Back to All Events

Transparency and Trust in Algorithmic Hiring Procedures

  • Amsterdam Leadership Lab 7 Van der Boechorststraat Amsterdam, NH, 1081 BT Netherlands (map)

Susan Niessen

We present four experimental studies aimed at investigating the effect of transparency on perceptions of mechanical hiring procedures. It is well established that combining information mechanically (via a pre-defined decision-rule) results in better hiring decisions than holistic combination. Nevertheless, mechanical combination is perceived negatively and rarely used in practice.  Inspired by research conducted on AI-based hiring tools, we investigated the effect of various operationalizations of transparency in mechanical procedures on the perceptions of two important stakeholder groups: decision-makers and applicants. For each group, we used between- and a within subject’s design. Moreover, we explored the relative importance of several perception constructs in relation to trust and use intentions, and supplemented quantitative analyses with thematic analysis of open-ended responses. In the between-subjects studies, we mostly found near-zero differences. However, the open-ended responses indicated a strong emphasis on ‘filler information’. Within-subjects replications conducted to solve this issue did show the expected positive effects of transparency on perceptions, but mostly with small effect sizes. However, both groups indicated a strong preference for transparent algorithms over holistic judgment when asked to choose. Moreover, the open question responses indicated that both groups generally perceived algorithms positively regarding fairness, but negatively or mixed in terms of validity.