Access to knowledge is crucial for the development of new and innovative ideas. In the last 20 years, an increasing number of scholars have explored the role of scientific collaborations in connecting a diverse pool of knowledge and in reducing the cost of access to information, as well as their effect on scientific productivity. Along this line of research, in this paper we explore the importance of intellectual interactions among the scientists at the University of Florida, the state’s flagship university, as channeled by their network of publication co-authorship between 2008 and 2014 (approximately 5,000 nodes and 20,000 edges). Assessing the significance and the magnitude of the effect of interactions on academic productivity contributes to the literature on social interactions and scientific networks, and provides policy implications encouraging academic collaborations. An inherent selection effect arises in the formation of co-authorships. We use data on spatial and social proximity of investigators to account for this endogeneity problem. After controlling for endogenous co-authorship formation, unobservable heterogeneity, and time varying factors, we find a positive impact of intellectual collaboration on individual performance, measured in terms of success rate for grant funding.