contexts of ethnic diversification.

In view of this limited evidence on the attitudes of pure liberals, this IMR Research Note utilizes a novel convenience sample from Denmark with an embedded list experiment (fielded in 2018). Our study has three aims. First, we contribute to research on ethnic prejudice toward Muslims. Muslims represent a fast-growing ethnic minority group in Europe (Gusciute, Mühlau and Layte 2021), but they may face important barriers of inclusion, such as prejudice. Second, and more broadly, we contribute to research on the complexity of liberal attitudes toward immigrant inclusion processes. Self-interest and disapproval of specific Muslim values may “infect” liberal attitudes and ideals with prejudice. Third, as we utilize data from a survey of social-science students, we can indicate how parts of the educational elite in a well-established democracy react to immigration and Muslim immigrants.

Our results are remarkably different from Sniderman and Hagendoorn’s (2007) consistency claim, as we show that many pure liberals endorsed immigration but remained prejudiced toward Muslims. This gap between attitudes toward immigration and toward specific immigrants underpins what we call the inconsistency hypothesis. In terms of theoretical implications, our results suggest that the notion of pure liberals is a misnomer because it underestimates the psychological power of prejudice. We conclude by tentatively discussing how pure liberals’ prejudice toward Muslims should be interpreted.

Pure Liberals and the Power of Prejudice

In Europe, the inflow of migrants from the Middle East and Northern Africa has increased considerably over the past four decades (Gusciute, Mühlau and Layte 2021). How does the ethnic majority in Europe react to these significant developments and to Muslim immigration? In a cross-national study, Abdelaaty and Steele (2022) report that majority members’ opposition to Muslim immigration is common, but they also found that mass-level reactions toward immigration and immigrants are not uniform (see also Bohman 2018). How should this non-uniform pattern at the mass level be conceptualized? Sniderman and Hagendoorn (2007) strove to understand the motivations triggering non-uniform reactions among the Dutch majority group toward ethnic minority members. To accomplish this objective, they distinguished between three categories of majority members: “anti-Muslims,” who included majority members objecting to Muslim norms and disliking Muslims; “liberal critics,” who objected to some Muslim norms but had sympathy with Muslims in general; and “pure liberals,” who accepted Muslim norms and had sympathy with Muslims.

In this IMR Research Note, we focus on the pure liberal category because it most clearly rests on the assumption that its members remain unprejudiced. Yet this assumption is debatable. Indeed, Allport’s (1979, 17) seminal work comes to mind, particularly his claim that “human beings slip so easily into ethnic prejudice” because hostility is “a natural and common capacity of the human mind.” If prejudice has these powerful properties, majority members’ support for greater ethnic diversity may not necessarily convert into sympathy toward ethnic minority members. To qualify as pure liberals, however, majority members must consistently support greater ethnic diversity and remain unprejudiced toward Muslims.

We question the attitudinal consistency of the pure liberals but are not suggesting that they systematically offer untrue answers in surveys. Rather, and inspired by Jackman and Muha’s (1984) “superficial-commitment perspective,” we argue that it is possible that many pure liberals honestly support the general idea of a multiethnic society but remain quite prejudiced toward some of its ethnic minority members. Such a possible inconsistency, however, is not trivial, as it suggests that these majority members fail to recognize the implications of a liberal orientation; liberal is antithetical to prejudice (Sniderman and Hagendoorn 2007).

Our “inconsistency conjecture” builds on the observation that ethnic prejudice is surprisingly common across time and space and probably best described as a deep-seated, instinctive rejection of those who differ from the average majority member (for an extensive review of the prejudice literature, see Paluck et al. 2021). If ethnic prejudice includes an instinctive “driver,” it should be observable across different predispositions, and not only among anti-Muslims. Building on this rationale, we examine the following hypothesis:

The inconsistency hypothesis: Although pure liberals endorse greater ethnic diversity and lenient immigration policies, many remain prejudiced toward Muslims.

