![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
ABSTRACT
Research on attitudes towards immigrants devotes much attention to the relative effects of economic and social-psychological factors for understanding sentiment towards immigrants, conceived in general terms. In this article, we advance this work by arguing that the context framing immigration concerns leads publics to associate different types of immigrants with different threats. An issue context that diminishes support for one ‘type’ can boost it for another. Evidence from an original survey experiment in Britain supports this claim. Security fears affect attitudes towards Muslim immigrants but economic concerns bear on views towards Eastern Europeans. While concern about crime adversely affects sentiment for East Europeans but casts Muslims more positively, cultural threats have the opposite effect. By shifting the focus onto the qualities of different types of immigrants, we highlight the importance of the target immigrant group for understanding public attitudes.
Acknowledgements
The authors thank Harold D. Clarke and the British Elections Study for incorporating their experiment and questions into the June 2011 Continuous Monitoring Survey.
Notes
‡ The underlying research materials for this article can be accessed at http://www.sinno.com/publications
1. Lack of attention may be due to limited data – few public opinion surveys allow analysts to discern individuals’ attitudes towards different types of immigrants relative to their views of immigrants with other attributes. The 2002 New Zealand Election Study asks respondents about support for immigrants from ‘Muslim countries’. The 1998 and 2001 Australian Election Studies include an item pertaining to ‘migrants who are from the Middle East’. The 2006 Pew Global Attitudes Survey asks respondents from four Western democracies whether it is ‘a good thing or a bad thing that people from the Middle East and North Africa come to live and work in this country’. The 2003 British Social Attitudes (BSA) survey asks for attitudes that relate to individuals from three immigration areas: South Asia, the European Union, and Australia/New Zealand. The Transatlantic Trends Survey (Citation2011) includes an embedded experiment to gauge sentiment towards Muslims and generic immigrants in five European countries and Hispanic versus generic immigrants in the US. The survey also gauges differences of attitudes towards the integration of each dyad of immigrants but does not address threats that are not cultural in the experiment. It does however address them in non-experimental survey questions about generic immigrants.
2. Examples include McLaren and Johnson Citation2007; Sniderman and Hagendoorn Citation2007; Brader, Valentino, and Suhay Citation2008; Strabac and Listhaug Citation2008; Ford Citation2011; Hainmueller and Hopkins Citation2015; and Adida, Laitin, and Valfort Citation2016.
3. A case in point is the emerging consensus pointing to socio-psychological factors as driving immigrant sentiment among European and North American publics (Hainmueller and Hopkins Citation2014). This finding may stem from a particular identity respondents attach to ‘immigrants’. In the US, for instance, biases in media coverage imply that individuals are far more likely to view immigrants as Hispanic than as East Asian (Brader, Valentino, and Suhay Citation2008).
4. These expectations should apply to publics in many immigrant-receiving European countries. Applying the analyses on the British case to other contexts is grounds for future research.
5. Ford (Citation2011) addresses the complex cultural factors that affect Britons’ attitudes towards immigrants from different regions in a longitudinal study of BSA survey data. Ford’s study is limited, however, in that the BSA data do not allow him to provide precise mechanisms linking wide cultural proxies with sentiment.
6. Hames (Citation2004), reported in Cekalova (Citation2008)
7. On Muslims, see Bolognani (Citation2009). On East Europeans, see Campbell (Citation2012). Another group, Nigerians, is also often associated with fraud. One complication is that some East European immigrants (e.g. Albanians or Kosovars) who are frequently covered in media outlets could be of ‘Muslim’ background, but they are generally considered East Europeans rather than Muslims. See also Endley (Citation2014).
8. The questionnaire and data are available at http://bes2009-10.org/cms-data.php.
9. Of the 1,009 respondents who took the online survey, 353 received the generic control treatment, 349 the ‘Muslim’ treatment, and 307 the ‘East European’ treatment. We report the distributions of key demographic variables for participants in these three randomly assigned groups in Appendix A in the supplementary information file.
10. Item construction follows that used in waves of the European Social Survey (e.g., Hainmueller and Hiscox Citation2007). ‘Don’t know’ and no answer responses, comprising 7.6 percent of the sample, are omitted from the analysis.
11. In all cases, subjects were asked to respond to items with respect to the same immigrant group, such that each frame contained separate randomisations. We considered an alternative design with subjects randomized according both to type of immigrant and one of four issue frames (economic, culture, security, crime). However, this would require dividing subjects into twelve groups and would not provide a sufficient number of cases per treatment on which to gauge statistical inference. The substantive differences we find suggest that our design is capable of isolating the effect of both the type of immigrant and type of issue.
12. We also included a set of items that are not theoretically driven to maintain parity between negative and positive frames (see Appendix B).
13. Question wording appears in the supplementary file. We did not have a suitable retrospective question for culture because of the complexity of cultural considerations. Note that items were put to respondents earlier in the omnibus CMS questionnaire, thereby posing little risk of affecting responses to the experimental items questions.
14. See Appendix B. The distribution across response categories is similar to that for similarly worded items from the European Social Survey modules.
15. Education is coded as 1=14 or fewer years of formal education, 2 = 15 years, 3 = 16, 4 = 17–18, 5 = 19–20, and 6 = 21 or more; Age is coded 1 = 18–24 years old, 2 = 25–39, 3 = 40–54, and 4 = 55 and over. Demographic items are included for bases of comparability with other studies (e.g., Hainmueller and Hiscox Citation2010) and have no effect on the results of the main coefficients of interest. Analyses use demographic weights.
16. As a robustness test, we ran Model 4 with indicators of issue salience instead of retrospective performance assessments. Results were generally similar but with wider confidence intervals because responses clustered at the higher (‘more important problem’) end of the scales.
17. The likelihood ratio test of the null hypothesis that all interactive parameters in Model 4 are zero is rejected at χ2 (6) = 14.49, p = 0.02.
18. Comparing this positive frame with the previous two must be done with care, however, because positive frames generally elicit a more positive reaction.
19. This finding agrees with McLaren and Johnson (Citation2007) and Sniderman, Hagendoorn, and Prior (Citation2004).
20. This is not surprising considering that the social threat of Muslim gender inequality is one of four dominant themes in the British tabloids’ coverage of Muslims (Richardson Citation2004).
21. There is no statistical significance to the difference between the means for generic immigrants and for East European immigrants but a statistically meaningful one between the mean for the Muslim treatment and the means for the two other treatments.
22. Our findings indicate a difference between the Britain of 2011 and the Netherlands of 1997–8. The analysis by Sniderman, Hagendoorn, and Prior (Citation2004, 38) of the latter case find that ‘threats to safety are the least important in accounting for hostility to ethnic minorities’ (emphasis added). It is consistent, however, with McLaren and Johnson’s (Citation2007) findings based on a 2003 British survey on the importance of crime.
23. Eurostat, ‘Foreign citizens and foreign-born population’, Reference: STAT/12/105, 11/07/2012.