Racial inequalities in journals highlighted in giant study

Scientists from minority ethnic groups experience various forms of inequality throughout the academic-publishing process and are poorly represented on journal editorial boards, according to new research¹.

Editorial boards dictate which papers — and by extension, which researchers — drive the scientific discourse, the study authors note. As a consequence, they suggest, journals should work to ensure that their editorial boards are diverse. Yet many studies have identified a lack of gender, racial and ethnic diversity among editors overseeing the publishing process. One such analysis, which examined 81,000 editors across 15 disciplines, found that women accounted for 14% of editors and 8% of editors-in-chief². And a survey of 368 editors across 25 medical and science journals revealed that more than 75% are white³.

The current study builds on these findings by tracking such discrimination over the past two decades and highlighting other ways in which publishing disenfranchises members of marginalized groups.

The results show that scientists from minority ethnic groups are systematically under-represented on editorial boards compared with their share of authorship of papers. Researchers submitting papers from Asia, Africa and South America, for example, account for 35% of authors but only 19% of editors. The study also found that they often experienced longer waiting periods between the submission of a paper and its acceptance for publication. In the United States, Black scholars face the longest delays, and papers written by teams with a majority of Black or Hispanic scientists are cited less frequently than are textually similar papers published by white researchers.

What’s in a name?

To study the inequities facing scholars from minority ethnic groups, Bedoor AlShebli, a computational social scientist at New York University Abu Dhabi in the United Arab Emirates, and her colleagues amassed publication and citation information for more than one million papers published in more than 500 journals between 2001 and 2020. Specifically, the study focused on six publishers: Frontiers, Hindawi, the Institute of Electrical and Electronics Engineers, the Multidisciplinary Digital Publishing Institute, PLOS and the National Academy of Sciences. The authors used an algorithm to infer people’s race or ethnicity.

They tested three metrics — editorial-board composition, review time and number of citations — and identified 13 nations in Asia, Africa and South America that were under-represented on editorial boards. Of the 20 countries with the longest paper-acceptance delays, 19 were in these regions. In the United States, the study highlighted that Black authors have routinely faced the longest delays over the past two decades. “That disparity in the time spent under review was the most surprising to us,” AlShebli says, “especially since no other study has ever documented any similar finding before.”

Using a metric called citation lensing⁴ that tracks the textual similarity of publications, the researchers were also able to show that, globally, scientists in Africa, the Middle East, Latin America and the Caribbean are cited significantly less often than expected across all disciplines, whereas those in North America and Oceania are referenced more. Looking at only the United States, the study found that papers from teams in which most authors are Black or Hispanic are not cited as often as those in which most authors are white, even when the content is very similar.

AlShebli notes that although two independent analyses produced the same results, the tools used to infer demographic information are imperfect. Jeffrey Lockhart, who studies the sociology of knowledge and science at the University of Chicago in Illinois, published a study on 17 April⁵ outlining how gender-, race- and ethnicity-detecting algorithms can have substantial error rates. In the United States, for example, name-based algorithms misclassify Black names roughly 65% of the time.

Lockhart doesn’t doubt the conclusions of AlShebli’s study, some of which “fit very well with established sociological literature going back decades”. But without a direct measure of the algorithms’ accuracy, he says, it might be more appropriate to frame the results as measuring discrimination against names that seem unfamiliar to reviewers. “I find that idea even more interesting, because there are all kinds of social and personal and psychological dynamics that go into how you pick a name,” Lockhart says. “It could open the door to a lot of very intriguing studies.”

Mechanisms of discrimination

With these results in hand, AlShebli says, a next step will be to identify the causes of such disparities, which are probably rooted in broader societal issues (see also ‘Tips for combating citation bias’). One explanation for the findings might be that scientists from minority ethnic groups are not invited to sit on editorial boards at the same rate as their white counterparts are. But it could also be that “researchers are hesitant to accept requests to review papers” from editors with foreign affiliations, she says, creating barriers that members of marginalized groups must overcome to complete their editorial work. If that’s true, it might not be enough to simply invite more people onto boards without policies in place to support and retain them.

