Beyond bean counting : using patent information to investigate inventive productivity in academia

Abstract

Contemporary universities have multiple missions: to provide education, to conduct research and to contribute to innovation and growth through so-called third stream activities. Examples of the last include both commercial undertakings, such as patenting, licensing, and new business creation, and non-commercial accomplishments, such as social outreach. Since many policymakers and other stakeholders have sought to further promote third stream activities, there is an ongoing debate about how to appropriately measure and evaluate such activities. Rather than the issue of individuals’ or organisations’ scientific productivity, the promotion of third stream activities highlights, among other things, how to assess inventive productivity, which is broadly understood here as the inventive output of individual academics, research groups or organizational units during a specific period of time. In countries like the UK and the Netherlands, where research evaluation frameworks have been introduced that include assessment of third stream activities, inventive productivity has typically been linked to the number of patent applications that have been filed or granted in a given year. However, various studies have questioned if the use of such volume-based measures appropriately represents qualitative and process-related aspects of invention and patenting processes.

Building on this previous research, the present thesis asks how measurement of inventive productivity in academia can be undertaken. This overarching question is addressed in four papers, which are positioned at the intersection of research streams covering academic commercialisation, research evaluation and the use of patent-based measures in innovation studies. Data on academic inventors and patents from a single faculty medical research university are used to explore different methods for measuring inventive productivity and their consequences.

Three papers elaborate on the use of available patent information for measurement of inventive productivity. Specifically, inventive productivity is operationalized based on an alternative definition of patent counts (Paper I), measures of patent survival (Paper II) and routes for patent transfer (Paper III). The findings suggest that existing patent information, when analysed longitudinally, can be used to construct more representative measures of inventive productivity compared with single-period patent counts. Paper IV extends the discussion on measurement of inventive productivity by problematizing whether volume-based patent measures capture behavioural differences among academic inventors in patenting processes. Through the tracing of inventors’ experiences at the process level using interviews, behavioural similarities and differences between individual academic inventors of varying inventive productivity emerge. These findings suggest that established categorisations of academic inventors as occasional and serial inventors, based on e.g. single-period patent counts, can obscure meaningful differences in individuals’ behaviours.

In summary, the realization of a more tailored and targeted innovation support and assessment system would benefit from further methodological and organizational development. To move beyond bean counting based on single-period patent counts, this thesis proposes that policymakers and other stakeholders further develop volume-based measures using existing patent information and also complement such data with process-related information concerning inventor behaviours and views.