Using a Factorial Survey Experiment to Unbundle Ethnicity and Phenotype in Welfare Attitudes

Abstract

This seminar discusses factorial survey experiments as a methodological strategy for identifying the causal effects of multiple attributes on political preferences. The methodological discussion is illustrated with an experiment on support for redistribution in Spain. The design is particularly useful in contexts where different social cues tend to overlap in the real world and are therefore difficult to separate observationally. In the empirical example presented here, a factorial vignette experiment randomly varies the ancestry (native, Latin American, or North African) and phenotype (light-skinned Caucasian or Black) of a potential beneficiary of a housing subsidy for low-income young people, while holding constant the claimant’s socioeconomic circumstances. Using an online sample of 2,386 Spanish citizens, the seminar shows how this design makes it possible to estimate the independent effects of ancestry and phenotype and to examine whether these effects vary by political ideology. The findings suggest that ancestry cues matter more than phenotype on average, with support declining most clearly when the claimant is of North African background, especially among right-wing respondents. The seminar therefore highlights both the methodological advantages of factorial experiments and their substantive relevance for the study of redistribution, deservingness, and diversity.

Chair

Simona Piattoni, Università di Trento