It’s a familiar stereotype: Asian people are good at math and science. This belief has pervaded American pop culture and media for decades, perhaps best exemplified in a now-infamous 1987 Time magazine cover that showed six young students, sitting behind a computer and books, with the caption “Those Asian American Whiz Kids.”

Since the stereotype ostensibly is a compliment, there’s a temptation to think that pursuing careers in science, technology, math, and engineering is easier for Asian Americans. Indeed, diversity initiatives in these STEM fields tend to focus solely on supporting African Americans and Latinos. But often the opposite is true.  

In recent years, the authors of this article, alongside our colleagues at the Center for WorkLife Law at the University of California’s Hastings College of the Law, have investigated gender and racial biases that pervade STEM professions. Our research has found that Asian Americans, especially women, often face significant career hurdles tied to perceptions about ethnicity and race.  

For one approach, we developed a 10-minute survey that picks up major patterns of racial and gender bias. When we gave an early version to more than 3,000 American engineers, Asian American men and women were much more likely than white men to report that they had to prove themselves more than their colleagues. Most of the 3,000 respondents were women, which makes it hard to draw conclusions about Asian American men. But our data are clear that Asian American women, at least, face the same kind of “prove it again” bias that has been documented for decades in studies of women and black people. Despite being stereotyped as competent, Asian American women still report that they have to provide more evidence of competence than white men in order to be seen as equal.

“If you’re perfect, we might accept you. But if you’re not perfect, forget it,” summarized an Asian American woman in a 2014 study of science professors by our center, with contributions from Katherine W. Phillips of Columbia University and Erika V. Hall of Emory University. We interviewed 60 women of color in the field and documented their experiences with bias. (As scholars, we are ethically bound to protect our interview subjects’ identities as a condition of our research.)

In this study, Asian American women engineers were significantly more likely than white women to report that they were held to higher standards than their colleagues. Taken more broadly, this suggests a pernicious bind: Though Asian Americans might be seen as having a specific set of technical skills, white men with identical skills may be assumed to have a broader range of skills they haven’t demonstrated. So, like women and other people of color, Asian Americans in STEM may have to be more skilled than white men to be seen as equally competent. “Not a whole lot is taken on promise,” said an Asian American astrophysicist. “You have to prove yourself all the time.”

Asian Americans also face bias stemming from assumptions not just about how they do act, but about how they should act. At work, white men generally have more leeway in their behavior: They can shout and scream when they’re angry; they can brag when they’ve accomplished something. For women and people of color, a narrower range of behavior is often accepted. Just as white women are, Asian Americans of all genders who behave in dominant ways tend to be disliked, according to a study by Jennifer Berdahl and Ji-A Min. As The Atlantic’s Olga Khazan writes:

The most notorious double standard is that women can’t break into important jobs unless they advocate for themselves and command respect. But they’re also reviled unless they act like chipper and self-deprecating team players, forever passing the credit along to others. Laurie Rudman, a social psychologist at Rutgers University, said the “poster woman” for this predicament is Hillary Clinton, who, according to surveys, was more popular when in office than when she was vying for office.

Because white Americans typically see Asian Americans as more feminine, Asian American women may well seem even more out of order when they behave in dominant ways. Our study of science professors found that Asian American women report more pressure than any other group to act feminine—and more pushback when they don’t, in the form of negative performance evaluations and assessments that they have personality problems. “[I] immediately started, I guess, having a reputation of being a dragon lady,” said one Asian American biologist. “Early on when I started working here, a faculty colleague of mine said ... ‘I can hear you way down the hall because your balls clang.’”

Our research found that Asian Americans, both male and female, are expected to be good team players and worker bees. Asian American women engineers were significantly more likely than white women—and much more likely than white men—to report the expectation that they work hard, keep their heads down, avoid confrontation, and let others take the lead. This makes it difficult to get ahead as a scientist, a job that demands assertiveness, aggressive lobbying for research opportunities, and leadership of teams of people.

“If you’re a young man, you’re a boy genius,” said a female Asian American astrophysicist. “But if you’re a young woman, you are so threatening that, in order to be able to cope and to be liked ... I have had to become as amiable as possible and a group player all the time.” She felt that she could not promote her accomplishments and thus rarely spoke about the awards and media attention she received for her work.

Another study that our center conducted found that postdoctoral researchers of Asian heritage (including Asian Americans and Asian immigrants) face more bias based on parenthood than any other group. Among 741 postdocs who have children, Asian scientists were twice as likely as white postdocs to report being discouraged from taking maternity leave by their principal investigators. Asian mothers were also more than twice as likely as white postdocs to report that these bosses were not supportive of their new parental status. Perhaps as a result, Asian parents were less likely to request pregnancy accommodations than any other postdocs in the survey. Asian fathers were less likely to request paternity leave than white postdocs as well; and when they did ask, they were less likely to be granted paternity leave than any other group.

All this has consequences. A recent ASCEND report of the Bay Area tech sector found that while Asian people became the largest racial population of professionals between 2007 and 2015, they were the least likely of any racial group to become managers and executives. Our research points to a clear reason why: Asian Americans have been invited into STEM workplaces, but only to play a specific role.

Looking ahead, our center is developing evidence-based, metrics-driven “bias interrupters” to reduce bias through tweaks to hiring, performance evaluations, and other basic business processes. So far, one-shot bias-training programs have mostly been ineffective. Our hope is that bias interrupters can change that.