Bias in Hiring: Experimental Evidence on Discrimination Based on Partisanship

• In Event: Group-Based Reasoning, Prejudice, and Political Discrimination

Abstract

Do employers discriminate based on partisanship? Partisan bias---a preference for in-partisans and against out-partisans---is rampant, leading to discrimination based on partisanship in both the political and the non-political sphere. As such, discrimination in hiring based on partisanship seems likely. Yet, we know very little about whether this is actually the case. Only one study to date has investigated, and documented, discrimination based on partisanship in hiring---specifically in the United States. While informative, this study raises several questions about the conditionalities of and underlying explanations for this finding. First, because the United States context is characterized by exceptionally high levels of affective partisan polarization, and partisan bias more generally, it is an open question whether partisan discrimination in hiring occurs in less polarized contexts and, if so, to what extent. Second, and relatedly, because the United States is a two-party system, it is unknown how discrimination in hiring based on political affiliations work in multiparty contexts, where there are different, more complex constellations---and likes and dislikes---of parties. Third, the existing study is silent on the mechanisms through which discrimination based on partisanship occurs in the hiring process.

To extend our knowledge of whether and how discrimination in hiring based on partisanship exists in countries outside the United States, this paper leverages an online survey experiment and a field experiment in the United Kingdom and Denmark, respectively. While the United Kingdom is a relatively polarized country compared to other European countries, Denmark is considered a moderately polarized country, therefore, making for a compelling comparison vis-à-vis establishing the potential scope conditions of discrimination in hiring based on partisanship. Furthermore, the fact that the two countries have different party systems provide interesting variation in relation to understanding potential discrimination along partisan lines.

In the first stage of the survey experiment, respondents are asked to rank two fictitious applicants for a number of job vacancies based on their resumes. For each vacancy, the main feature manipulated in the resume is previous experience with voluntary work in a political party (with conditioned (i) mainstream right-wing, (ii) mainstream left-wing party, (iii) far-right party, or (iv) in an apolitical organization). All other features of the candidate (education, skills, gender etc.) are randomized in the resumes and will match the qualifications and the expertise required by the vacancy. In the second stage of the survey experiment, the respondents are asked in an open-ended question to (i) elaborate on the suitability of the fictitious candidates for the vacancy, and (ii) to rate the candidates on several dimensions relevant to social perception (competence, warmth, open-mindedness, and social skills). This, in conjunction with respondents’ own partisanship, allows for an unpacking of the underlying mechanisms, explaining potential discrimination based on partisanship.

In the field experiment, applications are sent to real vacancies (in various industries) and partisan cues are again randomly assigned in the resumes. While the political affiliation of (representatives of) employers are unknown, this can reasonable be approximated by sector, as sectors vary dramatically in average political leanings (indicated, e.g., by union rates).

The complementary qualities of the field experiment and the survey experiment allow us to rigorously study the potential role of partisan bias in an understudied, yet highly important non-political domain (the labor market) in a European multi-party context, as well as unpack the underlying mechanisms.

Author

• Diana Galos, Aarhus University