Social network analysis comparing researcher collaborations in two cardiovascular cohort studies

Abstract

The development of social network analysis techniques and comprehensive online bibliographic databases has led to studies of scientific co-authorship networks. This study compares collaboration among researchers associated with two epidemiological cohort studies, the Cardiovascular Health Study (CHS) and the Strong Heart Study (SHS), using journal article citations published from 1990 through 30 June 2011. Descriptive analyses of their publications and authors were computed, network graphs were produced, and network statistics were calculated. While similar in scientific methodology, these studies differed in ways that influenced researcher collaboration and entry into the networks. As a result, the number of unique authors for the CHS was three times greater than for the SHS (1,749 versus 571, respectively), driven by the CHS' larger number of research institutions and the higher influx of non-formally affiliated authors into its network. The CHS network had more ports of entry and also had more components, or collections of authors collaborating independently from the main network. In contrast, even at similar sizes, the SHS network was denser, indicating a greater cohesiveness than the CHS. The SHS had higher network centralization scores, suggesting the network contained relatively more authors that acted as gatekeepers. Indeed, the SHS had the majority of authors with the highest individual centrality scores, a measure of their collaborative prominence in the network. Differences in the studies' designs, populations, and organizations; funding mechanisms and policies; and awardee institutions likely explain these findings.This work provides a preliminary glimpse into understanding the mechanisms of collaboration among research studies. Copyright , Oxford University Press.