The Young and the Old, the Fast and the Slow: A Large-Scale Study of Productivity Classes and Rank Advancement

TL;DR

This large-scale study of Polish STEMM scientists reveals that younger promotion age and faster promotion speed are associated with higher current productivity, highlighting the importance of biographical factors in academic advancement.

Contribution

The study introduces a novel methodological approach using promotion age and speed, providing new insights into productivity patterns across career stages.

Findings

Younger promotion age correlates with higher current productivity.

Faster promotion speed is associated with increased productivity.

Patterns are consistent across disciplines and career levels.

Abstract

We examined a large population of Polish science, technology, engineering, mathematics and medicine (STEMM) scientists (N = 16,083) to study rank advancement and productivity. We used two previously neglected time dimensions - promotion age and promotion speed - to construct individual biographical profiles and publication profiles. We used a classificatory approach and the new methodological approach of journal prestige-normalized productivity. All scientists were allocated to different productivity, promotion age, and promotion speed classes (top 20%, middle 60%, and bottom 20%). The patterns were consistent across all disciplines: scientists in young promotion age classes (and fast promotion speed classes) in the past were currently the most productive. In contrast, scientists in old promotion age classes (and slow promotion speed classes) in the past were currently the least productive. In the three largest disciplines, the young-old promotion age productivity differential for associate professors was 100-200% (150-200% for full professors); and the fast-slow promotion speed productivity differential for associate professors was 80-150% (100-170% for full professors). Our results were confirmed by a regression analysis in which we found odds ratio estimates of membership in top productivity classes. We combined data collected from the national register of all Polish scientists and scholars (N = 99,935) and publication metadata on all Polish articles indexed in Scopus (N = 935,167).