Is there gender bias in peer review?
When a scientific paper is submitted for publication, an editor at the journal will usually select - a number of authority figures in the relevant field and ask them to “peer review” the manuscript. So, who gets picked to do the job? Specifically, are male scientists more likely to be asked compared with women? Markus Helmer let Chris Smith know about the issue...
Markus - We were looking into, if there is any kind of bias in the way that editors or reviewers have chosen. In our specific case, we were looking into if gender matters for how reviewers are chosen. We use one specific series of journals, the Frontiers series of journals. They do publish alongside each paper who was the reviewer of that paper and who was the editor of that paper. They have been existing since 2007 and over the years, they have published around 40,000 articles. These 40,000 articles, they were handled by 9,000 editors and around 43,000 reviewers. So altogether, that gives a huge dataset in which we can study the behaviour in the peer review process and look for potential bias.
Chris - How did you interrogate that dataset? What questions did you ask of it?
Markus - We were interested in, if women are unrepresented beyond expectation. It’s a well-known fact that there are just less women than men in science. If there are less women than expected, given that just numerical under-representation and that seems to be the case. We saw that for authors and reviewers while for the editors, the actual number was within the expectation.
Chris - What was the other question you asked?
Markus - The other questions we asked was if there's any bias in the way editors select the reviewers. What we found is that it seemed that editors are for both genders, preferentially chose reviewers of their same gender.
Chris - That’s interesting, isn’t it? So we have got a bias but we’ve got both people being biased in favour of their own sex.
Markus - That's true.
Chris - By how much?
Markus - It appeared that this bias is present for both editor and genders, but it appears also that there's a difference in the way female editors prefer their own gender than the way male editors prefer their own gender. It seems that the same gender preference is much more widespread among male editors. Whereas for female editors, it appears to be restricted to relatively few who were then highly preferential.
Chris - Why do you think we’re seeing this because this is what we dub ‘homophile’ – preference for your own gender?
Markus - Yes. So it appears that the same gender preference is something that might be just human nature in a sense. So you see that in small children already, for example, the way small children form friendship networks, in work environments. It really seems to be a widespread pattern of human behaviour.
Chris - Were you surprised by what you found?
Markus - We were surprised by what we found, yeah. We did expect to find that women are underrepresented but on top of that underrepresentation, there is also the same gender preference. We did not expect that. This is potentially also important because it seems that the number of women in science increases over the years. Very slowly, but it does increase. In a couple of years, even when numerical equity might be reached between the genders, the same gender preference might still be there. So, by just improving the numerical ratio between men, women, the gender bias as a whole will not go away.
Chris - In other words, the system is a bit broken and people’s approach to the system is wrong and if we top up the system with women to achieve parity, it will still relapse to type if we leave it alone anyway.
Markus - It appears so, yeah. So this same gender preference would have to be tackled on top of the numerical underrepresentation.