Table 5. Logit regression results for the probability of pursuing a postdoc training in a highly ranked university, after graduation.

	Odds ratio
Female	0.673*
	(0.140)
Other controls	✓
Pseudo R2	0.09

* p<0.10.

We restrict the sample to Ph.D.s who pursued postdoc training after graduation. N = 1,137. The coefficient is an odds ratio. Standard errors clustered around supervisors are in parentheses. Controls include Ph.D. demographic and predetermined characteristics, Ph.D. number of publications and involvement in applied projects, supervisor characteristics, labor market characteristics at graduation, university-research field fixed effects, and graduation-year fixed effects.