A new study conducted by the National Bureau of Economic Research (NBER) has concluded that machines are more capable of hiring successful employees, with regards to low skill, service sector jobs such as data entry and call centre positions, than humans.
Are algorithms the future of recruiting?
The study, which covers more than 300k hires, at 15 companies, reveals that employees hired based on the suggestion of a computer algorithm, through means of a job test, are more likely to stay in their positions for longer than those hired by recruiters against the suggestions of the algorithm.
The algorithm used analyses answers given by candidates to a job test, and then places each candidate into one of three groups; green, yellow, and red, based on their appropriateness for the role in question. On average, results show that candidates from the green group stay in their roles for 12 days longer than those from the yellow group, who in turn stay 17 days longer than those from the red group.
On the occasions that recruiters went against the algorithm, hiring candidates based on their own analysis instead, the results were less successful. When opting to hire a candidate from the yellow group, rather than one available from the green group for instance, once said candidate from the green group was hired for a different position, they stayed at their job 8% longer.
Adopting algorithms can eliminate bias
Recruiters can argue that although candidates suggested by the algorithm remain in the job longer, that they may suffer from lower productivity levels than human picked candidates. This doesn’t seem to be the case however, as the study tested productivity levels of hires at 6 of the 15 companies surveyed, and found that the exceptions to the algorithm were no more productive than others.
According to the researchers, although algorithms are slowly gaining ground in the world of recruitment, there are still some preconceived notions, and prejudice, that are affecting widespread adoptability. A fear of the unknown, and a comfort found in the human factor, are restricting the benefits that algorithms could provide to the industry. By hiring candidates through an algorithm, there is a greater chance for a successful hire, and furthermore, you are able to eliminate any potential personal biases that may affect the process when dealing with a human element.
To read the National Bureau of Economic Research’s full study click here.