University patenting activities and their link to the quantity and quality of scientific publications

Abstract

Integrating data from three independent data sources––USPTO patenting data, Shanghai Jiao Tong University’s Academic Ranking of World Universities (ARWU) and the Times Higher Education Supplement’s World University Ranking (WUR), we examine the possible link between patenting output and the quantity and quality of scientific publications among 281 leading universities world-wide. We found that patenting by these universities, as measured by patents granted by the USPTO, has grown consistently faster than overall US patenting over 1977–2000, although it has grown more slowly over the last 5 years (2000–2005). Moreover, since the mid-1990s, patenting growth has been faster among universities outside North America than among those within North America. We also found that the patenting output of the universities over 2003–2005 is significantly correlated with the quantity and quality of their scientific publications. However, significant regional variations are found: for universities in North America, both the quantity and quality of scientific publications matter, but for European and Australian/NZ universities, only the quantity of publications matter, while for other universities outside North America and Europe/Australia/NZ, only quality of publications matter. We found similar findings when using EPO patenting data instead of USPTO data. Additionally, for USPTO data only, the degree of internationalization of faculty members is found to reduce patenting performance among North American universities, but to increase that of universities outside North America. Plausible explanations for these empirical observations and implications for future research are discussed.

Appendix

Following Trajtenberg (1990), a linear weight was used, with the citation-weighted patent count (WPC) in year t being

${\text{WPC}}_{t} = \sum\limits_{i = 1}^{{n_{t} }} {\left( {1 + C_{i} } \right)}$, where n_t = number of patents issued to the university in year t for the years 2003–2005, and C _i is the number of citations received by each patent i up to the year 2006.

This is a somewhat crude approximation of the true citation-weighted patents count, for two reasons. Firstly, truncation bias means that citations to more recently issued patents are under-represented. Secondly, citations received by patents typically peak 4–5 years after the patent is issued (Mowery and Ziedonis 2002). Since our patents are those issued between 2003 and 2005, and data availability restricts our citation data to 2006, we have captured only a small fraction of the citations that will eventually be made to the patents in our database.

Table 6 Distribution of leading universities by region and country (n = 281)

Table 7 Estimated regression model of factors affecting citation-weighted patent activity

Table 8 No. of patents by leading universities versus all USPTO patents, 1976–2005