How recruiters discriminate on employment websites

Researchers funded by the Swiss National Science Foundation conducted a large-scale study of discrimination on an online recruitment platform.

Discrimination in hiring is a major societal problem. A new analysis of extensive data from an online recruitment platform has enabled researchers to determine why and to what extent the ethnic origin of a job seeker and their gender influence the chances of finding a job. The findings are from a study funded by the Swiss National Science Foundation (SNSF) and published in the prestigious scientific journal Nature (*).

Daniel Kopp and Michael Siegenthaler, economists at the Swiss Institute for Business Cycle Research at ETH Zurich, together with political scientist Dominik Hangartner, analysed not only which people were invited for an interview, but also how they were selected.

Stereotypes have a greater effect in the evening

The researchers showed that on average, foreigners were 6.5 per cent less likely than Swiss nationals to be contacted by recruiters for an interview. “This discrimination was particularly pronounced among immigrants from the Balkans, Africa, the Middle East and Asia, who often have to battle prejudice”, says Daniel Kopp.

The research team also found that the negative impact of foreign origin on jobseekers was greater around noon and late in the day – when recruiters spend less time evaluating CVs. “This pattern supports the hypothesis that unconscious discrimination plays a certain role”, says Kopp.

In contrast, the study found on average no discrimination on the basis of gender. However, there are huge differences across occupations. Women are discriminated against in typical male professions (7 per cent lower probability of being contacted) and men in professions dominated by women (13 per cent lower probability of being contacted). According to Kopp, these results suggest that certain recruiters still adhere to traditional gender roles.

Three million cases

In order to estimate labour market discrimination, the researchers analysed more than three million decisions made by recruiters over 10 months. Rather than applying the familiar yet controversial method of submitting fictitious CVs for vacant positions, the researchers cooperated with the State Secretariat for Economic Affairs (SECO) to gain access to anonymised data from Job-Room, one of the largest recruitment platforms in Switzerland. The researchers were thus able to peer over recruiters' shoulders when selecting candidates. “The wealth of data allowed us to study the preferences of recruiters for different job seeker characteristics and in many different professions. Conventional methods are much more limited in this respect”, says Kopp.

The results of the new study are not necessarily representative for all job seekers in Switzerland. For instance, management occupations are underrepresented. Nevertheless, based on the findings of the study, it is possible to formulate recommendations for re-designing online recruitment platforms to increase equal hiring opportunities. For example, experience and skills could be prioritised over characteristics such as country of origin and gender, which are irrelevant for work performance.

The research was supported by the SNSF through project funding. This funding scheme enables researchers to carry out research projects on self-chosen topics and with self-chosen goals independently and under their own responsibility.

Contact by mouse click

The online recruitment platform Job-Room contains profiles of more than 150,000 job seekers. Recruiters specify the criteria required for a particular job. They then receive a list of suitable candidates and information about them. If they are interested in a particular candidate, recruiters can open the profile to learn more about the person and eventually click on a contact button to access the detailed contact information and invite them for an interview. In their study, the researchers used these clicks to assess the role of gender and ethnic origin in explaining the likelihood of contact. In the case of ethnic origin, researchers distinguished between “typically” Swiss jobseekers (Swiss-sounding name, Swiss nationality and command of at least one national language) and “typically” foreign jobseekers (foreign-sounding name, foreign nationality and command of at least one language of the country of origin).

The analyses are based on data from over 450,000 searches by recruiters between March and December 2017. Recruiters were directed to 16.9 million candidate profiles, of which they opened 3.4 million and contacted some.

(*) D. Hangartner, D. Kopp and M. Siegenthaler: Monitoring hiring discrimination through online recruitment platforms. Nature (2021). Link Icon
(Available to journalists in PDF format from the SNSF: 
com@snf.chExternal Link Icon)


Daniel Kopp
ETH Zurich
Tel.: +41 44 633 22 22

kopp@kof.ethz.chExternal Link Icon