The Unintended Effects of Recruitment Technology: Lessons from Asynchronous Interviews

Abstract

Labor is one of the most important parts of a company’s production function and hiring mistakes are expensive. Thus, employers want to get more information about applicants but are limited by cost. Recent technological advances have produced a range of application assessment tools with very low marginal costs, leading to a proliferation of additional screening stages during the application process. However, these additional stages also change the experience of applicants and recruiters in ways that may change who applies and how they are evaluated. In this paper we study how asynchronous interviews impact who applies and who could get hired using a natural field experiment with over 3,000 applicants for 3 different jobs. Asynchronous interviews deter 53% of applicants from continuing with the application process. Due to differential deterrence, the final applicant pool is more male and less qualified. We also find large differences in evaluations of the interviews by gender, race, and qualification. Combining application decisions and evaluations, asynchronous interviews substantially change the composition of who would be considered for the position: the shortlist becomes more male and racial minority and less educated, experienced, and Asian. These findings indicate that while novel recruitment technologies can vastly cheapen the process of acquiring information about applicants, they also substantially change who applies and who is considered for hiring in unintended ways.