Skip to content

Do photos help or hinder field experiments of discrimination?

Research output: Contribution to journalArticle

This paper assesses field experiments of labour and product markets that have attached photos to identify applicants (in the case of labour markets) or sellers/crowdfunders (in the case of product markets). The experiments seek to identify the contribution of attractiveness, race/ethnicity, skin colour, sexual orientation or religion to the behaviour of agents in markets. These experiments attach photos to CV to signal attractiveness, or the basis being tested such as race/ethnicity, skin colour or religion. Many experiments report significant findings for the impact of attractiveness or the identity revealed on positive callbacks to applicants. The issue considered here, however, is to what extent it is attractiveness or other perceived characteristics that may have had an impact on the behaviour recorded in the experiments. The results of the studies covered in this paper, to a lesser extent those of Weichselbaumer (2004) and Baert 2017, are compromised by including photos, with the possibility the responses received were influenced not only by the basis being tested such as attractiveness, race/ethnicity or religion but by some other characteristic unintended by the researcher but conveyed by the photo. There is evidence in experimental work of a range of characteristics that photos convey of individuals and their impact on labour and product market outcomes such as success in obtaining a positive response to job applications and success in obtaining funding to finance projects in the product market. Suggestions are made for future experiments: evaluation of photos for a range of characteristics; use of a ‘no photo’ application together with the photo applications; evaluation of responses for any bias from unobservable characteristics using Neumark (2012). This paper discusses for the first time three question with some tentative answers. First we face introducing further unobservable characteristics by using photos. Second, we cannot fully control the experimental approach when using photos. Third, we are able to accurately evaluate the impact of the photos used on the response/probability of callback. Field experiments using photos need to ensure they do this for the range of factors that have been shown to affect judgments and therefore potentially influence call back response. However, the issue remains whether we have in fact identified all potential characteristics conveyed by the photos.
Original languageEnglish
Pages (from-to)502-518
JournalInternational Journal of Manpower
Volume39
Issue number4
Early online date13 Jul 2018
DOIs
Publication statusPublished - Aug 2018

Documents

Related information

Relations Get citation (various referencing formats)

ID: 10044171