party differences in minority-related discourse in german election manifestos: a text analysis approach
My master thesis examined gendered dropout in biomedical scientific careers and asks whether the gender composition of a research field moderates the relationship between gender and the risk of leaving science, using bibliometric data from the PubMed Knowledge Graph and author gender inferred from first names with the Wiki GenderSort database. Thereby, it complements prior work on gender differences in science by focusing on the role of field characteristics for retention within biomedical research. Dropout is operationalized as publication inactivity over different time windows, and Cox proportional hazards models are used to estimate differences in dropout risk between male and female researchers. The results show a robust positive asso- ciation between being female and scientific career dropout. Across model specifications and dropout operationalizations, female researchers have a higher estimated dropout risk than male researchers. However, the size of this association becomes smaller once field, entry time, and gender composition at entry are included. In contrast, the effect of female share at entry and its interaction with gender are highly dependent on model specification and do not show a stable pattern. This article therefore finds clear evidence of gendered dropout, but only limited evidence that field-level gender composition systematically moderates this relationship. These findings suggest that women’s underrepresentation in science cannot be understood only as a problem of entry into scientific fields, but must also be studied as a problem of retention.