People’s tendency to exhibit discriminatory behaviour and favour those with shared characteristics leads to suboptimal outcomes at the individual and societal level. While the characteristics that induce biased decisions in diverse societies are well-documented, their relative and intersectional nature is less well understood. In this study, we address this limitation and conduct a combined multifactorial and split-sample choice experiment, implemented online with an ethnically diverse group of 816 participants based in the United Kingdom.

Our design, thanks to the random assignment of economic status across participants, allows us to address the problem of endogeneity resulting from an overlap between social and economic characteristics among individuals. As a result, we are able to disentangle taste for discrimination from statistical discrimination.

What is more, it addresses potential social desirability bias thanks to the trade-offs included in the conjoint experiment, while observing high validity as the profiles in the conjoint experiment represent actual respondents in our sample that made choices on additional financial incentives that were distributed among respondents after the survey was completed.