The study looked at data from Swiss a government-affiliated online recruitment platform, used by the public employment service. Recruiters using the platform were 4% – 19% less likely to follow up with equally qualified job seekers with foreign sounding names.
Discrimination was particularly pronounced among immigrants from Africa, the Middle East, the Balkans and Asia.
Researchers applied machine learning algorithms to 452,729 searches by 43,352 recruiters and analysed them based on time spent looking at each profile and whether contact was made. The data were not representative of all job seekers in Switzerland. For example, management occupations were underrepresented.
The study also revealed gender bias, something that cut both ways. Women were 7% less likely to be contacted for roles in professions dominated by men, and men were 13% less likely to be contacted for positions in fields dominated by women, a finding that suggests some recruiters are still guided by traditional gender roles
Recruiters appear to be more likely to default to discriminatory stereotypes at particular points in the day. Before lunch (11am – 12 noon) and towards the end of the day (5pm – 6pm) recruiters spent less time looking at CVs and this was associated with ethnic minorities faring 20% worse than at other times of the day.
This finding mirrors research done in 2016 on the decisions of Israeli parole judges. They found that the probability of a favourable decision dropped from about 65% to almost 0% from the first ruling to the last ruling within each session, and that the rate of favourable rulings returned to 65% in a session following a food break.
Hangartner said that we are optimistic that at least part of the discrimination that we document in this study can be overcome by re-designing recruitment platforms. For example, more relevant information such as a candidate’s work experience and education could be placed at the top, and details which might indicate ethnicity or gender, such as name or nationality, could appear much lower down the CV.
However, the research also found that there were only small differences in the time spent viewing the profiles of individuals from minority ethnic groups compared to those from the rest of the population, a finding that suggests ethnicity is not being used as a shortcut to screen out applicants. Instead it seems more likely it is being applied at the end of the process.
Research report (in English)