ABSTRACT
In this paper, we examine the association between contact with migrant populations and support for the populist radical right (PRR) in Switzerland. Building on group threat and intergroup contact theories, which offer opposing predictions, and drawing on Appadurai’s thesis of the ‘fear of small numbers’, we propose a new theoretical framework to explain this association. We predict that the relationship between the size of the migrant populations and PRR voting is nonlinear: a small but noticeable minority triggers the formation of anti-immigrant attitudes, which soften as the minority grows and people start having meaningful interactions with foreigners. To test these theories, we combine individual-level data with municipality-level information. Mixed-effects multilevel models confirm that individuals in municipalities with a moderate proportion of foreigners are more likely than those with fewer or a greater number of migrants to cast their vote in support of PRR parties; this is particularly so for certain stigmatised minorities. We further explore the effect of perceived immigrant threat in moderating these relationships.
Acknowledgements
We would like to thank Jeffrey Murer for his comments in an earlier version of this manuscript.
Disclosure statement
No potential conflict of interest was reported by the authors.
ORCID
Javier García-Manglano http://orcid.org/0000-0001-7233-8770
Effrosyni Charitopoulou http://orcid.org/0000-0001-9647-1234
Notes
1. Note that Hjerm (Citation2009) explores anti-immigrant attitudes rather than PRR voting behaviour; his findings are of relevance as PRR parties have often been defined as ‘anti-immigrant’ (see Boomgaarden and Vliegenthart Citation2007; Fennema Citation1997; Van der Brug, Fennema, and Tillie Citation2005).
2. Due to data availability, there is a one-year difference between contextual- and individual-level variables. Given the low year-to-year variation of context-level variables, this discrepancy should not be of greater concern.
3. This measure's reliability is high, with a Cronbach's alpha of 0.76.
4. We also tested models coding threat as a continuous variable. Results were similar for the main multilevel models (see ), but generated problems with convergence when adding random effects and cross-level interactions between threat and proportion of migrants. Thus we used threat as a dummy variable for high threat, as described in text.
5. This variable has 139 missing cases (5% of the sample); instead of creating an additional category for these cases as with other variables, we dropped them. Hence, the sample size of models including ‘perceived threat’ is reduced to 2541 individuals in 916 municipalities.