Tastes, expectations and gendered jobs: a field experiment with pink-collar men

About this event

Taste-based and statistical discrimination theories have been extensively used to explain demand-side determinants of occupational segregation by gender and race. However, evidence on the way these forces affect the supply-side is still limited. To what extent do preferences for co-workers gender versus performance expectations affect minority's decision to apply for a job? To answer this question, I designed and implemented a nationwide recruitment experiment in collaboration with one of the main UK organizations in social care. Men's share in this job has historically been below 25%. I use random variation in emails to potential applicants to manipulate two dimensions: 1) salience of workers' gender (through a male or female photograph) and 2) performance expectations (describing the aggregate performance of two previous cohorts of workers). I find no evidence for a taste-based explanation of men's application decisions. Disclosing information that past performance was poor encourages 14% more men's applications than information of an outstanding past performance. These effects are stronger for men with limited experience in social care. Despite this initial quantity/quality trade-off, the treatment which triggers more applications also attracts men that perform better in the first month on the job. Women apply more when seeing the female than the male photograph and are insensitive to information on average, but this masks heterogeneity depending on the photograph they saw.