Research routinely fails to account for sex and gender differences

Research funders and scientists routinely fail to consider the crucial impact of sex and gender with respect to the way research is designed, carried out and implemented.

A new report published by the League of European Research Universities (LERU), ‘Gendered research and innovation: Integrating sex and gender analysis into the research process’, describes the negative effects of failing to take sex and gender into account, as well as the positive effects of including them.

“We don’t systematically say to ourselves at the point of funding, at the point of doing a survey – should this research make a distinction between men and women?” said Simone Buitendijk, vice-rector of the University of Leiden, at the report’s launch on Wednesday in Brussels.

Perhaps as a result, research that includes more male subjects has long been the norm, Buitendijk noted. Some of the most striking examples of the effects of the failure to take sex and gender into account can be found in health and medicine.

Diseases, treatments, and chemicals may affect the sexes differently, yet there is a long and storied tradition of ignoring gender when it comes to health research.

For example, research shows that women tend to present different heart attack symptoms than men. This knowledge was absent until recently, simply because most research had been carried out on men only, and the assumption was made that there were no significant differences between the sexes.

“Lives are being lost because diagnostics and therapeutics are not as good for women as they are for men,” said Buitendijk. In the US, 10 drugs were withdrawn between 1997-2000 because of life-threatening health effects. Eight of these posed greater health risks for women than for men.

It is not just in health research that female subjects have been excluded. The automotive industry did not routinely test its cars with pregnant crash dummies until 1996. Until then, cars were fitted with seatbelts that didn’t fit pregnant women properly, putting the lives of unborn children at risk.

Research that does not adequately include both sexes opens up a space where stereotypes can grow, Buitendijk said.

Because of the perception that eating disorders are the province of ‘womens’ illnesses’, men are underdiagnosed and undertreated for anorexia, despite making up about a quarter of all cases.

Without proper market analyses, manufacturers in the tech world often wrongly assume products will appeal to women if they undergo ‘shrinking and pinking’ exercises.  

Even in ornithology, there is a false belief that only male birds sing, because research in birds has almost exclusively been carried out on males. “Lo and behold, research from University of Leiden shows us that female birds also sing,” Buitendijk deadpanned. 

Challenging bias

The report stresses repeatedly that gender-enhanced science benefits science overall.

For example, the development of seatbelts for pregnant women has in fact lead to designs that are directly better for everyone.

This message is being carried forward by an initiative led by Stanford University called ‘Gendered Innovations’, which gets funding from the European Commission and the US National Science Foundation.

The director of the project, Londa Schiebinger, leads a team of 70 scientists, engineers, and gender experts from across the US, Europe, Canada, and Asia, with the aim of exploring how gender analysis can open doors to discovery.

She makes it her business to challenge gender bias wherever she finds it. For instance, it was revealed to her that Google Translate had “a masculine default”, meaning it massively overused masculine pronouns, even when the text it is working on specifically refers to a woman.

“You’d try translating a speech and you get ‘He wrote’ ‘He said’ and occasionally ‘it said’,” reported Schiebinger. “With one algorithm, Google wiped out 40 years of revolution in language.”

Schiebinger and her team organised a workshop with Google’s language processing experts, who told her it could be fixed.

The episode encouraged her look a little closer at Stanford’s engineering faculty, which is responsible for hatching many of Silicon Valley’s future stars, but she did not find clear sex and gender analysis methods in its curricula. When she pointed this out to faculty members, Schiebinger faced some initial resistance. “They said ‘we have no time for this’ – but I have strategies,” she said.

To change things, the best incentive is probably money, said Schiebinger. Borrowing an idea from Curt Rice, head of Norway's committee on gender balance in research, Stanford introduced small seed grants of $25,000 a year, for research with a strong gender dimension. “It’s not a lot of money but it’s enough to interest researchers,” she said.

Commission ‘a global leader’ on gender

Researchers say the Commission is setting a good example when it comes to including a gender dimension in research policies and programmes, raising awareness and providing information and training on the topic.

According to Schiebinger, “Currently, the European Commission is the leader in sex and gender research.”

The Commission has designated 137 subfields where data showed that gender analysis could benefit research. These range from computer hardware and architecture to nanotechnology, oceanography and geosciences, among others.

But this does not mean their work is done. Viviane Willis-Mazzichi, head of gender in the Commission’s directorate for research and innovation, said that while application forms for the Horizon 2020 research programme often stress requirements for gender expertise, some applicants skim over it.

“There’s people ticking the box for ‘gender expertise’ on applications and then you find out they don’t have any,” she said. 

Outside Europe, progress is being made. In June, the US National Institutes of Health rolled out new guidelines for sex inclusion in research. In August 2015, the Canadian Institutes for Health Research rolled out the first of three training programmes for sex and gender in biomedical research.

Broad responsibility

Schiebinger said she would like to see academic journals hold a tougher line when it comes to evaluating articles for sex and gender analysis methods.

Although journals like The Lancet, PLOS Biology, Nature and Science now ask that authors provide sex- or gender-specific reporting of research, there was a question at the LERU launch from the audience over whether this should really be journals’ domain. 

Stephane Berghmans, Elsevier’s vice-president of academic relations in Europe, argued that, because journals are at the very end of the research chain, it is hard for them to influence the data collection habits of researchers.

“I have been talking to journal editors – they will send a paper back if it hasn’t addressed gender issues,” replied Schiebinger. “It’s a re-iterative process and journals can hold the line on these issues.”

Clearly there is a need for broad responsibility and leadership, not just from governments and journals, but universities, funders and individual researchers, said Buitendijk.

“It needs a Chinese water torture approach – I mean the good kind,” she said.