I ended up at Nature Networks thanks to a post on my previous blog about a paper in TREE which claimed that the proportion of female authors was increased by double-blond peer review1. In a nutshell, I argued that there wasn’t enough data to show this. Well, things have progressed a bit since then, and now we have more data.
The whole analysis was based on one journal, Behavioral Ecology switching to double-blind reviewing, and claiming that the increase in proportion of female authors was due to the change in review system. Tom Webb and Rob Freckleton went back through the journal and looked at the data in a bit more detail2 (yes, I’m involved too, but my role was just to tidy up the analysis – the credit is theirs). They simply plotting the proportion of female authors against time. If double-blinding has an effect, there should be a visible change in the proportion of female authors just after the shift. Well, this is what they got:
Figure 1. The proportion of papers with a female first author in BE, volumes 1 (1990) to 18 (2007), with the years used by Budden and colleagues shown in blue. The vertical line is when double-blinding was implemented in the journal. Link
When I first saw that, I laughed. Who needs any complicated statistics? Visually, it’s clear that there is no change, just a constant upward trend.
For me this is pretty conclusive, that double-blinding didn’t have any effect (the same thing has been found in economics3), but of course that doesn’t mean it doesn’t have effects elsewhere, or that there might be other effects of double-blinding[4].
But I think the graph shows something more important: female representation in science is increasing, which I’m happy to declare is good for science. I wonder, though, how even this is – is it just in whole-organism biology, or does it spread across the sciences?
1 I decided to leave the typo in. Do with it as you wish.
2 T.J. Webb, B. O’Hara, R.P. Freckleton, 2008. Does double-blind review benefit female authors? Trends in Ecology & Evolution, online already
3 Blank, R.M. (1991) The effects of double-blind versus single-blind reviewing: experimental-evidence from The American Economic Review. Am. Econ. Rev. 81, 1041–1067. JSTOR
4 Footnote 3 gives evidence that agrees with the results of the model, which quite frankly is a surprise.
Bob, I’d be interested in learning about female last authors as well. That may provide an indicator of female PIs.
Well, you’ll just have to collect the data. 🙂
It would be interesting. There might be cultural differences about whether the senior author is last, but a signal will probably be visible. Hmm, that would be interesting – I would expect to see a lag in the change, but does it then increase as quickly?
Bob,
I suppose that the numbers would not be as encouraging. While a post-doctoral fellowship (which may result in first-authored publications) is sort of an equal opportunity position, an assistant professorsip, or more generally a PI condition, assuming that this is what make you become the last author on a publication, is still on the other side of the glass ceiling to many women.
I wonder how much of the glass ceiling is just the time lag (my guess – some but by no means all). But I know some women who have got fed up with the male-dominated culture of science, and who have left (for example one became a journal editor). That might change as more women get more senior positions, and change the culture.
A couple of years ago our university had a review of its research. One comment made about our department was that it was surprising how few senior women there were, and in particular that there were no women professors. I thought this was encouraging for academia as a whole if this was thought aberrant – especially as we are a maths department.
Massimo: authorship position on papers is not generally applicable across all fields and among all the groups within one field, so is not a completely reliable metric – though an indicator.
As mentioned in other discussions, the gender of authors is often unknown to editors, because intials are used, or the first name is not gender-specific. (For example, many Asian names are of non-apparent gender to a western editor.)
Although it is good news, Bob, that there is no evidence for a systematic gender bias in publications, I agree with the arguments in ref 2 about gender discrimination in general, from what I’ve observed over the years.
Double-brunette end of peer show?
I had to check to see what we had written in ref. 2…
Your fn. 4 (end of 2nd-to-last para) is broken.
Thanks, Richard. I’m still trying to work out how Textile thinks.
When you read counter-productive opinion pieces like this, it makes you not want to be particularly objective. But your own argument is compelling. Better to direct limited time and resources in combatting sexism in our disciplines to effective actions. What gets my goat in the linked article is this snide remark: “A world where women (and resocialÂized men) earn Nobel Prizes on flextime has no relation to reality.”
Any Nobel prize-winner with a person or more to run his or her household is already working on “flex-time”.
But I am getting off topic (and inexorably drawn back to this site despite other urgent things to do. Fie!)