Measuring Intellectual Diversity in Groups, and Its Effects on Scientific Impact and Productivity

Abstract

I answer questions regarding the relationship between the intellectual diversity of an individual’s collaborations and their impact and productivity. I propose two new measures of intellectual diversity∶ coauthor diversity score and group intellectual diversity. Coauthor diversity score measures how different an author is from their immediate collaborators. Group intellectual diversity measures how diverse a group is as a composite of each members difference to every other member. These measures allow me to test hypotheses regarding the relationship between intellectual diversity and career outcomes. I conduct two case studies, using data from the Web of Science database, to examine the measures in biomechanical modelling and nanotechnology. I bootstrapped negative binomial regressions with squared interaction terms for diversity measures to test for relationships between individual and group diversities against career citations. Increasing diversity is associated with more citations until an maximum after which citations exhibit negative returns as diversity increases. This behaviour is consistent in both case studies. I use a bootstrapped linear regression to show that individual and group diversities have a significant effect on publication rates in biomechanical modelling, but not in nanotechnology.