The organization of scientific collaborations

Abstract

Based on empirical analysis of 53 multi-institutional collaborations in physics and allied sciences, we find that generalizations about the essentially informal and collective social organization of collaborative projects in science stem largely from a narrow analysis of high-energy particle physics experiments. Cluster analysis reveals that the variety of organizational formats of collaborative projects can be grouped into four types, ranging from bureaucratic to participatory. Except for particle physics, which is overwhelmingly participatory and non-bureaucratic, the membership of the other three types is mostly cross-disciplinary. The four-fold typology discriminates collaborative projects with respect to their technological practices. The structure of leadership is related to the character of interdependence in data acquisition, analysis, and communication of results: greater interdependence leads to decentralization of leadership and less formalization. We conclude that extrapolation of the organizational characteristics of particle physics to scientific collaborations in general is unjustified.