Using a bibliometric approach to support research policy making: The case of the Flemish BOF-key

In this paper, we describe the development of a methodology and an instrument to support a major research funding allocation decision by the Flemish government. Over the last decade, and in parallel with the decentralization and the devolution of the Belgian federal policy authority towards the various regions and communities in the country, science and technology policy have become a major component of regional policy making. In the Flemish region, there has been an increasing focus on basing the funding allocation decisions that originate from this policy decentralization on objective, quantifiable and repeatable decision parameters. One of the data sources and indicator bases that have received ample attention in this evolution is the use of bibliometric data and indicators. This has now led to the creation of a dedicated research and policy support staff, called Steunpunt O&O Statistieken, and the first time application of bibliometric data and methods to support a major inter-university funding allocation decision. In this paper, we analyze this evolution. We show how bibliometric data have for the first time been used to allocate 93 million Euro of public research money between 6 Flemish universities for the fiscal year 2003, based on Web-of-Science SCI data provided to Steunpunt O&O Statistieken via a license agreement with Tbomson-ISI. We also discuss the limitations of the current approach that was based on inter-university publication and citation counts. We provide insights into future adaptations that might make it more representative of the total research activity at the universities involved (e.g., by including data for the humanities) and of its visibility (e.g., by including impact measures). Finally, based on our current experience and interactions with the universities involved, we speculate on the future of the specific bibliometric approach that has now been adopted. More specifically, we hypothesize that the allocation method now developed and under further improvement will become more criticized if it turns out that it (1) also starts influencing intra-university research allocation decisions and, as a consequence (2) introduces adverse publication and citation behaviors at the universities involved.