Page - (000369) - in Knowledge and Networks

367 Our findings lend support to the hypothesis that similarity in knowledge and accumulated capabilities enhance the attractiveness of collaboration options and link maintenance. Nevertheless, firms also seek some degree of heterogeneity in the controls DStatus and DPatAge, for the probability of repeated collaboration increases when the partner is not a firm or when the partner is significantly different in patenting experience. However, these findings can be partially attributed to the specificities of research in biotechnology. One reason is that relationships between industry and the university are prevalent in German biotechnology. Because the innovation process is rather linear, with discoveries being introduced by public research institutes, collaboration between industry and the university is an impor- tant mechanism of technology transfer and thus increases its likelihood. Furthermore, the influence of the difference between the patenting ages of the partners might reflect another widespread form of collaborative combination in biotechnology: young, small companies as the creative engine of joint research and large pharma- ceutical companies as a source of financial resources (McKelvey, 1997; Powell, Koput, & Smith-Doerr, 1996; Ter Wal, 2014). In summary, our findings generally suggest that both similarity and diversity of actors afford incentives to form alliances. Similarity plays a specific role in partner choice with regard to general collaboration experience (DyadCoopPAT5) and the accumulation of resources (DCentrality). Actors seek to connect to actors who can reciprocate their general collaboration expertise and provide a certain basis for mutual understanding. The reciprocity in knowledge gains and the amount of inno- vative capability seem to play a comparatively subordinate role. As far as organiza- tional similarity and patenting age are concerned, actors are inclined to choose diverse partners. Conclusion and Further Research The aim of this study was to elaborate on the coevolution of several attributes of cognitive proximity, social proximity, and similarity in competencies as collabora- tion between two actors progresses. We have contributed to the debate on whether networks are rather stable (i.e., with actors always cooperating with the same part- ners) or volatile (i.e., with actors changing partners regularly). Our findings suggest that firms are prone more to switching their cooperation partner than to repeating the collaboration with a given partner. We found no significant effect of knowledge transfer and prior common experience on repeated link formation. Instead, we found that firms prefer to cooperate with a partner whose knowledge bases and accumulated collaboration experience are rather similar to their own and whose organizational nature and patenting age are rather dissimilar to their own. We did not find evidence to support the hypothesis that potential for innovation and collabo- ration decreases as the overlap of the knowledge bases increases (Gilsing et al., 2008; Nooteboom, 1998; Wuyts et al., 2005). 16 Coevolution of Innovative Ties, Proximity, and Competencies