The compounding costs of being female in academia: individual-based modelling of career progression and interventions

We develop a novel individual-based population-dynamics model of academic career progression, using 15 years of data from over 1000 academics from one university. Our model improves on previous models, which, by homogenizing career progression, may underestimate the costs of being female. We find multiple effects that compound to slow career progression for women. Women are hired at lower ranks than men, then face the sticky floor problem of getting stuck at the bottom for longer. Further, individuals in science, technology, engineering and mathematics (STEM) fields are promoted more quickly; this disproportionately affects women who are more prevalent in non-STEM fields. Our model reveals that age is more complicated than others have found with ordinary differential equation (ODE)-based models. Women are older when hired, and promotions favour the young; hence, age costs women more. Finally, the probability of attrition rises with years spent at the same rank, regardless of gender. Since women are promoted more slowly, they experience higher attrition rates. We also deploy our model to test possible interventions. We find that just hiring more women will not work. A more nuanced set of interventions is required. Gender parity will only be achieved at the highest ranks if hiring rates are jointly equalized across gender, academic rank and discipline.