Web mining science–industry linkages: extending typologies and rethinking knowledge transfer

Abstract Governments around the world have encouraged science-industry linkages (SILs) to support new technological discoveries and the foundations for a country’s competitiveness through knowledge transfer. However, when identifying SILs, researchers mainly rely on co-patenting, leading to a distorted picture of the prevalence and types of SILs. To analyse how web mining may help SIL researchers establish a more comprehensive understanding of SILs, we rely on a dataset of 995,438 scraped news items from Chinese universities. Our analysis includes a comparison of web mining and co-patenting, which we document in an extensive literature review as the key identification strategy of SILs in prior research. We find that web mining allows us to identify substantially more firms and research linkages between science institutes (universities and public research institutes), and that firms’ characteristics tend to be different. Thus, web mining may be considered as a complementary identification strategy to co-patenting. Further, we find that Chinese firms are embedded in the Chinese innovation system by a diverse set of linkages (visiting, donation, recruitment and teaching), suggesting a rich network of knowledge transfer channels. Finally, we show heterogeneity in the linkage formation patterns of science institutes, with a small group of universities being highly specialized in distinct linkages, and a large group of rather inactive universities. We grant open access to data resources to encourage research in these areas.