Post Hoc Intervention as a Remedy for Gender Bias in Grant Application Reviews

The Swedish Research Council is a Swedish government agency and one of the major research financiers in Sweden. As such, its operations are governed by legislation and governmental instructions and policies, which prohibit discrimination and include equal treatment. Consequently, the Swedish Research Council has for a large number of years been working actively against gender bias in their reviews of grant applications. We use data published by the Swedish Research Council to analyze potential gender bias in reviews of grant applications. We can show using several multiple linear regression models, that after controlling for a number of other factors there is still a significant gender bias in the grants awarded, at least for some review panels. We explore the possibility of mitigating this bias by using statistical methodology through the post hoc intervention, GIIU. We argue that GIIU is applicable to the dataset and could be used to mitigate the bias. However, we believe that a post hoc intervention, such as GIIU, would be more effective if it were implemented earlier during the review process on the more detailed data available to the research council.