Making it to the PhD: Gender and student performance in sub-Saharan Africa

Abstract

Women's underrepresentation in science, technology, engineering, and mathematics (STEM) impedes progress in solving Africa's complex development problems. As in other regions, women's participation in STEM drops progressively moving up the education and career ladder, with women currently constituting 30% of Africa's STEM researchers. This study elucidates gender-based differences in PhD performance using new survey data from 227 alumni of STEM PhD programs in 17 African countries. We find that, compared to their male counterparts, sampled women had about one less paper accepted for publication during their doctoral studies and took about half a year longer to finish their PhD training. Negative binomial regression models provide insights on the observed differences in women's and men's PhD performance. Results indicate that the correlates of publication productivity and time to PhD completion are very similar for women and men, but some gender-based differences are observed. For publication output, we find that good supervision had a stronger impact for men than women; and getting married during the PhD reduced women's publication productivity but increased that of men. Becoming a parent during the PhD training was a key reason that women took longer to complete the PhD, according to our results. Findings suggest that having a female supervisor, attending an institution with gender policies in place, and pursuing the PhD in a department where sexual harassment by faculty was perceived as uncommon were enabling factors for women's timely completion of their doctoral studies. Two priority interventions emerge from this study: (1) family-friendly policies and facilities that are supportive of women's roles as wives and mothers and (2) fostering broader linkages and networks for women in STEM, including ensuring mentoring and supervisory support that is tailored to their specific needs and circumstances.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 49 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.