Researchers investigating the news media have consistently found that women are quoted far less frequently than men. Since October 2018, the Gender Gap Tracker (GGT), an automated system developed by a team at Simon Fraser University in Burnaby, Canada, has tracked seven Canadian news sites, finding that 71% of interviewees featured in articles were men.
Last November, four other media organisations from around the world (UK’s Reach Plc, Japanese newspaper Nikkei, German TV group Deutsche Welle and international news agency Agence France-Presse) collaborated with the GGT to track 5 days of their own news coverage and discovered that 73 percent of statements came from men. In March, the Global Media Monitoring Research predicted that just 25% of news sources and subjects in 2020 will be women, up from 17% in 1995, when the project began.
Nature's journalism has now been examined by two American researchers, who have come to equally dismal conclusions.
According to the study, men accounted for 69 per cent of direct quotes (not including paraphrased comments) in Nature journalism pieces in 2020. According to a software study of the gender of people quoted in over 16,000 Nature news, features, and careers items published between 2005 and 2020, were men. The percentage of men quoted in Nature journalism has been declining in recent years. Before 2017, it was around 80 per cent, and 87 percent in 2005.
These findings serve as a timely reminder of the problem of gender bias in journalism, which Nature's editors claim they are working to rectify. The figures also indicate how software may help writers and editors spot biases, and that they need to work harder at Nature to eliminate them.
Software was used to automate the analysis, which was posted before peer review. The team began by scraping articles from journalists that had been published on nature.com. They then created code to identify the identities of those who were quoted by reporters, counting just those whose quotes were in speech marks. Gender was assigned using another technique called genderize.io, which is a usual strategy for large bibliometric datasets.
The study's concept was conceived in collaboration with Nature, but Natalie Davidson and Casey Greene, both computational biologists at the University of Colorado School of Medicine in Aurora, designed and conducted the study independently.
In Nature's Careers feature articles, the duo discovered one exception to the general conclusion on gender bias. In this section, which includes journalistic reporting on many elements of research careers, quotes from both men and women occur in equal numbers.
The study does not evaluate all of Nature's non-primary-research output; content authored by invited expert authors, for example, is not included. This type of literature isn't known for directly quoting other people.
Nature, on the other hand, has been collecting data on gender diversity among the authors of such commissioned content for the past five years. Women accounted for 58 per cent of authors in Nature's World View column last year, up from 35 per cent in 2017 and 18 per cent in 2016. Women also accounted for 34 per cent of authors of News and Views pieces (which explain and analyse new findings) in 2020, compared to 26 per cent in 2015. The photo-essay feature, Where I Work, which portrays researchers in their locations of study, is another kind of journalism. Since its inception in 2019, this has included 56 per cent female scientists.
There are some caveats, as with all studies. Not all names were analysed, and Davidson and Greene admit that their software has a minor masculine bias when it comes to gender assignment. For example, it categorised 78 per cent of quoted speakers as male in a sample of papers from 2005 to 2015, although the correct number was 75 per cent when the authors checked. It also has no way of determining non-binary gender.
The researchers consider various ways of measuring the overall proportion of women in academic research to help contextualise their findings. UNESCO, the United Nations institution for research and education, published a global scientific report earlier this month that puts this at 33 per cent in 2018. Women accounted up roughly 20 per cent of last-author positions and 25 per cent of first-author positions on Nature articles, according to Davidson and Greene; the ratios are around 25 per cent and 37 per cent in a larger sample of papers in Springer Nature journals.
Davidson and Greene also looked at what they call the "name origins" of quoted interviewees in Nature's journalism, which is a linguistic study that assigns names to wide geographic locations where a certain name is over-represented. Because of the diversity of names in these nations, the authors employ an algorithm called NamePrism to remove the United States, Canada, and Australia.
According to the analysis, journalists at Nature tend to quote more researchers with names that are common in English-speaking nations and fewer researchers with names that the algorithm defines as being of East Asian origin (including China, Singapore, Vietnam and other south-east Asian countries). It also suggests that the disparity in name origins is bigger than that seen in last author names.
Nature's journalism staff has been working to track and enhance its representation of all under-represented groups, but this hasn't been a coordinated effort in the past. However, for the last year, the team has been creating and testing a prototype system with the purpose of gathering data on the gender, career stage, and location of journalists while adhering to data-privacy standards.
Although more than half of Nature's journalism staff is female, the majority of its personnel are based in Europe, the United States, and Australia. However, they have acknowledged that they must work harder to find a variety of sources from throughout the world.