Sex differences in research productivity among doctoral students in Sweden: A quantile regression approach

• A consistent sex productivity gap could be observed among doctoral students in Sweden. • Quantile regression enabled an analysis of the sex productivity gap over the whole productivity distribution. • The sex productivity gap increases at the higher end of the productivity distribution. • A large dataset allowed for a comparative design including the natural sciences, technology, medicine, and the social sciences. This study examines the sex productivity gap among doctoral students in Sweden using a comparative design. It focuses particularly on how the gap increases at the higher end of the productivity distribution, with men consistently publishing more than women. The study is based on a large dataset of 10,804 doctoral students who graduated between 2010 and 2019 in the research areas of the natural sciences, engineering and technology, medical and health sciences, and the social sciences. By applying multiple quantile regression analysis, we were able to conduct a nuanced analysis of the sex productivity gap across the whole productivity distribution. Results indicate a consistent productivity gap by sex across all research areas and that the gap increases towards the higher end of the distribution, i.e., the sex differences in productivity increase among the top performers. However, the comparison of research areas revealed some heterogeneity. In engineering and technology, the increasing sex gap levels off in the middle of the distribution but takes a leap at the extreme tail. In the social sciences, the gap peaks just before the extreme end of the distribution and then starts decreasing. The natural sciences and medical and health sciences show a more gradual increase in the gap towards the higher end. Taking into account the Swedish context – with its widespread adoption of the collective model of doctoral education and the thesis-by-publication format – our main conclusions are: (1) there exists a consistent sex productivity gap across all studied research areas, and (2) the increasing sex gap at the upper end of the productivity distribution, commonly seen in later career stages, can already be observed during doctoral studies.