Tiny amounts of gender bias in employee hiring decisions contribute to concerning rates of discrimination and productivity losses that together represent significant costs, financial and otherwise, for employers, a new study from Oregon State University has found.
Gender bias is a subtle, unintentional preference for one gender over the other. Despite significant efforts to reduce bias in hiring over the last several decades, it continues to persist and pose potential problems for companies, said Jay Hardy, an assistant professor of management in OSU’s College of Business and lead author of the study.
“The research has long shown bias exists. This study shows that it matters,” Hardy said. “When carried through the hiring decision-making process, tiny amounts of bias will balloon into a high probability of discriminatory hiring outcomes in the eyes of the law, which also has important financial consequences for companies that end up hiring less-qualified candidates.
“The message of this study is that you can’t ignore gender bias, even if you think its influence is so small as to not be concerning. Society generally recognizes bias as a moral issue, but we are now learning the extent to which it is a financial and strategic issue, as well.”
The findings were published in the Journal of Management. Co-authors on the paper are Richard Martell and Andy Olstad of Oregon State and Kian Siong Tey, Wilson Cyrus-Lai and Eric Luis Uhlmann of INSEAD, one of the world’s largest graduate business schools, with locations in Europe, Asia, the Middle East and North America.
Hardy’s research focuses on areas of human resources, including employee recruitment and selection and systematic bias in hiring. His goal for the study was to better understand the effects gender bias in hiring might have for employers.
“Bias is a problem ingrained in us. It is systematic,” Hardy said. “It’s not typically about bad intent. It is a natural response of human beings to make sense of our complex world by taking mental shortcuts. However, these cognitive shortcuts can have big consequences for us when high stakes rewards like a sought-after job are on the line. Our goal is to better understand it so we can find solutions.”
Hardy and his colleagues first reviewed about 30 years of studies on gender bias and hiring in the workplace. The good news is that there appears to be less gender bias effect than there used to be. A couple of decades ago, gender bias influenced about 4 percent of hires. Today, bias influences 1 percent or possibly fewer hires, Hardy said.
“The science shows that the effects of gender bias on hiring are very small,” Hardy said. “In broad strokes, it seems like hiring decisions are actually pretty fair — certainly more fair than they used to be.”
To better understand the impact this bias has and what it means for companies, Hardy and his colleagues ran a series of computer simulations. The researchers found that even a small amount of bias in hiring decisions can lead to discriminatory action against job candidates, putting companies at risk of costly legal action. They also found that biased hiring decisions can be costly for companies because a less-qualified candidate may not be successful in the position.
For example, a typical Fortune 500 company that hires 8,000 new employees a year with a 1 percent gender bias effect can expect 32 additional failed hires and many more sub-optimal hiring decisions, resulting in productivity losses of about $2.8 million per year. A 4 percent bias effect would lead to an additional 192 failed hires and an additional $17 million in lost productivity.
The study focused specifically on gender, but Hardy said it is likely that other types of bias would have similar or even bigger impacts. Past research has demonstrated that racial bias effects in hiring are much larger than they are for gender, so it is reasonable to think there would be more discrimination and more costs for employers as a result, he said.
“My estimates for gender bias were conservative and my simulations modeled near-optimal hiring conditions, so the impact of gender bias is likely to be higher in many real-world hiring contexts than my study indicates,” Hardy said.
Hardy and his colleagues also simulated the impacts of some common methods for reducing gender bias in hiring, such as targeted recruiting efforts to build a bigger pool of female candidates, and found those methods still pose challenges.
“The targeted recruitment of highly qualified candidates can increase representation, but if you’re not fixing the underlying bias problem, these methods do not address discrimination and can lead to other issues, including employee dissatisfaction,” Hardy said.
Companies looking to reduce gender bias in their hiring processes should look for ways to remove human judgments from the process as much as possible, Hardy said. Interviews are one of the most common but least objective hiring tools, he noted. Future work by Hardy will further explore how to address bias in hiring and other employment decisions.
“Hiring is always going to be an imperfect process because human beings are complicated,” Hardy said. “But the gold standard for any hiring manager should be to be as objective as possible in the hiring process.”