Data, Measures, and Experimental Design

To test our hypothesis, we initiated a web survey of 716 students in the Departments of Political Science and Psychology at Aarhus University (Denmark). Data collection took place November 5 − 17, 2018. Data qualify as a non-probability sample, since students were not randomly picked from various departments. Obviously, this student sample is not representative of the Danish population. Considering our theoretical objectives, however, we needed a sample biased in favor of greater diversity and unprejudiced reactions. As numerous studies have shown (e.g., Güemes and Herreros 2019; Thomsen and Olsen 2017), students and higher-educated individuals in consolidated democracies are high in social trust and disproportionately tolerant toward ethnic minority members. University students have an additional characteristic that suits our purposes: in democracies, universities recruit persons into the political and administrative elite, who often advocate lenient immigration policies, despite their unpopularity among lower-educated individuals (Thomsen and Olsen 2017). In democracies, steps toward greater ethnic diversity and multiculturalist arrangements are usually elite-driven (Statham and Geddes 2006; Green-Pedersen and Krogstrup 2008). Thus, it is important to ascertain to what extent such elite preferences also involve acceptance of ethnic minority members at a personal level. Ultimately, our convenience sample adds to understandings of the tensions between elite decision-makers and lower-status groups over globalization and increasing diversity in European societies (also Kriesi et al. 2006). These tensions are highly real if many elite decision-makers wish to avoid the personal consequences of their own immigration policy preferences.

Identifying Pure Liberals

Our empirical analysis involves identifying pure liberals according to specific measures that serve as selection criteria. According to Sniderman and Hagendoorn (2007, 23‒24, 35), pure liberals both like ethnic minority members and accept their social norms (“their way of life”). “Like” means unprejudiced in this context. Building on Sniderman and Hagendoorn’s (2007) item battery, we apply five measures with the following wording to identify “pure liberal attitudes” toward a specific ethnic minority group in the country: 1) “Muslims have something to offer to Danish culture,” 2) “The Danish and Muslim lifestyles are irreconcilable,” 3) “Most Muslims in Denmark respect other cultures,” 4) “Muslim men in Denmark dominate their women,” and 5) “Muslims in Denmark raise their children in an authoritarian manner.” All five are Likert items with five response categories. Pure liberals are deemed to be those who offered “tolerant” responses to all five items, implying a positive evaluation of Muslims (item 1 (agree/partly agree), item 2 (disagree/partly disagree), and item 3 (agree/partly agree)) together with rejection of claims about Muslims’ conservative conception of the liberties of women and children (item 4 (disagree/partly disagree) and item 5 (disagree/partly disagree)). Table A1 (Online Appendix) shows some descriptive details of the five items meant to identify pure liberals. Identifying pure liberals in this manner constitutes the basis of the analysis presented in Tables 1-2 (see below).

Table 1.

The Relationship Between Multiethnic Attitudes and Exclusionism/Inclusionism.

	Immigration should be made more difficult		Immigration should not be made more difficult
Pure liberals	2.8%	(4)	97.2%	(139)
Liberal critics	18.8%	(93)	81.2%	(401)
Combined	15.6%	(97)	84.4%	(540)

Note: Entries are percentages, and the absolute number of observations is in parentheses. A t-test showed that the means of the pure liberals and liberal critics are significantly different in both columns (p < 0.001; two-tailed).

Table 2.

Means Tests of the List Experiment.

	Average	Standard error	N	Test of difference (µ_stimuli-µ_control)
Among pure liberals :
Control group	1.216	0.095	74	t = 2.75
Treatment group	1.554	0.098	74
Difference (treatment minus control)	0.338**	0.137		p = 0.0067
Among liberal critics :
Control group	1.364	0.042	272	t = 2.64
Treatment group	1.566	0.065	244
Difference (treatment minus control)	0.202**	0.076		p = 0.0085

Note: t-test (two-tailed). **p < 0.01.

Table A1 (Online Appendix) also distinguishes between pure liberals and liberal critics. Although not fully consistent with previous criteria from Sniderman and Hagendoorn’s (2007) work, the label “liberal critics” covers those who did not offer five tolerant responses. Unlike pure liberals, liberal critics had mixed feelings toward greater ethnic diversity. Moreover, liberal critics are probably still more tolerant than the average Dane, due to our sample’s composition, as indicated by the fact that 241 of 516 liberal critics only differed from pure liberals on two items (Table A1, Online Appendix). Most importantly, the category, liberal critics, serves as a reference point meant to indicate the distinctively unprejudiced stand of pure liberals. Even more, according to the theoretical criteria offered earlier, both pure liberals and liberal critics should remain unprejudiced toward Muslims. Table 2 below examines this claim.

