Scientific Sexism at the University of São Paulo

Livia Oliveira-Ciabati and colleagues examine scientific output at the University of São Paulo. They include 3,067 of the university’s 8,325 faculty and compare publications, citations and H-indices using Web of Science. Names serve as a proxy for gender in the study. Less than half of the faculty enter the analysis, making database coverage part of the result’s interpretation.

Gaps at the top of a metric

Men have higher averages on several publication and citation measures, and they are especially concentrated among faculty with high H-indices. The authors describe the widening as a scissors effect and consider years of employment. An H-index accumulates publications and citations. It is sensitive to time, discipline, collaboration and research resources, so the distribution cannot directly reveal differences in individual ability.

Among the 100 people with the highest H-indices in the analysed sample, 83 are classified as men. That denominator is the top of the observed database sample. Reading the share alongside years in post and field makes it possible to separate entry into the database, opportunities to accumulate citations and the kinds of work the metric is best equipped to retain.

The patterns raise questions about mechanisms. Which projects gain early funding, whose work appears in the author list, and who does teaching or service that these metrics barely register? The study’s future citation curves are projections. Database coverage and classification by name also constrain whose experience the analysis can distinguish.

Who makes it into the ranking?

Figure 2 brings together several populations: faculty with records, faculty in different bands of the index, and citations over time. Start with the change from the university roster to people with usable records before comparing who reaches the upper bands. The filters are different. Coverage of journals and languages influences which work enters the database. Career length and access to collaborators influence who can accumulate visible output. Table 3 groups faculty by years at the university, and the imbalance remains in the upper bands. Time at one employer can still conceal differences in discipline or previous career paths.

Ask what the indicator can count. Maintaining a community partnership or repeatedly checking data may make a study stronger without producing citations attributable to the person doing that work. The paper describes the distribution of visible output. If funding panels subsequently allocate money and time using the same metrics, that distribution may shape the next round of research. Testing such a feedback would require following actual allocation decisions. A projection of citation trends alone cannot establish it.

Ross and colleagues’ study of scientific authorship offers a complementary comparison between people who contribute to research and people who receive credit. Read the two together to see how visible outputs are distributed and how they become attributable in the first place.

Original text