This week's Nature includes a letter with a provocative title.

Mysterious disappearance of female investigators
Darach Watson, Anja C. Andersen and Jens Hjorth

Don't expect CNN or Fox to add this to their lineup of "Where the Women At?" stories though—it's just another troubling example of discrimination.

The results of the first year's European Young Investigator (EURYI) awards are worrying. The awards provide 25 young scientists with up to €1.25 million to establish research teams in Europe. Only three of the 25 initial recipients were women, far below the percentage working in science at the targeted career stage. This was not because women didn't apply in sufficient numbers: nearly a quarter of the applicants were female. Rather, male applicants were twice as likely to succeed.

EURYI applications were first submitted to the relevant national research councils, who could nominate a specific number of candidates. This selection cut the proportion of women from one-quarter to one-fifth. Each national research council oversaw a drop in the number of selected women. In Spain, where nearly a third of the applicants were women, not one was nominated. The all-male Spanish list emerged with the highest success rate in the later European rounds, nearly three times the average. From 133 national nominees, European evaluation committees created a shortlist of 67, causing the largest drop in the proportion of women: 9.9±0.5% of men applying made the European shortlist, but only 4.7±1.4% of women did.

The working paper Evaluation of the EURYI Awards Scheme by L. Langfeldt and K. E. Brofoss (NIFU STEP, Oslo, 2005), commissioned by the European Science Foundation (ESF), with access to a limited data set including assessments of candidates at the European evaluations and sample applications, recently concluded: "the main problem remains that the scheme ... attracts far fewer female applicants". The report also suggested that "female applicants had a somewhat higher tendency to be filtered out" at the domestic level but that there was "no evidence of bias" at a European level.

However, the statistics are clear: the consistent attrition of women at each stage, and the large size of the sample, mean that women's lack of success cannot have occurred by fluke. The random chance probability of halving the female fraction from one end of the competition to the other is only 0.05%.

The 'leaking pipe' phenomenon (in which a disproportionate number of women leave the sciences at each career stage) is often attributed to a complex array of external factors that cause women to drop out. In this case, we believe we are seeing a leaking pipe in the stages of a single competition. Does this mean more straightforward explanations for the career leakage may be possible?

Without a detailed knowledge of applications and judging criteria, it is impossible to nail down the underlying reasons for the inequality in the awards. We would like to replicate the groundbreaking analysis of Christine Wennerås and Agnes Wold ("Nepotism and sexism in peer review" Nature 387, 341−343; 1997) at the European level, but the ESF has so far been unwilling to release the necessary data to us.

We consider that this attrition demands further independent scrutiny to uncover the cause.

I know that some will try to claim that it's merely a consequence of women's relative lack of ability to perform at higher level scientific levels, but that link to the Wennerås and Wold article (which should work for everyone, it's to a copy not at Nature) puts paid to that idea. The authors had access to the peer-review evaluations of grants submitted to the Swedish Medical Research Council, which include a score the reviewer gave for scientific competence. In these grants, women scored particularly poorly on competence compared to men, so the authors evaluated their real world performance more objectively, examining publications, journal impact scores, etc. The result was an eye-opener for me.

Did men and women with equal scientific productivity receive the same competence rating by the MRC reviewers? No! As shown in Fig. 1 for the productivity variable ‘total impact’, the peer reviewers gave female applicants lower scores than male applicants who displayed the same level of scientific productivity. In fact, the most productive group of female applicants, containing those with 100 total impact points or more, was the only group of women judged to be as competent as the least productive group of male applicants (the one whose members had fewer than 20 total impact points).

They also carried out a multiple regression analysis to try to identify the factors that were most predictive of the competency assessment by the grant reviewers.

Three out of six productivity variables generated statistically significant models capable of predicting the competence scores the applicants were awarded: total impact, first-author impact and first-author citations. The model that provided the highest explanatory power was the one based on total impact (r2=0.47). In all three models, we found two factors as well as scientific productivity that had a significant influence on competence scores: the gender of the applicant and the affiliation of the applicant with a committee member.

How much of an impact do you think applicant gender has on their competency evaluation? They were able to quantify that, and it's dreadful.

According to the multiple-regression model based on total impact, female applicants started from a basic competence level of 2.09 competence points (the intercept of the multiple regression curve) and were given an extra 0.21 points for competence. So, for a female scientist to be awarded the same competence score as a male colleague, she needed to exceed his scientific productivity by 64 impact points (95 per cent confidence interval: 35-93 impact points).

This represents approximately three extra papers in Nature or Science (impact factors 25 and 22, respectively), or 20 extra papers in a journal with an impact factor of around 3, which would be an excellent specialist journal such as Atherosclerosis, Gut, Infection and Immunity, Neuroscience or Radiology. Considering that the mean total impact of this cohort of applicants was 40 points, a female applicant had to be 2.5 times more productive than the average male applicant to receive the same competence score as he ((40+64)/40=2.6).

Gah. That's some leaky pipe. I feel like I have to apologize to all of my female colleagues now for benefitting from this bias, even though it's not entirely men's fault, directly: the article also mentions that other studies have found both male and female reviewers deprecate their evaluation of women's applications (there were relatively few female reviewers in this sample).

The authors have one suggestion for correcting the problem: end anonymous review and open up the process to more thorough inspection. I can see the virtue of anonymous review in that it makes it easier to criticize the work of the powerful, but it's clear it also masks patterns of chronic abuse. They've convinced me, at any rate.