To further indicate pure liberals’ distinctiveness in terms of unprejudiced reactions toward Muslims, Table A2 (Online Appendix) compares this group with a representative sample of Danish citizens as regards their respective negative reactions toward Muslims. The two items we apply to establish to what extent the two groups (pure liberals and Danes) differ are not identical but intimately related in terms of substance. Table A2 shows that pure liberals were much less negative toward Muslims than the average Dane; the difference was more than one scale point. Thus, students who were pure liberals had little in common with the average Danish citizen.

Finally, we include the following measure: “Do you think that immigration to this country should be made more difficult?” This measure taps pure liberals’ policy preferences, indicating support for inclusionary immigration policies and open borders (with two response categories: “yes”/“no”; “don’t know” responses were excluded). This variable’s distribution was highly skewed toward an overweight of inclusionary responses, confirming existing research on higher-educated individuals’ immigration preferences (Thomsen and Olsen 2017). We apply this item to specify the extent to which pure liberals consistently supported the transition to greater ethnic diversity in modern democracies (see below).

Designing the List Experiment

The list experiment tests our hypothesis, and this “item count technique” has been widely used to examine attitudes on sensitive issues (see Krumpal 2013). The central idea of the list experiment is to create a hidden measure of prejudice to reduce the incidence of social desirability bias (SDB) (Kuklinski, Cobb and Gilens 1997; Sniderman and Hagendoorn 2007). Participants in our survey were randomly sorted into the control and treatment groups; the only aspect that varied between the groups was the number of statements offered to respondents.

Our survey included a list experiment meant to address sensitive issues (i.e., prejudice). In this regard, it may be considered a quality that participants answered the questionnaire on their own personal computer screen and not in a traditional face-to-face interviewer situation. Indeed, Krumpal (2013) notes that on list experiments, some respondents find it difficult to decide which items to apply and to keep a running tally of them. It may be less psychologically stressing, as in our design, when the list of items appears on the computer screen without response-time limits.

Respondents in our sample received the following written prompt on the computer screen: “You are now going to read a list of things that sometimes annoy people. After you have read all of them, please tell me how many of them annoy you. I don’t want to know which ones—just how many.” Respondents, therefore, knew that the interviewer could not observe their exact answers, which provides the essential feeling of privacy. Figure 1 shows the wording of the specific items in our experiment. The control group received four different items about 1) professional football players’ income, 2) television commercials that interrupt movies, 3) parents who bring their children to the supermarket, and 4) shops that are closed on Sundays. The treatment group received these four and an additional one: the hidden measure of prejudice: “if one of your siblings or best friends found a Muslim boyfriend/girlfriend” (you would get annoyed). The highest possible score was, therefore, 4 in the control group and 5 in the treatment group. The average score can only be higher in the treatment group as a result of the hidden prejudice indicator’s stimulus. By subtracting the average number of the control group from the average number of the treatment group, we can identify the percentage of respondents who were annoyed by the prejudice indicator (Holbrook and Krosnick 2010, 44).

Figure 1.

Items in the list experiment.

The stimulus (in boldface in Figure 1) is a prejudice indicator meant to activate a) antipathy toward a specific ethnic minority group (i.e., Muslims) and b) social distancing. We could have referred to other minorities, but Muslims constitute the largest ethnic (and religious) minority group in Denmark (Sinclair 2022). Can we be certain that “Muslim” is causing the expected reaction, rather than the words boyfriend or girlfriend? Note that the word “Muslim” serves the semantic purpose of characterizing a specific type of girlfriend or boyfriend. Moreover, having a boyfriend or girlfriend is very common among persons in their early twenties (Chung 2005). Thus, this specific wording does not refer to something unusual for the persons involved in the experiment. We also followed traditional procedures recommending that the stimulus not be listed as the last indicator (Sniderman and Hagendoorn 2007). Finally, we chose a weak indicator of prejudice, which only taps hostility and exclusionary reactions toward Muslims in an indirect way. Very direct indicators were inapplicable, since respondents’ social lives differ from many other groups. Specifically, students do not have “colleagues,” and most are unmarried without children (Lairio, Puukari and Kouvo 2013). It also made little sense to examine their reactions toward new neighbors. Accordingly, the present test may be considered conservative, as very direct measures of prejudice probably trigger the most negative reactions.

