OLS regression for the whole data.

This paper analyzes 4,488 applications from a grant funding competition held in 2017 in Kazakhstan. The competition had a two-stage design: first, anonymous subject matter experts evaluated the applications’ scientific potential; then, open panels of local science managers made the final decisions. We analyze a range of bibliometric, institutional, and demographic variables associated with the applications and show that review scores account for only a small variation in success rates. The most important factor is the organizational closeness to decision-making. Gender also plays a role: we find that, net of academic merit, men and women investigators receive similar review scores, yet the panelists grant awards to men more often than to women. We further demonstrate that the gender gap emerges due to decisions made in a specific domain—Natural Resource Management.