How multiple networks help in creating knowledge: evidence from alternative energy patents

Different from most prior research studying the influence of networks on scientific knowledge creation by taking on a single social-based perspective, this study offers an integrated frame based upon both social-based and knowledge-based views. This study incorporates quantitative bibliometric methods of analysis and network analysis. Based on 51,972 alternative energy patents in 1998-2012, we build knowledge networks composed of technological classifications and social networks capturing collaborations. Based on patent citation networks, we use citations in every period to calculate researcher's knowledge creation performance. We hypothesize that centrality of researchers in the knowledge network and their knowledge creation display an inverted-U relationship. What's more, we propose that this curvilinear relation is moderated by their structural holes and centrality in the collaboration network. Based on the negative binomial model with fixed effects and robust tests, results indicate that (1) locations of the researchers in knowledge network contribute more to knowledge creation when their centrality is moderate, rather than high or low; (2) at a moderate level of knowledge network centrality, the researchers, who occupy more central positions or span more structural holes in collaboration network, will benefit more than the ones who do not. The findings of this research, besides having implications for research on knowledge networks, have implications on scientometric indicators, multiple networks study, and knowledge creation process.