This article was originally published on the Nature Soapbox Science blog. It was written by STEM Women managers Zuleyka Zevallos, Buddhini Samarasinghe and Rajini Rao.

In a New England pub after a conference, our male academic colleagues shrug their collective shoulders at the gender imbalance; in their opinion, women drop out of science because their hormones make them “different”. As women in science know all too well, similar examples of bias abound in academia. We read with familiar dismay, therefore, the arguments that girls find science “boring,” that attempts to bridge the gender divide “deny human biology and nature,” and that efforts to achieve gender equality in the Science, Technology, Engineering and Mathematics (STEM) fields are doomed. Attributing the gender gap to biology misses the obvious contribution of societal and institutional biases.

The “girls are not interested in STEM” mantra is itself an example. Knowledge of a prejudicial stereotype can lead to enough anxiety that it becomes a self-fulfilling prophecy. Although initially applied to racial bias in IQ tests, ‘stereotype threat’ can be extended to gender as well. Negative stereotypes are transmitted from parents and teachers to girls. Reminding girls that they are girls just before a math test can impede their performance. This effect can be seen in children as young as age five. The key point is that as adults, we are able to view stereotypes as generalisations about a group. Unfortunately, young children are more accepting of stereotypes, and may implicitly believe that girls are indeed poor at STEM subjects. As girls grow up, these stereotypes affect their identities as STEM professionals. So how do these stereotypes take shape?

The Draw a Scientist Test (DAST) shows that science stereotypes are socialised from an early age. The test has its origins in a pilot study from 1957, in which high school students were asked to describe their image of a scientist. Subsequent research from the 1960s onwards has examined a timeline of when this image is cemented. This research finds that children in kindergarten and the first grade are less likely to draw a stereotypical scientist; that is, a White man with facial hair, who wears a lab coat and glasses, and who is surrounded by lab equipment, formulae and books, making a “Eureka!” style exclamation. By the second grade, however, “the stereotype has began to take root,” due to a combination of how science is taught at school as well as through media images and social ideas and expectations that children pick up from parents, teachers and other influences. By the fifth grade, the stereotype is overwhelmingly fixed. A review study of 50 years worth of international research on the DAST demonstrates that, across cultures, “the stereotype of scientists being male has largely endured since 1957.” More specifically, the research shows that in Western cultures, this image is of a White male, even amongst minority students. However, children exposed to female scientists, via a combination of visits by women scientists in the classroom and by talks and readings about women’s contribution to science and their careers, are more likely to draw both women and men as scientists. Studies demonstrate that teachers themselves can contribute to these stereotypes, by giving boys more attention in class, and by rating their abilities higher than girls, even when girls get the same test scores as boys. Conversely, “intervention programs” for teachers, including career information and weekly visits by women researchers to the classroom, as well as short courses with follow-up visits, are subsequently less likely to result in stereotypical DAST results amongst students. This is because the teacher’s bias has been actively addressed, leading to the reinforcement of diversity in their teaching.

Despite the fact that people are socialised into believing that girls can’t do science, popular culture blames the individual; young girls are often chastised Don’t get your dress dirty, or Be careful, why don’t you hand that to your brother, as a recent viral video reminded us. If only girls were more confident. If only girls spoke up and asked more questions in class. If only they actively looked for mentors. Then they could easily overcome this stereotype threat and perform just as well as boys in STEM subjects. Unfortunately this “leaning in” viewpoint is naive because it ignores the institutional disadvantages contributing to the academic exclusion of women and minorities.

Sexual harassment is widespread in academic fieldwork. Women trainees are the primary targets with the perpetrators being predominantly senior professional males. Female undergraduates in male dominated fields report higher levels of sex discrimination, and are more likely to consider changing majors. Another study showed that high-achieving male biologists train fewer women than men in their laboratories, and that these men predominantly fill Assistant Professor slots in academia. In the same study, biomedical science male postdocs are 90% more likely than women to have an adviser who is a Nobel laureate. Not only is there a ‘leaky pipeline’ problem, the plumbing itself is broken.

It is social conditioning, unconscious biases and institutional practices that create an environment where girls feel unwelcome and insecure in STEM fields. UNESCO data show that women are disadvantaged in STEM, with only one in five nations achieving equality. But the cultural variation in itself tells us that it is socialisation and policy intervention, not biology, that matters. Research shows that institutional gender bias develops in several phases. First, children lack female scientist role models from primary school. Second, young undergraduates learn that science privileges a masculine culture, which makes it hard to imagine their career path. Third, diversity barriers are witnessed first-hand by early career researchers. Both male and female faculty are less willing to hire women applicants with the same credentials as men. Given these clear prejudices, we must move away from lazy explanations that attribute women’s under-representation in STEM to their biology. Instead, we must acknowledge that the system actively discourages women in ways both obvious and insidious. We must move away from the individual and address the broader narrative of everyday sexism.

Practical ways to tackle this problem include diversity training for hiring committees and better mentorship programmes for female graduate students and postdocs. Another avenue for change is to address stereotypes and their effects. Research mapping neurophysiology during tests on STEM subjects show that there are no cognitive differences in men and women’s performance in tests until stereotype threat is triggered. Women perform comparably well until they are reminded about their gender, at which point their working memory and performance are negatively impacted. When girls and women are made aware of their minority status, they become hypervigilant about negative feedback, discouraging them despite their success, even if they are high achievers. Professor Chad Forbes is a social neuroscientist from the University of Delaware who studies the impact of negative stereotypes on individuals. One aspect of his research is looking at different ways to combat stereotype threat. The most effective strategy remains acknowledging and understanding the existence of stereotype threat and addressing its consequences, such as through training. Active intervention at the institutional level also leads to positive change. Already, some colleges are reporting huge improvements: at Carnegie Mellon University, 40% of undergraduate incoming class in computer science are women, a welcome contrast to the dismal 18% of graduates in the U.S., and at Harvey Mudd College, more than half of the freshman engineering class this year were women. Their strategies ranged from featuring women on their brochures and as tour guides, to training teachers and hosting camps for high school students.

Why should we care if girls remain underrepresented in STEM? Apart from basic fairness, if we want our best and brightest working on innovative ideas and creative solutions, it makes little sense to potentially abandon half the population. We already face many hurdles; lack of funding, lack of jobs, and pushback from science denialists backed by populist politics. We need all hands on deck to forge ahead.

We must look to nurture, not nature, for change.