A case study on gender bias in science reporting
Gender disparities in science have garnered a lot of attention over the past decade, but prejudices against women in media coverage of science have not received as much attention. A recent study by NatureNews and feature articles highlight how often women are mentioned in science news. The study finds that women continue to be cited less often than men in the high-level journal, although the gap appears to be narrowing.
Led by Natalie Davidson and Casey Greene of the University of Colorado School of Medicine, study, published in June on bioRxiv, analyzed more than 22,000 journalist news and feature articles that appeared in the first half of Nature from 2005 to 2020. Researchers used software to roughly identify the gender and ethnicity of authors and sources. The software had a few limitations: it had a slight masculine bias, did not include a non-binary gender assignment, and could not identify all names. The researchers compared software-assessed demographics from sources cited in Natureof the news section with that of the authors of more than 13,000 research articles published in the second half of Nature during this period.
Davidson and Greene found a significant decrease in the proportion of men cited in Naturepages during these 15 years. The analysis found that in 2005, 87% of citations were deemed to be from men, while male researchers were the first authors of 73% of research articles in the study sample. By 2020, the likelihood of quotes coming from men had fallen to around 69%. Articles on career-related topics were the only ones to achieve gender parity, according to the study.
Although Nature publishes research in several disciplines with different proportions of female participation, the new study does not distinguish the results by field or by subject. Research papers published in Nature also may not be representative of the gender balance in the academic community or research in general, Davidson says.
To account for this possible bias, the researchers selected another random sample of 36,000 research articles published during the same period in other journals led by publishing giant Springer Nature, which owns Nature. Compared to this data set, the estimated proportion of men cited in Nature News and feature articles in 2020 (69%) is higher than the percentage of male first authors of articles in the sample (around 63%) but lower than the rate of male last authors (76%). In contrast, the citation rate for men is lower than the rates for the first and last male authors of Nature manuscripts, which are 74% and 80% respectively. Davidson says that Nature research papers are more likely to be male dominated (in early and late authors) than those published in other Springer Nature journals.
In an editorial published in response to analysis, Nature acknowledges that its reporters need to redouble their efforts to eliminate bias, noting that the new analysis has shown that software can be used to recognize such trends. The editorial also mentions that Nature has collected data on gender diversity in its commissioned content over the past five years.
External researchers say the new research, while not yet peer-reviewed, is solid. Luke Holman, an evolutionary biologist at Napier University in Edinburgh in the UK, said the new study had “a new, high-quality and transparent methodology.” Holman co-author of a 2018 study which revealed that at the current rate of change, it would take 16 years for researchers – on average in all scientific disciplines – to catch up with men and produce the same number of articles. In physics, the gender gap would take 258 years to disappear.
While Holman likes the new study, he notes that it doesn’t mention how many different people are cited in each sample article, how many quotes are from the same sources, and how much page space is given to each source.
“It’s really a good thing that more women scientists are being named, even though things like this don’t normally directly contribute to tenure decisions,” says Barton Hamilton, an economist at the University of Washington in St. Louis who wrote on the gender gap in National Institutes of Health grant applications. “It is very important that the teachers cited are representative of the teachers who do the work.
Other analyzes have also shown that women are more often cited in scientific news than in the past. For example, the World Association for Christian Communication published the last quinquennial report on July 14 as part of the Global Media Monitoring Project of the non-governmental organization. The report examined, among other things, the extent to which women are cited in the news media. Of all the hot topics, women feature most often as current science and health sources and topics, says study editor Sarah Macharia, Toronto-based gender, media and international politics consultant .
In 1995, women made up 27% of the subjects and sources of science and health articles in various types of media; that representation rose to 35% in 2015. But Macharia says that as scientific and health news has increased in percentage of all media coverage during the COVID-19 pandemic, the proportion of women as subjects and sources fell to 30%. “As this topic has become so dramatically in the limelight and interest in the news has grown, women have been displaced from this space,” says Macharia.
Davidson and Greene’s study also compared citation rates to first or last author rates for scientists with various name origins in manuscripts published in Nature and other Springer Nature journals. They reported a serious undercutting of their paternity rates for scientists whose names are from East Asia, and an overrepresentation of scientists with English, Irish, Scottish or Welsh names.
Deborah Blum, science editor and director of the Knight Science Journalism program at MIT, explains that the reporter is responsible for identifying who did what in the studies he reports on and finding commentators from a variety of backgrounds who have not worked on the studies. “And it’s not just journalists trying to be politically correct,” she said. “When you have a diversity of scientists, science is smarter. If you incorporate this into your reporting, your story is smarter, too. “
But choosing diverse sources is not always easy or straightforward. Journalists often have to produce articles under tight deadlines and are sometimes limited to experts in a particular geographic location or time zone.
Holman sympathizes with the work of sourcing, but says, “If you have the opportunity to name two equally qualified people and one of them is from an under-represented part of the world or is female,” it is better to quote this person.
Editor’s note: Dalmeet Singh Chawla writes news regularly rooms for Nature but had no involvement in the study. Madison Brewer conducted the research for the Physics today case study described in box.