When there is no gender discrimination in promotion in science
When there is no gender discrimination in promotion in science: An econometric case study
Women face a ‘glass ceiling’ effect during their careers in many sectors, and science is no exception. A recent Nature’s Editorial (claims that “[i]n the United States and Europe around half of those who gain doctoral degrees in science and engineering are female – but barely one-fifth of full professors are women” (Nature Editorial 2013). Although it occurs – even prevails – in many sectors, the existence of a glass ceiling for women in science is particularly jarring since, as claimed by Robert Merton, science should follow the norm of universalism according to which all scientists are judged objectively on the base of their scientific merits, regardless of their “personal or social attributes” (Merton 1942). The evidence of inequality in female and male scientists’ careers has urged many scholars to study whether observing few female scientists in highly ranked positions results from gender discrimination in the promotion or differences in productivity, personal choices, or constraints (Sonnert and Holton 1995, Ceci and Williams 2011).
In a recent paper (Mairesse et al. 2020), we propose a study assessing whether gender matters for promotion to the highest career rank at the Institut National de Physique (INP) of the Centre National de la Recherche Scientifique (CNRS), France’s leading public research organisation dedicated to basic research in Europe.1 We consider a long list of promotion determinants that encompasses most of the relevant factors identified in the literature to explain academic promotions: research productivity, family characteristics, size of the professional network, mentoring, fundraising, technology transfer, and project management activities. We rely on ‘survival’ or ‘event history analysis’ statistical modelling and methods to estimate the impacts of these factors on female and male scientists’ promotion rates (Allison 1984).
We find that female and male physicists have the same promotion rate from junior to senior positions at CNRS, once controlling for research productivity. Nonetheless, our results also show that factors such as family characteristics, research responsibilities, professional network, mentoring and fundraising activities have different impacts on promoting female and male researchers.
There is no gender discrimination in promotion at CNRS in physics
When inquiring if women suffer from gender discrimination in promotion, we investigate the case of the INP – physics being known as a very ‘manly’ field. Within CNRS, INP covers theoretical physics, condensed matter, and optics, and quantum physics. CNRS is the ideal setting for studying gender disparity in promotion, since researchers are ranked on a promotion scale based on well-defined and publicly available criteria. Permanent CNRS researchers are classified into two categories: Chargé de Recherche (CR) and Directeur de Recherche (DR). Each of these categories has two sub-categories: CR2 and CR1, and DR2 and DR1, respectively. The promotion from CR to DR, as well as within the two sub-categories, is based on scientific activity and seniority. In our analysis, we focus on the promotion from CR to DR, primarily based on the assessment of scientific performance.
In June 2017, we launched an online survey with all the active researchers working at INP, and after three waves we obtained a rate of response of 58%. This allowed us to construct an individual panel data sample for 604 scientists (139 women and 465 men), with information on their family characteristics and research responsibilities, i.e. children and, if any, their dates of birth; head of a research team and, if yes, over which period. Data collected from the administrative archive of INP allowed us to reconstruct researchers’ career paths, i.e. date of entry at CNRS; career position (CR or DR); and, if promoted to DR, the promotion date. Overall, during our study period, 276 (45.7%) researchers were promoted to DR after 14.3 years, on average, and 328 (54.3%) researchers stayed at CR without being promoted to DR for 11.7 years. Using data collected from Elsevier’s SCOPUS database, we reconstructed the publication record and the co-authorships’ network of each researcher. We have also gathered information on mentoring and fundraising activities.
Figure 1 represents the ‘Kaplan-Meier’ survival function plotted separately by gender. It shows that the female physicists’ average survival rate as CR is higher in all periods than for their male colleagues, meaning that the average promotion rate to DR is lower for females than for males.
Figure 1 Kaplan-Meier survival function estimates by gender
Note: The x-axis represents the years elapsed from the recruitment as CR to the promotion to DR.
In a more sophisticated analysis applying Cox’s proportional hazard regression, we find no statistically significant difference between the female and male physicists’ promotion rates when including researchers’ productivity as the only regressor. We can conclude that there is no gender discrimination in the promotion from CR to DR. Promotion is essentially based on the assessment of scientific performance.
The fact that many factors affect differently the promotion of female and male researchers does not preclude the absence of discrimination in promotion
The finding that female and male INP physicists are treated equally when judged for promotion from CR to DR is compatible with that many factors have different, direct or indirect, impacts on their promotion. This is indeed what we find for family characteristics, research responsibilities, professional network, and mentoring and fundraising activities. Since these different factors are not measured in the same units, we cannot assess their relative importance for promotion. We thus rely on comparing them in terms of the promotion impact of publishing one additional paper. For instance, we ask: What is the equivalent number of publications that shows an equivalent effect as supervising one additional PhD student?
In Table 1 we summarise, in terms of ‘equivalent number of publications’, the impact of each of the promotion factors we find statistically significant.
Table 1 Significant impacts of promotion factors by gender
We see that each additional child in the family is equivalent to having 3.15 additional publications for male researchers; however, it does not affect the promotion rate of female researchers. Each additional collaborator is equivalent to having 0.33 publications less in male researchers’ publication portfolio, while it is not detrimental for female researchers. For male researchers, mentoring one PhD student is equivalent to publishing 7.18 articles. This is not the case for female researchers. Looking at the fundraising activity, having obtained an Agence Nationale de la Recherche (ANR) or an EU grant boosts female scientists’ promotion rate as much as having published a total of 67.0 articles more before being promoted to DR. This is not the case for male scientists. One additional year as head of a research team is equivalent to publishing 5.05 articles more for female and 2.09 for male scientists.
Our findings show that there is no statistically significant gender-related discrimination in the promotion from CR to DR of INP physicists in the recent 10 to 15 years. They thus confirm that the best basic strategy to earn promotion to DR, for women and men alike, is to invest in increasing their peer-reviewed article publications. However, the findings also show that men can leverage several other factors to increase their promotion rate, such as focusing on a limited number of collaborators, investing in mentoring activities, and dedicating time to leading a research team, while women have a limited set of options available. Indeed, the two options leading to higher promotion rates for them, besides publishing more, are in developing responsibilities as head of a research team and in coordinating ANR or EU projects.
Our results are overall encouraging. They demonstrate that in the French context – a context with consistent public investments for policies supporting gender equality – science can be for all. Nonetheless, major gender-related issues persist in the sector. In particular, we observe that female researchers have, on average, weaker publication records on their CVs than their male colleagues. The publication productivity gender gaps, or biases, or puzzle, in spite of abundant and diverse literature, remain open-ended questions (Cole and Zuckerman 1984, Fox 2005, Mairesse and Pezzoni 2015).
Allison, P D (1984), Event History Analysis Regression for Longitudinal Event Data, Sage Publications.
Ceci, S J, and W M Williams (2011), “Understanding Current Causes of Women’s Underrepresentation in Science”, Proceedings of the National Academy of Sciences 108(8): 3157–62.
Cole, J R., and H Zuckerman (1984), The Productivity Puzzle: Persistence and Change in Patterns of Publication of Men and Women Scientists, Vol. 2.
Fox, M F (2005), “Gender, Family Characteristics, and Publication Productivity among Scientists”, Social Studies of Science 35(1): 131–50.
Mairesse, J and M Pezzoni (2015), “Does Gender Affect Scientific Productivity? A Critical Review of the Empirical Evidence and a Panel Data Econometric Analysis for French Physicists”, Revue Economique 66(1): 65-113.