Growth, Impact, Productivity, and Scaling in Research Institutions*

The past two decades have seen significant research into understanding scaling laws in biology, cities, and networks. We extend these observations to explore how research growth, productivity, impact, and collaborations scale with the size of research institutions by analyzing physics papers in APS journals (through 2017) and the Microsoft Academic Graph (MAG). The latter, one of the largest datasets of scholarly publications with the most complete citation data available, has never before been used to analyze the scaling laws of research activity.  In both datasets we discover that while mean productivity per researcher is independent of institution size, a paper’s impact, the number of authors per paper, and the total number of collaborations per researcher, especially within an institution, all scale positively with institution size. Because the number of authors correlates with the paper’s impact, affiliation size appears to partly affect a paper’s impact by improving the ability for researchers to interact. Large institutions, in other words, provide an economy of scale to improve collaborations and research overall. We also discuss initial mechanistic models to explain the scaling laws we observe.