The well-known 1997 study of Wennerås and Wold on gender bias in grant allocation found a ‘male bonus’ of three Science papers: women would need three additional publications in Science to get a similar score for their competence as men. The paper has been cited many times but was also faced with criticism and with demands for reanalysis of the data of W&W. This study is an attempt to do so, and after field-normalizing the performance scores and panel-standardizing the review scores, we found partly different results. We also did find gender bias in the competence scores, but the male bonus we found is only one-fifth of the male bonus reported by Wennerås and Wold. The same holds for the ‘reviewer affiliation (or nepotism) bonus’ and for the ‘recommendation letter bonus’. We have expanded the original study by investigating whether there is any bias in the grant decisions, an issue not discussed by Wennerås and Wold, and apart from two panels, this does not appear to be the case. Here our findings are in line with more recent studies suggesting that gender bias in the panel scores not necessarily lead to gender bias in the grant allocation. However, the two panels where we identified bias in grant decisions accounted for a substantial share of the female applicants, suggesting that research into gender bias in grant allocation should be conducted at the panel level.
Sandström et al. (Tue,) studied this question.