Control indicators were deliberately phrased as meaningful and with a varying capacity to upset the respondent. We also wanted to avoid excessively repugnant items capable of provoking contrast effects. According to the literature on list experiments (Wolter and Laier 2014), the semantic differentiation among the non-key items is important to avoid ceiling or floor effects. In particular, if most respondents are upset by all four statements, it follows that prejudiced individuals must choose the fifth indicator of prejudice in the treatment group. This type of psychological “pressure” negates the experiment’s very logic, as it becomes obvious that these individuals are prejudiced (Kuklinski, Cobb and Gilens 1997; Sniderman and Hagendoorn 2007). We pre-tested the list experiment, which showed that respondents were, on average, annoyed by 1‒2 of the control (non-key) indicators. This finding is consistent with the results from Sniderman and Hagendoorn (2007, 34). Table A3 (Online Appendix) shows the distribution of responses on the list experiment’s items.

The list experiment assumes that the specific experimental stimulus fully accounts for the differences between the control and treatment groups, which, in turn, requires that the two groups are identical on other key characteristics. Table A5 (Online Appendix) shows that the randomization procedure was successful, since there are no statistically significant differences between the control and treatment groups on key variables. More importantly, a critical assumption of the list experiment design is the absence of so-called design effects (Blair and Imai 2012, 49). The stimulus should not influence how respondents react to the control items. Obviously, this type of design effect ruins the control logic (Imai 2011; Blair and Imai 2012, 49). In technical terms, according to Blair and Imai (2012, 68), effects should not be significantly negative in sign. Fortunately, Table A4 (Online Appendix), which is based on Table A3, shows two negative effects that are not statistically significant. Accordingly, our analysis meets the critical assumption, implying that a difference between the treatment and control groups is attributable to the specific stimulus rather than to anything else.

Finally, there is the questionnaire’s structure. Consistent with conventional guidelines (Iyengar and Kinder 1987), the survey was organized to avoid priming particular ethnicity-related attitudes prior to the experimental stimulus. Respondents were, therefore, confronted in the interview situation with the experiment prior to their reactions to the questions addressing various characteristics about Muslims. Likewise, the experiment was introduced early in the survey to minimize fatigue-related measurement error.

Empirical Results

Recall that we had five items measuring attitudes toward Muslims and their social practices. The five items were intended to identify pure liberals. More specifically, the “pure liberals” category comprised only those individuals who responded positively to all items (fully/partly agree with “tolerant” items and fully/partly disagree with “intolerant” items). Remaining respondents scoring differently are labelled “liberal critics.” This identification procedure serves as the basis for the subsequent analyses.

Table 1 shows an initial and weak test of how consistently pro-immigration pure liberals were and focuses on the item, “Do you think that immigration to this country should be made more difficult?” Certainly, we are unlikely to find many respondents who supported more exclusionist immigration policies, as the sample only included higher-educated persons. Indeed, Table 1 shows that 97.2 percent of pure liberals in the sample opposed harsher immigration policies. As expected, pure liberals’ support for greater ethnic diversity was unmistakably clear, providing initial support for the inconsistency hypothesis.

The list experiment examines the extent to which support for greater diversity in every respect (both liking Muslims and accepting their social practices) implies unprejudiced reactions at a personal level. At issue is, therefore, whether pure liberals were fully committed to a multiethnic society. Before proceeding to the main analysis, we wish to emphasize that we have examined the key assumptions of the applied statistical model. The assumptions of the t-test about interval-scaled variables and the randomization of the two groups were met. Moreover, as we are performing t-tests, we made an additional analysis in which Levene’s test confirmed the occurrence of homoscedasticity (i.e., equal variances in the control and treatment groups).

Turning to the substantial results, Table 2 shows a statistically significant difference between the averages of the control (1.216) and treatment groups (1.554). The difference between the two averages means that a significant share of the treatment group would grow annoyed if their siblings or best friends found a Muslim girlfriend or boyfriend. More specifically, when the difference in means between the treatment and control groups is multiplied by 100, almost 34 percent of pure liberals would appear to be prejudiced toward Muslims. This percentage is our best estimate of the occurrence of ethnic prejudice among the most “tolerant.”

