T**his article was first published in the June 2016 issue of WIRED magazine. Be the first to read WIRED's articles in print before they're posted online, and get your hands on loads of additional content by subscribing online.
Blind evaluations are en vogue. Compose Inc, a cloud-storage firm in California, no longer asks job applicants for their CVs but instead has them complete tasks relevant to the job. Across the board, such tests have been found to be significantly better at predicting future performance than traditional job interviews.
In fact, unstructured interviews are particularly bad at predicting a person's future performance. Research by sociologist Lauren Rivera suggests that interviewers tend to want to replicate themselves, or as one recruiter said: "You use yourself to measure because that's all you have to go on."
Analysing hiring in investment banks, law firms and management consulting firms, Rivera found that the degree to which a candidate's backgrounds, hobbies and self-presentation was similar to a company's existing employee base was decisive in candidate evaluations.
Dozens of studies show that unstructured interviews primarily add noise because human minds are incapable of separating relevant data from irrelevant data. A few years ago, an opportunity to better understand this presented itself in Texas when the state realised that it did not have enough physicians. It asked the University of Texas Medical School in Houston to dramatically increase the number of admitted students late in the academic year.
At that time, only students who had not been admitted anywhere were left. But what seemed like a catastrophe to the university turned out not to matter at all. The bottom-ranked students performed just as well as the top-ranked ones - leading the researchers to question whether interviews should be replaced by a lottery among applicants.
Looks also matter. We tend to prefer the good-looking, and what is more, expect them to be better performers and more worthy of our trust. But attractive people are as trustworthy as the average person. However, when the good-looking do not live up to those expectations, they are punished. Generally, we do not like people who violate our beliefs - whether related to how we expect attractive people to behave or an engineer or a nurse to look.
This is why blind evaluations have played an important role in auditions. The Boston Symphony Orchestra, and later most of the major orchestras in the US, asked musicians to audition behind a curtain. Although in theory orchestra directors should not care about who plays the bassoon but just about the music that is produced, it turns out they do, often without their conscious knowledge.
In the 70s, only five per cent of musicians in the top five orchestras in the US were female. Today, women make up almost 40 per cent. Research by economists Claudia Goldin and Cecilia Rouse shows that the introduction of screens increased the likelihood that women would advance to future rounds by around 50 per cent and therefore increased the fraction of female hires.
Adopting what worked for orchestras is not always possible. But at a minimum, a practice still prevalent in many European countries and Israel that encourages applicants to include head shots with their CVs should be stopped. Research from Israel demonstrated that employers were more likely to call back attractive men than average-looking male applicants.
Attractive female applicants did not reap a beauty premium. Indeed, in western countries, looks appear to matter in particular for men - even though evidence does not suggest that beauty is a credible marker for underlying characteristics such as intelligence.
But you can do more. To guard against irrelevant factors clouding your judgement, evaluate job candidates comparatively. In a recent study with Max Bazerman and Alexandra van Geen, we found that evaluators who use a comparative perspective are much more likely to focus on relevant predictors of future performance, instead of less helpful information, such as a person's gender. Indeed, our minds have an intrinsic need to rely on comparisons to understand what is going on.
Whether or not you are drawn to a particular job candidate is related to the people you are used to seeing in this job. This is how stereotypes such as engineer equals man or nurse equals woman take hold in our minds. We then use them as measuring rods when evaluating others. To debunk these internal comparators, force yourself to compare a given candidate with real alternatives.
Comparative evaluation is the right and smart thing to do. Many companies have begun to take encouraging steps in that direction. Google, for example, compares a given evaluator's assessments with those of others and finds that such calibration diminishes bias. But this is not limited to large tech giants. You can do it too. Tools are available to assist employers in evaluating talent more objectively. Check out Applied, GapJumpers or Unitive recruitment platforms that anonymise applicants, introduce work sample tests and help you with comparisons and aggregation across evaluators. Instead of picking those who look the part, you can find the best talent and level the playing field.
Iris Bohnet is the author ofWhat Works: Gender Equality by Design(Harvard University Press)
This article was originally published by WIRED UK
