Here is an excerpt from an article written by Joan C. Williams for Harvard Business Review and the HBR Blog Network. I wholly agree with her: “If organizations are truly interested in retaining and advancing women, they will approach the issue of gender bias the same way they do other business issue: develop objective metrics and hold themselves to meeting them.” To read the complete article, check out the wealth of free resources, obtain subscription information, and receive HBR email alerts, please click here.

* * *

By now, we’ve all heard about the low numbers of American women in science, technology, engineering, and math (STEM). Some argue it’s a pipeline issue – that if we can interest more young girls in STEM subjects, the issue will resolve itself over time. But that’s not convincing. After all, the percentage of women in computer science has actually decreased since 1991.

Another theory is that women are choosing to forgo careers in STEM to attain better work-family balance—rather than being pushed out by bias. But evidence for that is also thin. Several new studies add to the growing body of evidence that documents the role of gender bias in driving women out of science careers. A 2012 randomized, double-blind study gave science faculty at research-intensive universities the application materials of a fictitious student randomly assigned a male or female name, and found that both male and female faculty rated the male applicant as significantly more competent and hirable than the woman with identical application materials. A 2014 study found that both men and women were twice as likely to hire a man for a job that required math.

My own new research, co-authored with Kathrine W. Phillips and Erika V. Hall, also indicates that bias, not pipeline issues or personal choices, pushes women out of science – and that bias plays out differently depending on a woman’s race or ethnicity.

We conducted in-depth interviews with 60 female scientists and surveyed 557 female scientists, both with help from the Association for Women in Science. These studies provide an important picture of how gender bias plays out in everyday workplace interactions. My previous research has shown that there are four major patterns of bias women face at work. This new study emphasizes that women of color experience these to different degrees, and in different ways. Black women also face a fifth type of bias.

[Here is the first of five patterns discussed.]

Pattern 1: Prove-it-Again. Two-thirds of the women interviewed, and two-thirds of the women surveyed, reported having to prove themselves over and over again – their successes discounted, their expertise questioned. “People just assume you’re not going to be able to cut it,” a statistician told us, in a typical comment. Black women were considerably more likely than other women to report having to deal with this type of bias; three-fourths of black women did. (And few Asian-American women felt that the stereotype of Asian-Americans as good at science helped them; that stereotype may well chiefly benefit Asian-American men.)

Experimental social psychologists have documented this type of bias over and over again in college labs, but this is the first time someone has taken that experimental literature and asked women whether it describes their experience in actual workplaces. It does.

* * *

Here is a direct link to the complete article.

Described as having “something approaching rock star status” by The New York Times, Joan C. Williams is a Distinguished Professor of Law, UC Hastings Foundation Chair, and the Founding Director of the Center for WorkLife Law at UC Hastings College of the Law. She is a graduate of Harvard Law School/Massachusetts Institute of Technology, J.D., and Master’s Degree in City Planning (1980); Yale University, B.A., History (1974); and Princeton Day School (1970). To check out all of her HBR articles, please click here.

I highly recommend her latest book, What Works for Women at Work: Four Patterns Working Women Need to Know, co-authored with Rachel Dempsey and published by NYU Press (January 2014).