Previous research on prejudice has addressed the SDB issue when reporting anti-Muslim sentiment. Blair, Coppock and Moor (2020), for example, conclude that measures of prejudice have no sensitivity bias. Creighton and Jamal (2015) find that Muslims enjoy no protection from SDB. Our prejudice indicator (avoidance of cross-group contact) was only included in the experiment. How respondents would have answered the prejudice item without privacy protection remains unknown. Consequently, we cannot tell whether pure liberals wished to hide their prejudice toward Muslims due to item sensitivity. In any case, it follows that the analysis supports our inconsistency hypothesis, claiming that many self-declared liberals were, in fact, prejudiced toward Muslims. The observed prejudice among pure liberals also means that we identify an interesting attitudinal inconsistency between their general support for immigration and rejection of particular immigrants.

An important issue in exploring the normalcy of prejudice concerns the extent to which the stimulus had a similar impact among pure liberals and liberal critics. Drawing on the superficial commitment perspective from Jackman and Muha (1984), we expect pure liberals to be just as prejudiced as liberal critics. Table 2 shows that the stimulus worked among liberal critics, as 20.2 percent of them were prejudiced. Moreover, Table A6 (Online Appendix) reveals that the difference between pure liberals and liberal critics (0.136) was non-significant. In view of this finding, it is reasonable to wonder if the analyses are underpowered, as indicated by a two one-sided equivalence test, applying an upper (d = 0.3) and lower equivalence bound (d = − 0.3) (see Lakens 2017). The equivalence test in Figure A7 (Online Appendix) shows that the means among pure liberals and liberal critics were neither statistically different nor equivalent. From the equivalence test, it follows that we cannot determine whether pure liberals responded differently to the stimulus than did liberal critics. The similarity in reactions to the stimulus is an essential observation that is consistent with Allport’s (1979) claim that prejudice is a common disposition.

Another check for robustness concerns the extent to which the analysis suffers from floor-ceiling effects. Table A3 (Online Appendix) suggests that floor and ceiling effects may not be a serious problem. Regarding ceiling effects, Table A3 shows that only three respondents (0.8%) in the control group scored the value 4, meaning that the corresponding share of respondents in the treatment group who found all control items annoying (and, thus, capable of generating ceiling effects) must also be small. Similarly, respondents capable of generating floor effects had the response value 0 in the treatment group. Again, the share of respondents scoring this value is small. Moreover, because the majority of this group were unlikely to be annoyed by the stimulus (in the entire experiment, only 1 in 5 were annoyed), it follows that floor effects can only have a limited impact. Finally, potential floor-ceiling effects most likely caused the underestimation of key estimates, not the opposite. By implication, our estimate of the scope of ethnic prejudice among pure liberals may be considered conservative.

Naturally, because of pretreatment effects, we cannot be sure that the estimate of the number of prejudiced individuals in the treatment group is entirely unbiased. Again, we may underestimate the number of prejudiced liberals because pretreatment in public communication can reduce the experimental stimulus effect (Druckman and Leeper 2012). Related to our experimental stimulus, the word “Muslim” may also cause unease among the most prejudiced, as it connotes politically controversial issues in Denmark (Meer et al. 2015). The true prejudice estimate remains unknown.

Conclusion

As this IMR Research Note utilized a convenience sample of Danish social-science students, our conclusions relate to contemporary elites more than to the average Danish citizen. On this empirical material, we have tested a new inconsistency hypothesis, addressing the characteristics of liberal attitudes. Results provisionally support our inconsistency hypothesis, since many pure liberals in our sample supported greater ethnic diversity but remained prejudiced toward Muslims. Even more, with respect to prejudice, pure liberals did not differ from liberal critics. Consequently, our list experiment suggests that the theoretical label “pure liberals”’ serves no clear analytical purpose because it fails to identify a uniform class of unprejudiced individuals. Rather, the results resonate with Allport’s (1979, 17) “normality of prejudgment” claim.

What accounts for the combination of support for greater diversity and prejudice toward Muslims among many pure liberals? We see three different interpretations. First, the “Not in My Backyard” (NIMBY) literature offers a plausible interpretation of the reasoning processes (McAvoy 1998). NIMBYism addresses opposition to developments that individuals believe will have harmful consequences for their personal lives (Wexler 1996). The NIMBY approach may offer a foundation for claiming that self-interest sometimes nurtures prejudice. In our case, many pure liberals supported greater ethnic diversity but also wished to prevent such social change from affecting their personal lives. Even so, the exact characteristics of the notion of self-interest in this context needs further specification.

