This article discusses the nature and structure of scientific collaboration as well as the association between academic collaboration networks and scientific productivity. Based on empirical data gathered from the CVs of 73 researchers affiliated with an academic research network in Canada, this study used social network analysis (SNA) to examine the association between researchers' structural position in the network and their scientific performance. With reference to Granovetter's and Burt's theories on weak ties and structural holes, we argue it is the bridging position a researcher holds in a scientific network that matters most to improve scientific performance. The results of correlation scores between network centrality and two different indicators of scientific performance indicate there is a robust association between researchers' structural position in collaboration networks and their scientific performance. We believe this finding, and the method we have developed, could have implications for the way research networks are managed and researchers are supported.