The impact of research collaboration on scientific productivity

Based on the curricula vitae and survey responses of 443 academic scientists affiliated with university research centers in the USA, we examine the long-standing assumption that research collaboration has a positive effect on publishing productivity. Since characteristics of the individual and the work environment are endogenously related to both collaboration and productivity, this study focuses on the mediating effect of collaboration on publishing productivity. By using the two-stage least squares analysis, the findings indicate that in the presence of moderating variables such as age, rank, grant, gender, marital status, family relations, citizenship, job satisfaction, perceived discrimination, and collaboration strategy, the simple number ('normal count') of peer-reviewed journal papers is strongly and significantly associated with the number of collaborators. However, the net impacts of collaboration are less clear. When we apply the same model and examine productivity by 'fractional count', dividing the number of publications by the number of authors, we find that number of collaborators is not a significant predictor of publishing productivity. In both cases, 'normal count' and 'fractional count', we find significant effects of research grants, citizenship, collaboration strategy, and scientific field. We believe that it is important to understand the effects of the individual and environmental factors for developing effective strategies to exploit the potential benefits of collaboration. We note that our focus is entirely at the individual level, and some of the most important benefits of collaboration may accrue to groups, institutions, and scientific fields.