Second, inconsistent attitudes among pure liberals may indicate that they support immigration and cultural diversity but reject Muslims (rather than any other immigrant group) because they believe that Muslims are illiberal and too religious. Obviously, illiberal attitudes and support for strict religious norms contradict liberal worldviews. Such “principled” reactions among pure liberals, however, should be related to the fact that the category of Muslims in Europe is highly differentiated (Statham and Tillie 2016). To what extent pure liberals recognize this differentiation remains to be examined.

Third, inconsistent attitudes among pure liberals are not straightforward to interpret, as the extent to which SDB governs their responses to items not included in the experiment remains unsettled. Previous studies reached mixed conclusions about SDB’s role in studies of prejudice. Creighton and Jamal (2015) concluded that Muslims enjoy no protection from SDB. In contrast, other studies have concluded that opposition to immigration is greater when measured by hidden indicators (Janus 2010, An 2015; Creighton, Schmidt and Zavala-Rojas 2018). In view of these conclusions, SDB likely also affects our findings related to support for greater diversity, but its influence may easily be exaggerated, as Creighton and Jamal (2015) indicate. Specifically, support for lenient immigration policies among pure liberals in our sample was so massive that it must indicate some genuine commitment to greater ethnic diversity (see Table 1). Expanding on this commitment, it is well established that “tolerant” voters systematically avoid (anti-immigration) radical right parties (e.g., Zhirkov 2014). We take this electoral behavior to be indicative of some commitment to a multiethnic society among tolerant citizens, albeit not necessarily indicative of unprejudiced dispositions.

The present analyses also speak to a longstanding debate about how the impact of higher education on liberal attitudes should be interpreted (e.g., Jackman and Muha 1984; Thomsen and Olsen 2017). Indeed, the positive relationship between education and liberal attitudes is a remarkably consistent finding across the social sciences (Cavaille and Marshall 2019). At first glance, our analysis supports this observation. Upon closer inspection, however, our list experiment qualifies the liberalizing impact of education by demonstrating that many (educated) pure liberals were only superficially committed to greater diversity because they literally rejected Muslims in their own private lives.

To conclude, the liberalizing effect of education is often related to socialization processes in the educational system through which citizens are taught that prejudice is wrong (Thomsen and Olsen 2017). The argument here, then, is that were it not for socialization processes in the education system, prejudice would probably have an even stronger impact on interethnic relations. Finally, and emphasizing the independent role of socialization, most of the pure liberals in our sample were unprejudiced. Nonetheless, the takeaway message from our empirical study is that prejudice toward Muslims flourishes among the otherwise cosmopolitan educational elite.

Supplemental Material

sj-docx-1-mrx-10.1177_01979183221116871 - Supplemental material for Do Pure Liberals Both Endorse Immigration and Reject Muslims? Examining the Inconsistency Hypothesis

Supplemental material, sj-docx-1-mrx-10.1177_01979183221116871 for Do Pure Liberals Both Endorse Immigration and Reject Muslims? Examining the Inconsistency Hypothesis by Jens Peter Frølund Thomsen and Mikkel Terkelsen in International Migration Review

Footnotes

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.

ORCID iD

Jens Peter Frølund Thomsen

Supplemental Material

Supplemental material for this article is available online.

References

Abdelaaty

Steele

L. G.

. 2022. “Explaining Attitudes Toward Refugees and Immigrants in Europe.” Political Studies 70 (1): 110–30. https://doi.org/10.1177/0032321720950217

Allport

G. W.

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

B. P.

2015. “The Role of Social Desirability Bias and Racial/Ethnic Composition on the Relation Between Education and Attitude Toward Immigration Restrictionism.” The Social Science Journal 52 (4): 459–67. https://doi.org/10.1016/j.soscij.2014.09.005

Blair

Coppock

Moor

. 2020. “When to Worry About Sensitivity Bias: A Social Reference Theory and Evidence from 30 Years of List Experiments.” American Political Science Review 114 (4): 1297–315. https://doi.org/10.1017/S0003055420000374

