Special Report
Gender Gap in Research Output Widens During Pandemic
Special Report
Gender Gap in Research Output Widens During Pandemic

Gender Gap in Research Output Widens During Pandemic

Experts identify childcare, which tends to fall to women, as one likely cause for the relative decrease in women’s scientific productivity compared with men’s.

Katarina Zimmer
Katarina Zimmer
Jun 25, 2020

ABOVE: © ISTOCK.COM, GOLERO

When New York City went into lockdown in mid-March and most research had to be put on hold, virologist Benhur Lee’s lab at Mount Sinai’s Icahn School of Medicine was one of the few still running. The lab and its 10 researchers had pivoted their studies of various viruses to SARS-CoV-2, the coronavirus that causes COVID-19.

But with schools and daycare centers also closed, Jillian Carmichael—a postdoc and the only mother in the lab—couldn’t join the team. Instead, she spent two months inside a 600-square-foot apartment in Queens, taking care of her six-year-old daughter and three-year-old son as her husband, who works in healthcare, conducted telemedicine over the phone.

Although noise-cancelling headphones allowed her to get some work done from home, and her husband took on his fair share of childcare, her research on herpesviruses was put largely on hold, and she wasn’t able to contribute much to her lab’s COVID-19 research. “This is what I wanted to do with my life—study viruses—and then to be at home and not have a lab [during a pandemic]. . . . It’s sort of like watching a ship pass you and you really want to be on it,” she says.

Eventually, she and her husband, who was recently furloughed, decided that he would move with the children to stay with family in the Midwest for the summer so they’d have more space and help with childcare, and so that Carmichael could continue her research. “I’m fortunate that I have family that can help out . . . but after August, I really don’t know what I’m going to do.”

While the number of male coauthors of arXiv preprints grew by 6.4 percent from 2019 to 2020, the number of female authors increased only by 2.7 percent.

Carmichael’s situation may be familiar to any academic who is a parent, but especially to women. A series of recent studies has demonstrated a significant drop in the productivity of female scientists, especially those early in their careers, relative to their male peers—and the gender gap is particularly pronounced for COVID-19 researchers. Most researchers whom The Scientist spoke with pin the blame on childcare, a burden that disproportionately falls onto women. Researchers say they worry that the pandemic will exacerbate the already existing underrepresentation of women in scientific research and impair women’s scientific careers as well as the quality of research itself.

“The results were not shocking, but they were an affirmation of what we already felt,” notes Cassidy Sugimoto, an information scientist at Indiana University who specializes in gender disparities in research and has investigated publishing trends in the face of the pandemic. “If institutions don’t take [this] seriously, I think we’re going to prolong the effects of these disadvantages for certain populations.”

Pandemic exacerbates existing inequalities

Women have long been underrepresented in research, accounting for around 31 percent of coauthors on scientific papers between 2008 and 2017, according to an unpublished analysis by Sugimoto and colleagues. This spring, concerns on social media started to surface that this gap may be widening. For instance, women were submitting a “negligible” number of articles to The British Journal for the Philosophy of Science since the start of the pandemic, the journal’s deputy editor Elizabeth Hannon noted on Twitter. “Never seen anything like it,” she wrote, according to Inside Higher Ed.

Anecdotal accounts like this have motivated researchers, including Megan Frederickson, an ecologist and evolutionary biologist at the University of Toronto, to see if a trend would be evident in publishing data. In the evenings after her own six-year-old went to bed, Frederickson adapted some code she had used for previous studies on gender disparities to plow through papers posted to the preprint servers arXiv and bioRxiv, and used an algorithm to identify coauthors’ genders based on their names. Although identifying gender based on names alone is criticized for misidentifying non-binary individuals and often also researchers who don’t have Western names, they can help reveal overarching trends in gender disparities when used on large datasets, a news article in Nature notes.

To her surprise, given her own lack of productivity with a “wonderful” but at times distracting son, she found that many more preprints were submitted between mid-March and mid-April this year than during the same time period last year, which she suspects is partially explained by a flurry of COVID-19–related studies rushing to make sense of the unfolding pandemic, she says. However, while the number of male coauthors of arXiv preprints grew by 6.4 percent from 2019 to 2020, the number of female authors increased only by 2.7 percent.

Researchers developed a real-time tool to track the representation of female coauthors on preprint studies in the months preceding and during the pandemic. Across different preprint servers, the largest dip among female first authorship was seen on medRxiv, from nearly 36 percent in December to around 20 percent in April.  
Philippe Vincent-Lamarre, Cassidy R. Sugimoto and Vincent Larivière 

Sugimoto found a similar trend in an analysis of 11 different preprint repositories. She and her colleagues even created a real-time tracker so anyone can monitor how these trends evolve over time. To her, the relative drop in women’s productivity is particularly striking given that in the months before the pandemic, the percentage of female authors on arXiv and bioRxiv preprints was on the rise. “It does look like when people have gone to look, they’ve basically found an effect. [They’re] different magnitudes of effects, but still in the same direction,” Frederickson notes.

Sugimoto and Frederickson tell The Scientist they suspect that childcare duties, which have increased with the closure of schools and daycare facilities, are a significant factor driving this trend. Male academics have children too, but they’re four times more likely to have a partner who is a full-time caregiver than are female academics, who are more likely to have partners who also work outside the home. Women are also more likely to be single parents than men are. Even in relationships with two partners, both of whom are academics, there’s often an unequal division of labor. In a recent survey on parenting among academics, Sugimoto and her colleagues found that women tended to take on more childcare duties, even if couples insisted that the work is split evenly between them.

In terms of how the situation is playing out in the pandemic, a recent poll published by The New York Times suggests that the division of childcare isn’t equal between men and women. Among 2,200 respondents, around half of men said they were doing most of their children’s homeschooling over the past few months, whereas only 3 percent of women agreed that their partner was doing the bulk, the poll found.

