Female and minority researchers ‘focus on less popular topics’

The advancement of female and ethnic minority academics is restrained by their concentration in research areas that tend to attract fewer citations, a study suggests.

The authors of the US-focused analysis, which examined more than 5 million journal articles published by 1.6 million authors between 2008 and 2019, warn that their findings on the unequal distribution of scholars across research topics means that some areas which are important for marginalised groups, such as gender-based violence and mental health, “become systematically less studied”.

Writing in the Proceedings of the National Academy of Sciences (PNAS), an international team of researchers writes that analysis of the Web of Science database indicates significant variation in the gender and ethnic make-up of a research area’s scholarly workforce: for example, black, Latinx and white women are highly underrepresented in physics, mathematics and engineering, and overrepresented in health, psychology and the arts. Asian women are underrepresented in the arts, humanities and social sciences, and overrepresented in biomedicine, chemistry and clinical medicine. Black, Latinx and white men are underrepresented in psychology and health, together with Asian men, with the latter group also underrepresented in humanities and social sciences but overrepresented in physics, engineering, mathematics and chemistry.

Closer examination reveals more detailed distinctions. In social sciences, for example, Asian authors publish most on economics and logistics, where white and black authors are least represented; while black authors are highly represented on racial discrimination, African American culture and African studies. Latinx authors are highly represented on topics relating to immigrants, political identity and racial discrimination.

In general, Asian authors tend to be highly concentrated in particular topic areas, as are black and Latinx authors, while white authors are present in a wider range of topics, write Diego Kozlowski, Vincent Larivière, Cassidy Sugimoto and Thema Monroe-White.

Significantly, they find that the average number of citations that a topic attracts is positively correlated with the presence of Asian and white men among its research base.

“Our results show that minoritised authors tend to publish in scientific disciplines and on research topics that reflect their gendered and racialised social identities…there is a privilege of choice in scientific knowledge production, wherein research on a particular topic is influenced by scientist’s race and gender,” the authors write.

“The ubiquity of white men in science and across topics implies that this demographic group has a wider range of possible strategies to follow and an advantage in the way their scientific capital can be invested, reinforcing inequalities in scholarly outcomes.”

The study echoes the results of an analysis conducted by the US National Institutes of Health in 2019 which found that black scientists won fewer grants from the funder because they chose research topics – such as health disparities and patient-focused interventions – that scientific experts considered less important.

The PNAS paper, whose authors are based in Canada, Luxembourg and the US, indicates a potential cause in the underrepresentation of female and ethnic minority scholars in the higher ranks of academia, given that research impact can be a significant factor in hiring and promotion decisions.

But the authors also argue that the “compound effect of different citation rates of topics and unequal distribution on topics by race and gender leads to negative effects for marginalised groups and for science itself, as some topics become systematically less studied”.

They say that the “history of science is ripe with examples of understudied topics, such as female genitalia, which had direct implications on the life expectancy of women”, and argue that if topics’ research workforces mirrored the ethnic and gender distribution of the US census, there would have been significantly more research published in topics such as public health, gender-based violence, gynaecology and mental health over the past four decades.

The paper, published on 3 January, uses authors’ given and family names to infer their probable ethnicity and gender, computing the probability fractionally to create weighted aggregates.

chris.havergal@timeshighereducation.com