Machines Beat Humans At Hiring Best Employees

The use of hiring analytics leads to better outcomes for companies, according to a National Bureau of Economic Research study.

Machines make better hiring recommendations than people, at least when it comes to employing low-skilled workers, according to a National Bureau of Economic Research study released this week.

The study, “Discretion in Hiring,” conducted by Mitchell Hoffman from the University of Toronto, Lisa B. Kahn from Yale University, and Danielle Li from Harvard Business School, examined the hiring of almost 300,000 low-skilled service-sector workers, such as call center operators, standardized test graders, and data entry clerks at 15 companies.

By evaluating algorithmic hiring recommendations from a job test alongside human hiring decisions that deviated from the recommendations, the researchers found that hiring managers who exercise discretion produce worse results.

Specifically, employees hired based on algorithmic recommendations had job tenures that lasted 15% longer than people hired without testing. Managers who overruled machine-based recommendations hired workers less well-matched to the job, as measured by job tenure.

The researchers conclude, “[M]anagers systematically make hiring decisions that are not in the interest of the firm.”

(Image: Pixabay)

In other words, hiring managers have biases, something seen in other studies. The study cites a 2012 paper, “Hiring as Cultural Matching: The Case of Elite Professional Service Firms,” that found one of the most important factors in hiring was shared leisure activities.

Such bias might be acceptable if it led to the hiring of employees who were, for example, more productive. But the researchers of this latest paper found “no evidence that firms are trading off [employment] duration for higher productivity.”

They also found that job testing can improve hiring results not only by providing more information, but also by making information verifiable, thereby providing a yardstick by which hiring decisions can be evaluated. They argue that firms should put more weight on hiring analytics “to mitigate errors and biases in human [judgment] across a variety of domains.”


Author: admin


Lost your password?