Women also tend to take on domestic labor and the responsibility of caring for old or sick relatives, and some studies suggest that female academics do more teaching than men, on average, so perhaps the transition to online courses also affected women’s research productivity, Frederickson adds. An additional factor could be demographic, as many more men were hired as professors than women in the 1980s, and hiring women faculty only became more common in recent decades. “As a result of that, the average female faculty member at a university is younger than the average male faculty member. Just because of that, it’s possible that women are more likely to have kids at home than male faculty members,” she suggests.

Early-career researchers most affected

Like Carmichael, Kishana Taylor, a postdoc at the University of California, Davis, has been struggling to get work done with a one-year-old at home and no daycare to send him to. While she and her husband try to split childcare evenly, her son prefers care from her most of the time, often running into her office when she’s on a call or working, she says. “Maybe on a good day I’m getting like two-and-a-half, three hours of work, compared to the eight or nine that I would be getting otherwise.”

That has affected her ability to wrap up some of her remaining postdoc projects on influenza evolution and to apply for new academic positions and attend online seminars and conferences, she says. And it’s unclear how long the situation will last. Taylor says she sent her son back to daycare last week, but “I don’t know that we’ll keep him there in the long term, especially if cases are going back up again.”

Sugimoto’s analysis suggests that women early on in their research careers are particularly affected by this loss in productivity. When she analyzed gender disparities during March and April among first authors—which tend to be early-career researchers—she found that the differences between men and women were more pronounced than looking at overall gender disparities during that time. “That worries me a lot, because we know how cumulative effects work in science: the best predictor of having a publication is having a previous one. . . . This could have a huge effect on this cohort of young women.”

Both Carmichael and Taylor expressed concerns that the pandemic will leave gaps in their CVs and make them less competitive on the academic job market. “And I know that’s the concern of . . . the postdoc women that I know who are also mothers, or caring for other family members,” Taylor says.

Gender disparity could have consequences for COVID-19 research

Two independent analyses indicate that the COVID-19 research field may be particularly affected by the gender gap. One examined authors with typically male and female names on nearly 2,000 COVID-19–related medical studies and found that the proportion with female first authors was nearly 20 percent lower than for medical studies published in 2019. Another study of 1,445 COVID-19 papers on PubMed found that, overall, women made up a little more than one-third of authors on coronavirus research, which was in keeping with figures across other fields, but this proportion was much lower for first and senior authors on COVID-19 papers than in other fields, notes Ana-Catarina Pinho-Gomes, a doctoral student at the University of Oxford’s George Institute for Global Health and a coauthor on the analysis. Sugimoto’s interpretation is that women could be relegated from first and senior author positions into less-influential middle authorship positions. Likewise, female economists, who have also seen a recent drop in productivity during the pandemic, were less likely to publish on COVID-19–related topics relative to their male peers, analyses have shown.

Having more women researchers may be particularly important with COVID-19, a disease that is increasingly recognized as affecting men and women differently.

As to why there is less representation by many women coauthors on coronavirus papers than on studies before the pandemic, Frederickson says that could reflect the constraints of familial obligations limiting their ability to pivot their research to a time-sensitive, fast-moving field. “I would love to have a COVID first-author paper, but I don’t think it’s going to happen,” Carmichael says. “The science is really being pushed at warp speed.”

Pinho-Gomes says she worries that this could have consequences for COVID-19 research itself. Some of Sugimoto’s research has shown that study methodologies and conclusions depend, in part, on the authors. In an analysis of more than 11.5 million medical studies published between 1980 and 2016, she found that that the presence and number of women authors on a paper influence the likelihood that the scientists will examine sex as a variable in their analyses. “Female authors were more likely to look at women as populations of study in their medical research and more likely to take both men and women into account,” Sugimoto says, noting a 2001 government accountability report estimating that of 10 prescription drugs withdrawn between 1997 and 2001, eight were taken off the market due to adverse effects for women, likely because they were tested on men only.

Having more women researchers may be particularly important with COVID-19, a disease that is increasingly recognized as affecting men and women differently, Pinho-Gomes notes. If “women are not shaping the research response to the pandemic. . . there is a gendered lens through which we’re looking at [the coronavirus],” she says.

The same goes for other groups that are also underrepresented in science, such as Native Americans, Blacks, and Hispanics, who are disproportionately affected by the pandemic. Sugimoto says she expects that researchers from those communities may also experience research productivity losses, because they are more likely to become sick or care for family members who are sick. However, that effect is harder to study, Frederickson adds.

Having a diverse community of researchers at the table during the design of drug and vaccine trials and other COVID-19 research “will allow you to get better and more complete results,” Taylor says. “How I would approach recruitment for vaccine trials might be very different from someone else.”

Preventing the pandemic from worsening existing inequalities in research will require action, as well as compassion, at all levels of the scientific community, Sugimoto says. Lab leaders could give certain members of the lab who may have less time to do experiments more time to publish their work, for example. Funding agencies could put more money toward early-career and minority researchers. Institutions could take into consideration that many scientists may not be able to return in the fall because childcare is either unavailable or because they feel that putting their kids back in childcare may be a risk to themselves or family members. 

“I think also just letting us know that you’re . . . trying to find ways to mitigate some of the potential fallout [from the pandemic.],” Taylor says. “I think that would be comforting, and relieve us of some of the stress and worry that we have at the moment.” 

Correction (June 25): A previous version of this story erroneously stated that from 2019 to 2020, the number of male coauthors of bioRxiv preprints increased by 6.3 percent, compared to only 2.7 percent for women. That statistic referred to an analysis of arXiv preprints, not of bioRxiv preprints. The Scientist regrets the error.