How Social Networks Affect Scholars’ Productivity and Influence
Dimitrina Dimitrova, Barry Wellman, Tsahi HayatTo get your work known, it pays to cultivate large and diversified networks of fellow scholars. This is not a feel-goodism: it is based on a research we have done on a large scholarly network in Canada. We found a big difference between how you get your research done—by working in cohesive teams—and how you get your work known—by being part of multiple diversified networks of scholars. Our conclusions are based on a study of a large Canadian network of scientists, social scientists, and humanists. Although it is based on a single study, we believe that its face validity makes it instructive for scholarly communication. Here is what we found out when we studied the GRAND scholarly consortium of over 100 Canadian computer scientists, social scientists, and humanists. The consortium required each scholar to be a member of at least three research projects that contained members from multiple universities and diverse disciplines. These requirements gave us an excellent opportunity to see how the scholars’ social networks affected their productivity and influence. We surveyed 101 scholars twice, once near the start of the GRAND consortium in 2010 and several years later in 2014, and we also were participant observers in many GRAND activities. We analyzed the effects on productivity and influence of different types of scholarly connections, both in-person and online: acquaintanceship, advice, and co-authorship. We studied whether the networks were cohesive, if the scholars were central in their research networks or linked to central players, and whether their work had more opportunities to be disseminated.
We found that scholarly productivity and influence were associated with different social network properties. Densely knit, bounded networks were the most strongly associated with producing publications. Dense networks created and maintained long-term trust, especially among people with similar disciplinary backgrounds, leading to greater output. Yet, dense networks were not associated with scholarly influence, as measured by citation counts. It was the effective size and diversity of the scholarly networks that was associated with their influence. Diverse, large networks built awareness and spread the word further afield, across interdisciplinary and interuniversity boundaries. Thus, the very densely knit boundedness associated with productivity may have limited the diffusion of influence of the research they produced. Our conclusion is that if you want to get it done, have a small, tight network; but if you want to get the word out, also connect to large, diverse networks.