The idea that the very machines man has created may one day do us in has remained a fixture of pop culture for decades, inspiring countless books, movies and television shows. But it’s not just in the fictional world that humans may be pitted against machines. While many of the most popular, real-life man versus machine stories (Kasparov vs. Deep Blue; Ken Jennings vs. Watson) shows man ultimately triumphing over their mechanical foes, that’s not always the case.
In fact, a study released late last year by the National Bureau of Economic Research suggests that, at least in the case of hiring decisions, machines might have the advantage on their human counterparts. In the study, researchers compared the quality of hiring decisions made by experienced hiring managers and recruiters with those made according the recommendations of a computer algorithm that analyzed the candidates’ results on pre-hire job assessments.
These screening tests asked candidates to answer a number of questions designed to assess their cognitive and technical skills, as well as whether they possessed key personality traits that were considered beneficial for the position. These answers were then run through the algorithm to assess each candidate’s potential fit for the particular job for which they were applying. The algorithm then assigned each candidate a rating based on his or her probability to be a fit for the job: green, yellow or red. “Green” candidates were those considered to have the highest potential to be successful in the position, “yellow” candidates were those with moderate potential, “red” candidates were those the algorithm predicted to be the least successful in the position.
Based on the study’s metric of employee retention, the algorithm’s rating system was proven to be, on the whole, accurate. Candidates with the highest recommendation from the algorithm had longer tenures with the company compared to candidates who received the lowest recommendations.
The study went one step further than proving the algorithm’s accuracy, however. Researchers also compared the hiring decisions made solely based on the algorithm’s recommendations to those made by hiring managers from the participating companies. What did they find out? In cases where hiring managers used their discretion to hire candidates who received lower assessment ratings over a higher-rated candidate, the study showed that the higher-rated candidates (i.e. the one the hiring manager decided not to hire) actually tended to stay with the company about 8 percent longer than the candidates the hiring managers picked.
Could hiring managers and recruiters really one day be replaced by computer programs? Eleesha Martin, senior recruiting specialist for G&A Partners, shared her thoughts on the topic.
“In my experience, pre-hire assessments can be very helpful when it comes to assessing an individual’s related skill set and interpersonal skills, but I’ve also seen them become a hindrance when hiring managers treat the results as deal breakers for qualified individuals who score lower than expected.”
Martin went on to say that the value of the data these assessments provides depends heavily on the type of position for which the recruiter is hiring.
“Certain types of positions may benefit more from the data these tests yield than others. Recruiters looking to fill lower-level positions that don’t require a lot of previous experience may not benefit from having assessment data, but the information that assessments provide about a candidate’s personality and computer skills would be critical when narrowing down the applicant pool for an administrative or customer service role.”
Industry may also play a role how recruiters use assessment data.
“When it comes to IT or Accounting positions, you’d want to make sure to assess candidates’ computer skills, but their personality type might be less important in the final decision-making process as long as the candidate is a cultural fit,” Martin said.
As a recruiting enablement tool, assessments can provide all kinds of relevant data about an applicant: personality type, cognitive reasoning skills, proficiency in job-specific skills, behavior tendencies, personal motivations – the list goes on and on. Personally, Martin believes tests that help a hiring manager build a profile of a candidate are the most useful during the initial screening process.
“Cognitive and skills-based assessments are better used as part of a candidate’s interview. A hiring manager will ask more specific questions about a candidate’s skillset, intellect and ability to perform a job successfully than an assessment is able to measure,” she said.
So what does she think of the study’s findings?
“I don’t think that the metric used in the study [employee retention/tenure] is the most accurate way to measure the quality of a hire. Individual employees are motivated by different factors, and each employee’s job tenure is going to be driven by those factors, which often have little or nothing to do with how they scored on pre-hire assessments,” Martin said.
Martin also thinks the future of the recruiting profession is safe from becoming the next job to be made obsolete by machines.
“The information a hiring manager or recruiter gleans from meeting a candidate in person (cultural fit, demeanor, appearance, communication style, etc.) isn’t something an assessment can tell you. Just like people tend to buy from people they like, managers want to work with people they like. An algorithm can’t feel emotions, and so therefore can’t determine whether a candidate is likeable or will really fit in with the organization.”
Make sure to check out Eleesha’s previous article: “5 Ways You’re Sabotaging Your LinkedIn Profile.” Click here to view the article.
This article is not intended to be exhaustive nor should any discussion or opinions be construed as legal advice. Readers should contact legal counsel for legal advice.