Participant researchers in the 2nd edition of the Serrapilheira Scientific Retreats, in 2018. Photo: Diego Padilha

In Brazil 57% of undergraduate students are women, 55% of introduction to scientific research scholarships have been granted to women and only 36% of those granted with productivity scholarships for CNPq research (scholarships granted to PhDs with a sound academic background) are women. The data show the hurdles female researchers are faced with in reaching top careers – this is not just chance, as women account for only 30% of scientists.

This is why Serrapilheira especially encouraged women applicants for the 3^rd Research Support Public Call, opened in 2019. The most recent invitation for research applications reflected the actual situation when it comes to Brazilian science: 56,6% of the applicants were men, 42,6% were women and 0,8% did not inform gender or ticked the “other” box.

After the selection process, the issue turned out to be even more complex: out of all successful candidates, only 29% of applicants were women, leading to the question: could there be a bias in application assessment, despite secured applicant anonymity throughout the process? In this context, for the 2^nd Public Call for research support a study was carried out during the No-Budget Science Hack Week, a project intensive workshop seeking to solve science problems at low cost, held in July in Rio de Janeiro, Brazil.

The group in charge of the study summarized the 2^nd Public Call data, examined the grades given to both the projects as well as the applicants resumés by gender and reached the conclusion that there was no difference in performance between men and women during the evaluation.

Then, an analysis was conducted by area of work. It was verified that in the “harder” sciences, there were less women submitting applications. In Computer science, for instance, there were only four women among the 52 applicants (7,7%); in Physics, 9 out of 51 (18%) were women. Thus, the selection process ended up reproducing the gender effect found in the various scientific areas. The only area where this trend was not observed was in life sciences, as 49% of registered applications were submitted by women (and 56% of successful applications were female scientists).

This is a well-known statistical phenomenon called the Simpson’s paradox: when observing data groups separately, they show a certain trend, however, when the data is combined the trend disappears. So, the study reached the conclusion that there is no gender bias in the Public Calls, but rather, to achieve more gender balance in the selection of applicants, what we need is more female candidates submitting applications in areas such as mathematics, physics, and computer science.

The study was conducted by researchers PhD Aline Melina Vaz; Ana Elisa Ribeiro Orsi (both form the University of São Paulo USP); PhD Janaína Dutra Silvestre Mendes (National Cancer Institute – INCA-MS); Karina Lobo Hajdu and Raiane Ferreira (both from the Federal University of the State of Rio de Janeiro – UFRJ), who participated in the No-Budget Science Hack Week.

Take this opportunity to get to know our “Best practices guide for diversity in science” drafted by a committee of experts in this issue.