Many correspondence audit studies on labour market discrimination have recently been published. My research with Siel Vermeiren and Stijn Baert, which appeared in the January 2023 issue of European Economic Review, synthesises the data from nearly all correspondence audit tests conducted around the world between 2005 and 2020. It thus gives a bird’s-eye view of the extent of hiring discrimination on various grounds of discrimination. How do, for example, age discrimination and ethnic discrimination compare? Are there regional differences? And does hiring discrimination decrease over time?
Recent scientific research on hiring discrimination relies for a large part on the tried-and-true method of the correspondence experiment. This method typically involves sending out pairs of fictitious resumes to actual job postings, randomly assigning a particular minority characteristic to either candidate. The responses of employers or recruiters to these resumes are then compared to estimate hiring discrimination.
Such research has been conducted worldwide and on numerous grounds of discrimination in recent years. More than 900,000 fictitious job applications have been carried out between 2005 and 2020 and used in analyses in international scholarly publications. However, each correspondence experiment captures only part of the hiring discrimination in a given labour market context.
Therefore, we used a meta-analytical approach to synthesise the available data from these field tests. In this meta-analysis, we calculated weighted averages of the discrimination statistics that consider, among other things, the sample size of the individual correspondence tests and variance around the estimates. We discovered a few remarkable differences between discrimination grounds, across regions, and over time.
Physical attractiveness pays off in the labour market
First and foremost, we find that hiring discrimination based on (i) employment disability (ca. 41% fewer positive responses than those without disabilities), (ii) low physical attractiveness (ca. 37% fewer positive responses), and (iii) older age (ca. 34% fewer positive responses) is at least as problematic – if not more problematic – as hiring discrimination based on ethnicity (ca. 29% fewer positive responses), which is also the most studied discrimination ground.
Furthermore, there is no overall evidence of gender discrimination in hiring. In contrast, significant differences exist between correspondence tests; in some cases, hiring discrimination against women is detected, while in other cases, it is men who are discriminated against. Very recent large-scale U.S. research confirms this discrepancy – occupational gender composition differences presumably drive this phenomenon.
Moreover, across correspondence tests conducted worldwide, candidates who are open about their LGB orientation receive, on average, about 30% fewer positive responses to their job applications.
It is important to note that there are many differences within discrimination grounds and between subgroups. Considering ethnic discrimination, for example, applicants with a Maghreb, Arab or Middle Eastern background receive, on average, about 41% fewer positive responses than applicants with no migration background. At the same time, there is hardly any hiring discrimination against applicants from Northern or Western European minority groups.
Europe versus the United States
A breakdown by region reveals that hiring discrimination based on age differs substantially between Europe and the United States. In European correspondence tests, older applicants (typically around 50 years old) receive, on average, about 48% fewer positive responses than younger applicants (typically around 30 years old). This percentage is much lower in the United States: about 31% fewer positive responses. Interestingly, the ages and age differences used in the American correspondence tests were much higher than in the European experiments.
Still, this result is consistent with the employment rates of older workers in the respective regions. Earlier research pointed to regional differences in legal and social norms regarding working at older ages. These norms tend to be more lenient in the United States and more focused on encouraging older workers to stay in the workforce than in many European countries.
Is discrimination declining globally?
Finally, discrimination rates have barely improved over time. This finding is quite remarkable given the anti-discrimination policies in many countries and the attention paid to the problem, even at the supranational level. This finding aligns with earlier meta-analyses.
The one bright spot the study uncovered, namely that ethnic discrimination would have declined in European correspondence experiments, seems mainly due to trends in minority groups studied. In the 2015-2020 period, applicants of European and West Asian descent received proportionately more attention, and applicants of Maghreb-Arab or Middle Eastern descent received less attention in correspondence tests than in the 2005-2014 period. At the same time, we know that the latter group is more strongly disadvantaged in the application process.
The meta-analysis allows policymakers and employers to target those minority groups most disadvantaged in the selection process specifically. The insights from the research should also raise awareness of the issue; the structural nature of hiring discrimination cannot be questioned based on this extensive analysis.
Crucially, diversity promotion measures, such as information campaigns and diversity training, should focus not only on ethnicity and gender but also on, say, age. In other words, the meta-analysis underscores that diversity in the labour market must also be given a diverse interpretation.
This post also appeared via UGent @ Work and in various Belgian media, in Dutch. This page was last updated on 29 November 2023.