Blair

Imai

. 2012. “Statistical Analysis of List Experiments.” Political Analysis 20 (1): 47–77. https://doi.org/10.1093/pan/mpr048

Bohman

2018. “Who’s Welcome and Who’s Not? Opposition Towards Immigration in the Nordic Countries, 2002–2014.” Scandinavian Political Studies 41 (3): 283–306. https://doi.org/10.1111/1467-9477.12120

Careja

2016. “Party Discourse and Prejudiced Attitudes Toward Migrants in Western Europe at the Beginning of the 2000s.” International Migration Review 50 (3): 599–627. https://doi.org/10.1111/imre.12174

Cavaille

Marshall

. 2019. “Education and Anti-Immigration Attitudes: Evidence from Compulsory Schooling Reforms Across Western Europe.” American Political Science Review 113 (1): 254–63. https://doi.org/10.1017/S0003055418000588

Chung

2005. “Violence, Control, Romance and Gender Equality: Young Women and Heterosexual Relationships.” Women’s Studies International Forum 28 (6): 445–55. https://doi.org/10.1016/j.wsif.2005.09.005

10.

Creighton

M. J.

Jamal

. 2015. “Does Islam Play a Role in Anti-Immigrant Sentiment? An Experimental Approach.” Social Science Research 53: 89–103. https://doi.org/10.1016/j.ssresearch.2015.04.001

11.

Creighton

M. J.

Jamal

. 2020. “An Overstated Welcome: Brexit and Intentionally Masked Anti-Immigrant Sentiment in the UK.” Journal of Ethnic and Migration Studies 48 (5): 1051–71. https://doi.org/10.1080/1369183X.2020.1791692

12.

Creighton

M. J.

Schmidt

Zavala-Rojas

. 2018. “Race, Wealth and the Masking of Opposition to Immigrants in the Netherlands.” International Migration 57 (1): 245–63. https://doi.org/10.1111/imig.12519

13.

Druckman

J. N.

Leeper

. 2012. “Learning More from Political Communication Experiments: Pretreatment and Its Effects.” American Journal of Political Science 56 (4): 875–96. https://doi.org/10.1111/j.1540-5907.2012.00582.x

14.

Green-Pedersen

Krogstrup

. 2008. “Immigration as a Political Issue in Denmark and Sweden.” European Journal of Political Research 47 (5): 610–34. https://doi.org/10.1111/j.1475-6765.2008.00777.x

15.

Gusciute

Mühlau

Layte

. 2021. “All Welcome Here? Attitudes Towards Muslim Migrants in Europe.” International Migration 59 (5): 149–65. https://doi.org/10.1111/imig.12810

16.

Güemes

Herreros

. 2019. “Education and Trust: A Tale of Three Continents.” International Political Science Review 40 (5): 676–93. https://doi.org/10.1177/0192512118779184

17.

Holbrook

A. L.

Krosnick

J. A.

. 2010. “Social Desirability Bias in Voter Turnout Reports: Tests Using the Item Count Technique.” Public Opinion Quarterly 74 (1): 37–67. https://doi.org/10.1093/poq/nfp065

18.

Imai

2011. “Multivariate Regression Analysis for the Item Count Technique.” Journal of the American Statistical Association 106 (494): 407–16. https://doi.org/10.1198/jasa.2011.ap10415

19.

Iyengar

Kinder

D. R.

. 1987. News That Matters. Chicago: University of Chicago Press.

20.

Jackman

M. R.

Muha

M. J.

. 1984. “Education and Intergroup Attitudes: Moral Enlightenment, Superficial Democratic Commitment, or Ideological Refinement?” American Sociological Review 49 (6): 751–69. https://doi.org/10.2307/2095528

21.

Janus

A. L.

2010. “The Influence of Social Desirability Pressures on Expressed Immigration Attitudes.” Social Science Quarterly 19 (4): 928–48. https://doi.org/10.1111/j.1540-6237.2010.00742.x

22.

Kriesi

Grande

Lachat

Dolezal

Bornschier

Frey

. 2006. “Globalization and the Transformation of the National Political Space: Six European Countries Compared.”.” European Journal of Political Research 45 (6): 921–56. https://doi.org/10.1111/j.1475-6765.2006.00644.x

23.

