As faculties become more diverse, these researchers want to help citations catch up
This spring, an editorial appeared in Annals of Biomedical Engineering encouraging authors to acknowledge the unconscious biases that may have shaped a particular section of the manuscript — the list of references at the end.
The editorial, written by editors of the four journals published by the Biomedical Engineering Society (BMES) and the chair of the society's publications committee, suggests that authors add a citation diversity statement. These statements highlight widespread inequalities in whose work is getting cited — inequalities that reinforce broader imbalances in gender, race, and ethnicity.
Annals of Biomedical Engineering is the society’s flagship journal; Bethany Rowson, a research assistant professor of biomedical engineering and mechanics and the journal’s managing editor, led the citation diversity initiative with editor-in-chief Stefan Duma, the Harry Wyatt Professor of Engineering and director of the Institute for Critical Technology and Applied Science.
Studies of citation patterns reveal that women are cited less often than men, and researchers of color less often than their white counterparts. These imbalances — driven, at least in the case of gender, primarily by the citation habits of male authors — pop up across a broad swath of academic fields and tend to get more deeply entrenched over time, as citations attract prestige and then more citations in a pernicious feedback loop of perpetuated privilege.
“I think most people aren't aware of gender or race when they're citing papers unless they personally know the person they’re citing,” Rowson said. “But whether people realize it or not, these biases exist, and if even a small number of people reinforce them, they get propagated and magnified.”
Structural inequalities and conscious and unconscious biases have shaped university faculties where women and faculty of color are underrepresented — particularly so among the tiers of senior researchers whose work typically attracts the most citations. In part, disparities in citations reflect the formal and informal professional networks that reinforce these demographic disparities — people tend to cite people they know. Stereotypes casting women and people of color as less innovative can also lead to an unconscious devaluation of their contributions.
As faculties diversify, efforts to encourage parity in citations can help give newer researchers from underrepresented groups a more level playing field.
“The primary goal is to get people looking at this as an issue,” said Duma. “If authors become more aware of unintended bias in who they’re citing and referencing, that’s going to help us make some progress.”
Citations have only recently begun to attract attention as a signal of inequality. In 2020, former BMES president Lori Setton spotted a seminal paper on citation diversity statements in the journal Neuron. She shared the article, authored by a biomedical engineer at the University of Pennsylvania, with the BMES publications committee. Committee chair Naomi Chesler, a professor of biomedical engineering at the University of California, Irvine and director of the Edwards Lifesciences Center for Advanced Cardiovascular Technology, suggested implementing something similar across the society’s four journals — Cellular and Molecular Bioengineering, Cardiovascular Engineering & Technology, Biomedical Engineering Education, and Annals.
“As a society, BMES has been leading the way in promoting diversity, equity, and inclusion in STEM fields from the longstanding Women in BMES social (now luncheon), to the Celebration of Minorities event and LGBTQ dessert hour,” Chesler said. “I’m proud that we’ve taken this next step in raising awareness of citation disparities and hopefully better recognizing the important work of women, Black and Indigenous people, and people of color in our field.”
Duma and Rowson pointed out that biomedical engineering, which has a higher proportion of women than most engineering fields, is a natural origin point for some of these efforts: If the field itself is more diverse, then — in theory — citation lists can be, too.
“It’s a little easier for us to take that step forward,” Rowson said.
The BMES editorial board developed a plan that gave authors some latitude in how to approach the issue. Though a citation diversity statement isn't required, it's encouraged, and authors who submit one have two options. The first is a general statement that acknoweldges bias in citation practices and states that the authors have worked to avoid it.
The second option, modeled after the statement suggested by the Neuron paper, is more detailed. In this version, the author would supplement the basic statement with a quantitative analysis assessing what percentage of that paper's citations credit research from underrepresented groups across gender, race, and ethnicity.
Pinning down those numbers can be challenging. Some journals and academic presses are beginning to offer authors the option to supply their own demographic information, which will make those metrics easier to track. For now, though, diversity analyses rely largely on public databases that correlate names with social media profiles, voter rolls, and other records that may list gender, race, or ethnicity. These sources aren't error-proof, and often fail to capture categories like nonbinary genders and mixed-race identities. And the analysis itself is cumbersome, as Duma discovered when he tried it for one of his own papers — “it takes days,” he said.
Authors who might be deterred by these factors can opt to just include the generic statement — which, Duma said, still achieves the primary goal of elevating the issue.
“What's important is that people are at least acknowledging it,” he said. “It starts the discussion and starts the process of becoming aware of the citation parity as part of larger discussions we’re having in the scientific community around diversity and inclusion.”
Bringing diversity to the bibliographies of academic manuscripts, so that they more accurately reflect who’s doing influential work, isn’t a solution to systemic inequalities in research. But by shining a light on disparities that are usually invisible, compelling authors and readers to reckon with them one manuscript at a time, it may help chip away at them.
“If people aren’t getting cited, that’s really going to affect whether they’re able to get funding and other resources,” Rowson said. “It’s more than a number — it’s influencing people’s careers.”