Bibliometric indicators might be inappropriate tools to evaluate public funded research whose aim is to deal with the socio-economic challenges endured by a nation. By the means of a bibliometric analysis (Web of Science, period 1981 – 2011 for Belgium) combined with qualitative material, it has been argued that the kind of research that is more likely to address societal issues (based on its language, its type of collaboration and its type of research) is the one with the least scientific impact as measured by bibliometric measurements. The outcome of this study helped us realise that there exists a paradox of incentives for academic researchers due to an inadequacy between the production of knowledge that address societal issues and its evaluation: (1) traditional evaluation methods do not produce enough incentives to address societal challenges and that (2) scientists are intensifying their effort to produce valuable results for their evaluation.
