Managers looking to mitigate gender bias when hiring and promoting might want to consider evaluating candidates in groups.
Researchers from Harvard University found that managers are much less likely to stereotype employees by gender during an evaluation process if done in a group setting as opposed to one-on-one.
Because group settings provide managers with a total picture of each individual’s qualifications side by side, gender bias is less likely to affect the decision-making that leads to a hire or promotion, said Iris Bohnet, a professor and academic dean of the Harvard Kennedy School (HKS) of Government and one of the study’s co-authors.
The findings were originally published in a March 2012 paper titled “When Performance Trumps Gender Bias: Joint Versus Separate Evaluation.” The study was also co-authored by HKS doctoral student Alexandra van Geen and Harvard Business School professor Max H. Bazerman.
The authors’ conclusions came as a result of a two-stage experiment at the Harvard Decision Science Laboratory. About 550 male and female participants played the role of an employer, while 100 played the role of employee.
The first stage included those classified as “employees” completing two rounds of either a math or verbal task — activities highly associated with gender bias, because “females are believed to be worse at math tasks and better at verbal tasks than males,” the study said.
They were then paid based on their performance. Employees also completed a demographic questionnaire that included their gender.
Those in the employers group were then asked to choose candidates qualified to move on to a second round of tests. In some instances, employers were presented with information about a single candidate.
Here the employers were more likely to choose women for additional verbal tasks and men for math tasks. This was the case even for those whose first-round performances had been considered weak.
When employers were presented with information about a male and female candidate together, however, past performance — not gender — were the primary determinant in which additional tasks employees would be asked to do.
In other words, in the former test gender played a role in figuring out for which positions men and women were most qualified.
Bohnet likened a possible explanation to shopping for different products. When evaluating a product individually, it is more difficult to compare without preconceived judgments than when you’re able to look at two products simultaneously.
“I think that the host of evidence suggests that looking at things in bundles rather than one at a time leads to more informed judgments and decision-making,” she said.
Bohnet said it’s common for companies to use group evaluation in the hiring process; it is less common, however, in cases of promotion or in assigning job assignments.
She said the researchers are now in the processes of designing similar behavioral interventions for testing with companies in the field, where certain variables cannot be controlled as closely as in a lab environment.
Frank Kalman is an associate editor at Diversity Executive. He can be reached at firstname.lastname@example.org.