Low publication output for female scientists

Analysis suggests that male physicists and mathematicians author nearly 50 per cent more papers than female colleagues

August 21, 2014

The average male physicist and mathematician authors nearly 50 per cent more papers than his female colleagues, an analysis of nearly a million preprints has suggested.

The study of 938,000 preprints on the arXiv open access repository, used mostly by researchers in physics and maths, was carried out by Emma Pierson, a statistician at DNA analysis company 23andMe and soon to be a Rhodes scholar in computational biology at the University of Oxford.

According to Ms Pierson’s account of her research on the FiveThirtyEight website, the proportion of authors who were identifiably female rose from 8.7 per cent in 1991 to 15.5 per cent in 2014. However, over that period the average male published 45 per cent more papers than the average female, and twice as many solo papers.

Ms Pierson suggests that women’s lower publication output and consequent lower number of collaborators could blight their careers since “being friends with a scientist who reviews a paper, grant or job application can provide a crucial bonus”.

One explanation for women’s lower output is that men are more likely to be principal investigators – who are typically authors on all of their labs’ publications. Women accounted for 17 per cent of authors overall, but only 12 per cent of PIs. But Ms Pierson also found that male PIs published significantly more papers than female PIs.

She told Times Higher Education that possible reasons for this were female scientists’ childcare responsibilities and their tendency to spend more time teaching. She also said that gender bias might make it harder for female scientists to get published.

Ms Pierson was surprised to find that women were more likely than men to secure first authorship of papers, indicating that they did most of the work. But she was unconvinced that this was a positive sign since “it might just mean women are underrepresented among middle authors”.

She said that the most “clearly positive sign” from the figures for female scientists was that their numbers were increasing – especially given that women are more likely to be first authors in fields where more women work.

“People just tend to like and relate to those similar to themselves, and I think it’s easier to feel comfortable as a woman scientist if you’re not always the only [one] in the room,” she said.


Times Higher Education free 30-day trial

Please login or register to read this article

Register to continue

Get a month's unlimited access to THE content online. Just register and complete your career summary.

Registration is free and only takes a moment. Once registered you can read a total of 3 articles each month, plus:

  • Sign up for the editor's highlights
  • Receive World University Rankings news first
  • Get job alerts, shortlist jobs and save job searches
  • Participate in reader discussions and post comments

Reader's comments (3)

This, of course, says nothing whatsoever about what matters, the quality of the papers, But it may say something about the dishonest practice of lab heads putting their names on papers regardless of whether they have made much contribution. The fact that bibliometricians continue to ignore the content of papers is what makes them a menace to academic standards.
Does this research take into account the leaky pipeline effect (which is known to be more prevelant for women)? Relatively speaking, more women will leave academia than men at various stages in the career. So, relatively, more women will stop publishing after obtaining their Ph.D., after a first postdoc position, etc. If this is not taken into account, women will appear to have a lower output per year, even if the women that do remain active have comparable output rates. To be clear: it may well be the case that, on average, female researches do have a lower publication rate than male researchers; but it's not clear to me that this conclusion follows directly from what has been studied here. Moreover, it may also be the case that women use the arXiv less than their male colleagues. (No suggestion why this might be the case, just something that should be checked before drawing conclusions.)
In response to comments -- both good points (I am the author of the original article on FiveThirtyEight). @David -- indeed, quality is very important (and impossible to get from my data, unfortunately). One proxy for quality is arguably citation counts, as I discuss in the article -- some studies have found that papers with female authors have gotten fewer citations. But, obviously, citation count is also an imperfect metric for quality. @SylviaW -- Women who stop publishing at all should not be included in my dataset when we break it down by year, because I did not include a scientist unless her name appeared at least once that year. But seniority and the leaky pipeline almost certainly play a role, as I discuss in the original piece. More careful analyses of publication rates also find that women publish at lower rates.

Have your say

Log in or register to post comments