A language analysis of titles and abstracts in more than 100,000 scientific articles found that papers with first and last authors who were both women were about 12% less likely than male-authored papers to include sensationalistic terms such as ‘unprecedented’, ‘novel’, ‘excellent’ or ‘remarkable’. The study, published in The BMJ1, also found that papers missing such words garnered significantly fewer citations.
Researchers tracked 25 positive terms in clinical-research articles published between 2002 and 2017, and input the authors’ names into the Genderize database to estimate their genders. The team then created models that compared the citation rates and word choice of articles published in the same journals in the same year with the same subject keywords.
The articles in each comparison were presumably of similar quality, but those with positive words in the title or abstract garnered 9% more citations overall, and 13% more citations in high-impact journals.
The relative reluctance of female authors to use self-flattering words could contribute to a gender gap in citations and impact, says lead author Marc Lerchenmueller, an economist at the University of Mannheim in Germany and the Yale School of Management in New Haven, Connecticut. In the big picture, he says, the results should encourage scientific authors and editors to think about word choice and its consequences. “Scientists should discuss whether using such sales terms is a disservice to the scientific enterprise,” he says.
The discussion seems to be becoming more important: the analysis also found that such self-flattering words were 80% more common in 2017 than in 2002. Lerchenmueller notes that this time period marked an explosion in the number of published articles. “Authors are trying to present research as favourably as possible to attract attention,” he says.
At this point, it’s impossible to pinpoint exactly why male and female authors would take a different approach to promotional language, Lerchenmueller adds. He points to decades of studies suggesting that women are more likely than men to face a backlash from peers and society when they stray beyond stereotypical norms. Women who have been chastised in the past for being too forceful or boastful might edit themselves and tone down their language, he says. Sensationalistic words could also get added or removed during the editorial process — and Lerchenmueller thinks this possibility warrants closer examination.
This relative lack of inflated language in female-authored papers echoes a 2019 experimental study published by the National Bureau of Economic Research2, showing that women gave themselves relatively poor marks in interviews, performance reviews, job applications and other settings. “We found a large and robust gender gap in self-promotion,” says Christine Exley, a business-administration researcher at Harvard Business School in Boston, Massachusetts. In one measure, women were less likely to describe their performance favourably when selecting from a list of potential adjectives that ranged from terrible to excellent. Exley notes that in the experimental setting, women should have felt no fear of backlash due to over-hyping themselves, but the gender gap still persisted.
Lerchenmueller feels that his study touches on some important philosophical questions about the power and meaning of words. “Is language a mirror of society, or does it shape society?” In the world of science, he says, language seems to both reflect and promote bias — and female researchers are experiencing the consequences.
Nature 578, 328 (2020)