Krumpal

2013. “Determinants of Social Desirability Bias in Sensitive Surveys: A Literature Review.” Quality and Quantity 47 (4): 2025–47. https://link-springer-com-443.web.bisu.edu.cn/content/pdf/10.1007/s11135-011-9640-9.pdf.

24.

Kuklinski

J. H.

Cobb

M. D.

Gilens

. 1997. “Racial Attitudes and the ‘New South’.” The Journal of Politics 59 (2): 323–49. https://doi.org/10.1017/S0022381600053470

25.

Kuppens

Spears

. 2014. “You Don’t Have to Be Well-Educated to Be an Aversive Racist, But it Helps.” Social Science Research 45: 211–23. https://doi.org/10.1016/j.ssresearch.2014.01.006

26.

Lairio

Puukari

Kouvo

. 2013. “Studying at University as Part of Student Life and Identity Construction.” Scandinavian Journal of Educational Research 57 (2): 115–31. https://doi.org/10.1080/00313831.2011.621973

27.

Lakens

2017. “Equivalence Tests: A Practical Primer for t Tests, Correlations, and Meta-Analyses.” Social Psychological and Personality Science 8 (4): 355–62. https://doi.org/10.1177/1948550617697177

28.

Linos

Jakli

Carlson

. 2021. “Fundraising for Stigmatized Groups. A Text Message Donation Experiment.” American Political Science Review 115 (1): 14–30. https://doi.org/10.1017/S0003055420000787

29.

McAvoy

1998. “Partisan Probing and Democratic Decision-Making: Rethinking the NIMBY-Syndrome.” Policy Studies Journal 26 (2): 274–92. https://doi.org/10.1111/j.1541-0072.1998.tb01899.x

30.

Meer

Mouritsen

Faas

de Witte

. 2015. “Examining ‘Postmulticultural’ and Civic Turns in the Netherlands, Britain, Germany, and Denmark.” American Behavioral Scientist 59 (6): 702–26. https://doi.org/10.1177/0002764214566496

31.

Paluck

E. L.

Porat

Clark

C. S.

Green

D. P.

. 2021. “Prejudice Reduction: Progress and Challenges.” Annual Review of Psychology 72: 533–60. https://doi.org/10.1146/annurev-psych-071620-030619

32.

Sinclair

2022. “What Does It Mean to be Danish? The Integration of Muslims in Denmark in a Historical Perspective.” Journal of International Migration and Integration., 1–17. https://doi.org/10.1007/s12134-021-00926-x

33.

Sniderman

P. M.

Hagendoorn

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

34.

Statham

Geddes

. 2006. “Elites and the ‘Organised Public’: Who Drives British Immigration Politics and in Which Direction?” West European Politics 29 (2): 248–69. https://doi.org/10.1080/01402380500512601

35.

Statham

Tillie

. 2016. “Muslims in Their European Societies of Settlement: A Comparative Agenda for Empirical Research on Socio-Cultural Integration Across Countries and Groups.” Journal of Ethnic and Migration Studies 42 (2): 177–96. https://doi.org/10.1080/1369183X.2015.1127637

36.

Thomsen

J. P. F.

Olsen

. 2017. “Re-Examining Socialization Theory: How Does Democracy Influence the Impact of Education on Anti-Foreigner Sentiment?” British Journal of Political Science 47 (4): 915–38. https://doi.org/10.1017/S0007123415000496

37.

Wexler

M. N.

1996. “A Sociological Framing of the Nimby (Not-in-my-Backyard) Syndrome.” International Review of Modern Sociology 26 (1): 91–110. https://www-jstor-org-443.web.bisu.edu.cn/stable/41421101?seq = 1#metadata_info_tab_contents .

38.

Wolter

Laier

. 2014. “The Effectiveness of the Item Count Technique in Eliciting Valid Answers to Sensitive Questions: An Evaluation in the Context of Self-Reported Delinquency.” Survey Research Methods 8 (3): 153–68. https://doi.org/10.18148/srm/2014.v8i3.5819

39.

Zhirkov

2014. “Nativist But Not Alienated: A Comparative Perspective on the Radical Right Vote in Western Europe.” Party Politics 20 (2): 286–96. https://doi.org/10.1177/1354068813511379

40.

Zick

Küpper

Höverman

. 2011. Intolerance, Prejudice and Discrimination: A European Report. Nora Langenbacher: Forum Berlin.

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

0.04 MB