Using Affiliation Networks to Study the Determinants of Multilateral Research Cooperation Some empirical evidence from EU Framework Programs in biotechnology

This paper studies multilateral cooperation networks among organizations and work on a two-mode representation to study the decision to participate in a consortium. Our objective is to explain the underlying processes that give rise to multilateral collaboration networks. Particularly, we are interested in how heterogeneity in organizations' attributes plays a part and in the geographical dimension of this formation process. We use the data on project proposals submitted to the 7th Framework Program (FP) in the area of Life sciences, Biotechnology and Biochemistry for Sustainable Non-Food. We employ exponential random graph models (p* models) (Frank and Strauss, 1986 ; Wasserman and Pattison, 1996) with node attributes (Agneessens et al., 2004), and we make use of extensions for affiliation networks (Wang et al., 2009). These models do not only enable handling variability in consortium sizes but also relax the assumption on tie/triad independence. We obtained some preliminary results indicating institutional types as a source of heterogeneity affecting participation decisions. Also, these initial results point out that organizations take their potential partners' participations in other projects into account in giving their decision ; organizations located in the core European countries tend to participate in the same project ; the tendency to preserve the composition of a consortium across projects and the tendency of organizations with the same institutional type to co-participate are not significant.