Attention Discrimination: Theory and Field Experiments

We link two important ideas: attention is scarce and a lack of information about an individual drives discrimination in selection decisions. We model how knowledge of ethnicity influences allocation of attention to available information about an applicant. When only a small share of applicants is accepted, negative stereotypes are predicted to lower attention, while the effect is opposite when most applicants are accepted. We test for such “attention discrimination” in two field experiments. We send emails responding to job offers and apartment-rental advertisements and monitor information acquisition, a new feature in this type of experiment. We vary the names of applicants to signal ethnicity and find that minority names are about half as likely to receive an invitation for an apartment viewing or a job interview. The novel finding is that minority names affect the likelihood of resumes being read on the labor market as well as an applicant’s personal website being inspected on the housing market, but the effects are opposite across the two markets. These results support the model's assumption of endogenous attention, which magnifies the role of prior beliefs in discrimination. The model implies persistence of discrimination in selection decisions, even if information about individuals is available and there are no differences in preferences, lower returns to employment qualifications for negatively stereotyped groups, and for policy, the important role of the timing of when a group attribute is revealed.