Sir

In the Commentary article “Does gender matter?” (Nature 442, 133–136; 2006), Ben A. Barres cites our article pointing out that the first round of the US National Institutes of Health (NIH) Pioneer awards was carried out in a way that would have predicted a bias against selection of women (M. Carnes et al. J. Womens Health 14, 684–691; 2005). Indeed, no women were selected in the first year, so when 43% of the second year's winners were women we examined the process again to see what had changed.

We identified several differences, including changes made by the NIH, that would predict a decrease in the activation of automatic gender stereotypes that may have discouraged women from applying and disadvantaged women applicants in the first round.

First, a reduction in the number of applicants (from 1,300 to 840) and greater familiarity with an application process that was no longer new may have reduced time pressure on the reviewers.

Second, the NIH removed the repeated mention of the need for applicants to engage in 'high-risk' research; we believe that this terminology encouraged male and discouraged female applicants. Similarly, the emphasis on 'intrinsic' leadership abilities and 'potential' of the scientist was removed, in favour of an emphasis on the scientist's research.

Third, there was a much higher proportion of women in the applicant pool, which may have been related to the change in language (26% in phase 1 and 35% in phase 2 in 2005, compared with 20% and 10% in 2004). There was also a greater proportion of women on the review panel: 44% in 2005, compared with 6% in 2004.

Fourth, the presence of accomplished women scientists on the review committee provided a positive role model for applicants.

Finally, women were specifically encouraged to apply — a particularly significant factor in the context of the outcry in the scientific community following the absence of women in the first round.

We applaud the NIH for taking an evidence-based approach. Regardless of the gender composition of the group selected in the forthcoming third round, removal of conditions that are known to activate automatic gender stereotypes ensures that the best science will be supported, regardless of the sex of the scientist.

See Nature 442, 510 (2006) for other letters on this topic. Readers are encouraged to add their comments on the Nature News Blog at: http://blogs.nature.com/news/blog/2006/07/ does_gender_matter.html