The present study addresses the ongoing debate concerning academic scientific productivity. Specifically, given the increasing number of collaborations in academia and the crucial role networks play in knowledge creation, we investigate the extent to which building social capital within the academic community represents a valuable resource for a scientist's knowledge-creation process. We measure the social capital in terms of structural position within the academic collaborative network. Furthermore, we analyse the extent to which an academic scientist's research specialization and ties that cross-community boundaries act as moderators of the aforementioned relationship. Empirical results derived from an analysis of an Italian academic community from 2001 to 2008 suggest academic scientists that build social capital by occupying central positions in the community outperform their more isolated colleagues. However, scientific productivity declines beyond a certain threshold value of centrality, hence revealing the existence of an inverted U-shaped relationship. This relationship is negatively moderated by the extent to which an academic focuses research activities in few scientific knowledge domains, whereas it is positively moderated by the number of cross-community ties established. Copyright © 2012 John Wiley & Sons, Ltd.