Hierarchical Bayesian model to estimate and compare research productivity of Italian academic statisticians

A new method for measuring scientific productivity is proposed. Each researcher is initially associated with a cumulative score over time, reflecting the quality of the papers based on the journals in which they have published throughout their career. The second measure, an average speed over time from varying production speeds, is derived through the estimation of a two-level hierarchical Bayesian model for piecewise linear regression. These productivity indicators are validated and compared to other commonly used bibliometric indexes. The proposed method is applied to compare the productivity of females and males at different career levels in Italian academia, with a focus on statisticians. The study also contributes to the literature on the gender gap, showing that among those who remain at the lower levels of the university career hierarchy, women tend to have higher and more consistent scientific production over time compared to their male colleagues.