The Harvard Kennedy School’s Women and Public Policy Program Gender Action Portal is a great website that aggregates academic research. One of the publications I found there could help managers and HR teams avoid unconscious gender bias and stereotyping that would result in bad hiring decisions. The take-away of this experiment: when you evaluate a candidate for a task/position, you’ll be more objective if you compare that person to another candidate.
The Harvard study involved two groups of people. One group, (“Evaluators”) assessed another group, (“Subjects”) based on the following information: gender, past performance, and a description of a “gendered task” (e.g. a math task or a verbal task). The assumption in this experiment was that Evaluators would hold common stereotypes around male and female abilities regarding math and verbal tasks, i.e. men are better at math tasks and women are better at verbal tasks.
When Evaluators assessed a Subject in isolation (i.e. without comparison or information on any other Subject), they weighed gender more heavily than past performance. For a verbal job, they were more likely to assess a woman favorably compared to a man, even if that woman had low, prior scores on verbal tasks. Similarly, Evaluators were more likely to positively view a man for a math job, even if he had low, prior math scores. To be clear, the Subjects’ history of low scores were not dramatically low, i.e. they were slightly below-average results. In other words, Evaluators were not considering anyone who looked like they were “failing” prior tasks. The experiment was designed this way in order to be more realistic as the real world often requires making “close calls”.
Gender bias seemed to disappear when Evaluators were asked to assess two Subjects at a time. If an Evaluator were given the same 3 pieces of information (i.e. gender, past performance and either a verbal/math task description) about two Subjects and asked to assess both, the overall assessment was predominantly based on past performance rather than gender. Regardless of whether a Subject was a man or a woman, the Evaluator primarily made their assessment of a Subject based on prior score histories! This pretty dramatic result implies that in addition to gender “blind” resume selection, it might be best to hire someone by comparing that person, head-on, with someone else.