A population model for academia: Case study of the computer science community using DBLP bibliography 1960-2016

Abstract

The academic community has seen a tremendous growth in number of authors and publications in recent times. Most previous studies of the academic community, whether on individual activity, productivity and influence, or collaboration patterns and their implications, have been based on a flat and static view of the system. However, the academic community resembles a dynamic growing population, with entry and exit of authors. In this paper, we study this systemic inflation by proposing a population model for academia. We use a generalized branching process as an overarching framework, which enables us to describe the research community from an evolutionary and structural perspective. Further, the observed patterns allow us to shed light on researchers' life-cycle encompassing their arrival, academic life expectancy, activity, productivity and offspring distribution in the community. In our study, we used data from DBLP for a case study in the computer science community, although our methodology can be adopted in a systematic manner to any other research domain with sufficient publication records. We believe that the results can help academics and policy makers alike to better understand and evaluate the development and evolution of their respective academic communities.