Careers Collection: Publishing

Edmond Sanganyado, an environmental toxicologist at Northumbria University in Newcastle upon Tyne, UK, began serving on editorial boards after his own experiences in academic publishing. After a move from the United States to China as part of his career, he noticed a shift in reviewer comments, including critical notes about his written English. Sanganyado, who is originally from Zimbabwe, thought that if he joined an editorial board, he could advise peer reviewers to focus on the quality of the science.

What he found, however, was that he often struggled to solicit reviewers at all. Whereas his colleagues lamented having to send a dozen e-mails, Sanganyado often contacted up to 40 people, stretching the acceptance process to months. This didn’t happen with journals that didn’t reveal the authors or editor’s names as part of the review process, leading him to resign from several boards because he thought he was experiencing discrimination because of his name. “I couldn’t continue to be the barrier to people who are trying to advance their careers,” he says.

Cristina Dorador, a microbial ecologist at the University of Antofagasta in Chile, has also experienced some of the struggles outlined in the paper, and says that she is glad to see that these issues are receiving attention. She and her colleagues struggle to publish their discoveries in high-impact journals — so much so that they often joke about the reasons reviewers have given for turning them down. “As Latin American scientists, I think we feel suspicious that something is happening,” she says, adding that such barriers seem to melt away when she and her team are collaborating with colleagues in North America. “But we have to include the visions and ideas of the Southern Hemisphere, because right now we are out of the conversation.”

Holding publishers to a higher standard

AlShebli’s study included only papers published until the end of 2020. Since then, many publishers, universities and professional organizations have brought about changes⁶ to address a lack of diversity in science, technology, engineering and mathematics. Currently, more than 50 academic publishers representing more than 15,000 journals have committed to tracking the gender and race or ethnicity of their authors.

Racism: Overcoming science’s toxic legacy

Shortly after researchers⁷ identified gender and regional biases in the demographics of people quoted in Nature’s news coverage in 2021, its news team began tracking the gender, geographical location and career stage of all of its sources. The news team reported in February that the percentage of quotes from sources identifying as male was brought down from 69% in 2020 to 55% in articles published since April 2021. The journal has also committed to addressing racism in its published research, putting out a special issue — the first in its history to be guest-edited — that included an acknowledgement of Nature’s role in perpetuating harmful science.

Among the efforts being advanced by publishers are flexible payment models for papers submitted by scientists in low- or middle-income countries. Frederick Fenter, the chief executive of Frontiers in Lausanne, Switzerland, says that “bias in scientific publishing is real, and all responsible publishers develop their programmes with full awareness of this reality”. Each editorial board overseeing a Frontiers journal is regularly audited to ensure balanced gender and geographical representation in its published papers, and authors are able to apply for financial assistance. In 2022, Fender notes, 44% of articles from South America and 38% from African nations received partial or full fee support.

In a statement, May Berenbaum, the editor-in-chief of the Proceedings of the National Academy of Sciences, says that the journal is “committed to enhancing DEI [diversity, equity and inclusion] in science publishing”. It recruits reviewers from “diverse backgrounds”, and all reviewers receive training in unconscious bias as part of their onboarding.

Other publishers are focusing their attention on new open-science publications, which can be more equitable. Madhukar Pai, one of two editors-in-chief at PLoS Global Public Health, says that when his journal launched in 2021, “we knew exactly what was wrong with the global-health journal landscape, and intentionally set out to do things differently”, including building diversity into every level of the publication process. Suzanne Farley, PLOS’s editorial director, confirms that the publisher is now collecting demographic information from its authors and plans to survey its 10,000-plus editorial-board members. “Everything we have done, we have done intentionally to make sure our editorial boards are diverse and inclusive and that we are platforming authors who are normally excluded,” Pai says.