Gender bias might be working at level of whole disciplines

Inequity between men and women in research funding and researcher performance evaluation has been firmly established as a problem in science policy for many years. But an emerging body of research is documenting a new piece of the puzzle, which some think has been hiding in plain sight: that whole research fields might be prone to gender bias.

“People just believe that there are some disciplines that are better than others,” says Alex James, a mathematician at the University of Canterbury in Christchurch, New Zealand, who led one the latest studies on the phenomenon. “And it turns out that ones that we think are a bit rubbish are all full of women.”

The study by James and her co-authors, published in eLife on 7 May¹, found that, the more women there were in a field, the lower the overall grant-application success rate and evaluation of researcher quality, according to the analysis of data from more than 30 countries. It builds on other studies published in the past decade that have investigated the various ways in which sexism and other biases might be playing out across and between disciplines. But researchers are divided as to how surprising the latest findings are.

Rich data

James and her colleagues looked at funding data from several organizations spanning tens of thousands of researchers and grant applicants in more than 40 disciplines. This included data from the previous three rounds of New Zealand’s Performance-Based Research Fund (PBRF), which evaluates the research quality of every academic in the country; ten years of data on grant-application success rates from the Australian Research Council (ARC); and similar data sets from the Canadian Institutes of Health Research (CIHR) and the European Institute for Gender Equality (EIGE), the latter of which collates data from government funders across the European Union, the United Kingdom, Turkey and Israel.

2024 Research Leaders

Using statistical modelling, the study authors investigated the relationship between the gender balance of a researcher’s discipline and their PBRF score, accounting for factors such as the researcher’s age, institution and publication record. With the other data sets, the authors analysed the gender balance of a discipline alongside funding success rates.

The New Zealand data set revealed that researchers of any gender working in female-dominated disciplines, such as nursing or education, tended to receive lower scores in the PBRF than did those working in male-dominated disciplines, such as physics and philosophy. According to the modelling, men working in male-dominated disciplines scored an average of around 40 points, out of 700, more than did men working in female-dominated disciplines. Meanwhile, women working in a male-dominated field scored some 70 points more than did women in a female-dominated field.

The latest ARC data set, from 2019, revealed that, in the same discipline, women had a slightly higher success rate in securing grants than men did. But the difference in success rates between disciplines was pronounced. For example, the average success rate for both men and women in philosophy (male-dominated) was 22–23%, but in nursing (female-dominated), it was 17–18%. Similar patterns were observed in the CIHR and EIGE data.

This was not the case in the oldest ARC data set, from 2010. At that time, men were generally more likely than women to secure funding across disciplines, and the differences in success rates between male- and female-dominated fields were limited. For example, success rates in philosophy were 31–33%, similar to those of nursing, at 30–33%.

Disciplinary bias

The authors are not able to explain these patterns, but suggest several reasons. It could be that grant-application reviewers are biased against women or female-dominated disciplines, or, as previously male-dominated disciplines began to attract more women, the perception of the fields’ quality dropped.

James says that many people are aware of how the gender balance of a field alters how others perceive it, but this has not been documented until now.

Virginia Valian, a psychologist at Hunter College at the City University of New York, in New York City, was surprised by the findings suggesting that women benefit from working in male-dominated fields. Women “appear to be more devalued in fields in which they are the majority and, if anything, more valued in fields in which they are a minority. I do not think that this is generally the case,” she says. This effect could be down to of a particular quirk of the ARC data set analysed, she adds.

But Ebony McGee, who studies structural racism at Johns Hopkins University in Baltimore, Maryland, was not surprised by the findings.

“Female-dominated disciplines are facing marginalization in evaluation due to biases that say that work is not rigorous or valuable,” she says. Ideas around what constitutes valuable research originate from entrenched power structures usually developed by white, male, non-disabled people from wealthier social classes, she adds.

McGee argues that using data to highlight potential biases is not enough. “One has to dismantle the field and rebuild it with women and women of colour leading the effort, but people do not want to give up or share their power,” she says.

James and her colleagues are not the first to have looked at how gender and other biases might be related to disciplines. In 2019, researchers at the US National Institutes of Health suggested that Black scientists more often proposed research projects in fields that had lower success rates for funding². Topic choice accounted for more than 20% of the funding gap between white and Black scientists when other variables that affect success, such as previous achievement, were controlled for.

And in 2015, researchers found that women were under-represented in fields in which success was perceived as requiring raw, innate talent — such as mathematics or philosophy — compared with disciplines in which people thought success could be gained through hard work³.

Rachael Murray, a biomedical scientist at the Queensland University of Technology in Brisbane, Australia, says that the studies in this area highlight the complexities involved. “We don’t yet fully understand where all the issues lie,” she says. She adds that institutions and funding bodies have an obligation to look thoroughly at their data — which are often not accessible to researchers — and identify problems that might need addressing.

Some of the organizations that were featured in the eLife study are taking action on the issue. An ARC spokesperson says that the agency is reviewing its grant-awarding mechanism, and will look at how to support “a strong and diverse workforce”. Research funders in Australia have previously made big moves to try to rectify gender bias. In 2022, the country’s main health- and medical-research funder, the National Health and Medical Research Council, overhauled its funding process, introducing quotas that specify that half of its grants will go to women, after an analysis found that men were disproportionately benefiting from the previous system.

In New Zealand, the study by James and her colleagues prompted discussions at the Ministry of Education, which administers the PBRF. Katrine Sutich, the general manager for tertiary and evidence policy at the ministry in Wellington, says an independent panel that reviewed the PBRF has met with the researchers and has taken their findings into account.