One area in which female scientists still have to confront sexism (Nature 468, 733; 2010) is in scholarly awards: women win fewer. Scientific societies must examine practices for selecting awardees.

The proportion of women receiving service or teaching awards in the past two decades is roughly equivalent to the proportion of women within the cohort-adjusted PhD pool in that discipline, but only half of these have won scholarly awards. Using data in the public domain on 13 disciplinary societies, we found that the proportion of female prizewinners in ten of these was much lower than the proportion of female full professors in each discipline (see also P. Leboy The Scientist 22, 67; 2008).

Our investigations reveal that practices for selecting awardees all tend to operate with few guidelines, minimal oversight and little attention to conflict-of-interest issues. Having women on selection committees helps recognition, but many panels have no female members and few have female chairs.

The pool of female nominees for an award is typically small. Their nomination letters tend to contain fewer descriptors of exceptionality, use stereotypically female adjectives (such as 'cooperative' and 'dependable') and mention personal details. Notices soliciting nominations, by contrast, tend to use language that fosters male images, such as 'decisive' or 'confident'. Unsurprisingly, bias thrives under these conditions.

To help push through key changes, the Association for Women in Science in Virginia, with funding from the US National Science Foundation, is collaborating with seven US science societies and the RAISE Project (http://www.raiseproject.org), which campaigns to raise the status of professional women through better recognition of their achievements.

We urge societies to use gender-neutral descriptors to broaden their candidate pool. Committees should establish selection criteria before reviewing nominees, and committee members need to understand the impact of implicit bias.