Résumé : In the large and complex debate related to the creation, diffusion and protection of academic research results, this paper intends to understand the characteristics of academics involved in the knowledge creation as measured by publications and patents. Moreover, it aims to produce some piece of evidence that it is possible to manage patenting activity without jeopardizing publishing. Analysing the publishing and patenting activity of the 326 faculty members of the School of Engineering at Columbia University between 1970 and 2005, we find out that more than the Bayh-Dole Act, it is the implementation of the IP policy at Columbia University that has created an incentive to patent at the engineering school. We also find out that the probability and propensity to patent is influenced by the scientific production of a researcher, his contacts with industry but also his mindset towards patenting. Analysing the scientific productivity of the researchers, we confirm that heterogeneity in the career might deter the productivity of a researcher. We find that scientific collaboration with industry and technological collaboration on application-oriented projects with public or industrial partners had a positive impact on the probability to be among the best scientists. Finally our results suggest that patenting activity undertaken by Columbia University does not divert academics from publishing and relay the recent findings of the literature.