In your Editorial “All things equal” (Nature 437, 296; 2005) and Special Report “Small steps towards campus child care” (Nature, 446–447; 2005), much was made of the need for women scientists to have access to good child care if they are to succeed. However, this recent attention to child care in the scientific workplace merely addresses a symptom, rather than a cause, of under-representation of women in science.

Childless women and those with children have strikingly similar patterns of salary disparity and lag in achieving tenure and promotion compared with men.

As your report highlights, nations differ in child-care facilities — but they all share a shortage of women scientists, particularly at higher levels. Furthermore, the proportion of women in different sub-disciplines varies dramatically, but child-care availability is no different for a microbiologist or an engineer.

We suggest that lying behind the paucity of women in science is an unconscious bias in evaluating the sexes. Research shows that both men and women tend to overrate men and underrate women in competence, particularly when women are in a non-traditional field such as science (V. Valian Why So Slow? MIT Press, Cambridge, MA, 1998). For example, when the heads of 147 psychology departments were sent fictitious resumés of prospective faculty members and asked to name the rank — assistant, associate or full professor — to which the candidate would be appointed in their department, the recommended rank was higher if the resumé had a male name than if the same qualifications had a woman's name attached (L. S. Fidell in Woman: Dependent or Independent Variable? 774–782, eds R. K. Unger and F. L. Denmark, Psychological Dimensions, New York, 1975).

More recently, women had to produce twice as many scientific papers of equivalent quality as men to be considered equally competent in a Swedish Medical Research Fellowship postdoctoral programme (C. Wennerås and A. Wold Nature 387, 341–343; 1997).