From an evolutionary economics perspective, knowledge networks are self-organizing systems. Therefore, studying changes of these systems requires an understanding of how such changes are influenced by both the behaviors and characteristics of key individual actors and the network structure. We apply this perspective to a network of investigators (i.e. lead scientists) and a sample of 9543 Phase 2 cancer clinical trials during the period 2002–2012, in order to examine the structure and explore the dynamics of the clinical trial network. Using temporal exponential random graph models, we examine whether preferential attachment, multi-connectivity, or homophily drive the formation of new collaborative relations to knowledge translators - i.e. investigators with basic and clinical research knowledge. Our results suggest that despite some increased connectivity over time the network remains fragmented due to the considerably growing number of investigators in the network. This fragmentation limits opportunities for knowledge transfer to advance clinical trials. We find that homophily in research fields and investigators’ country of affiliation and heterophily in terms of publication output promote the formation of ties to knowledge translators. We find also that multi-connectivity increases the probability of tie formation with knowledge translators while preferential attachment reduces this probability.
Exploring network dynamics in science: the formation of ties to knowledge translators in clinical research
Bastian Rake; Pablo D’Este; Maureen McKelvey
Journal of Evolutionary Economics