International research collaboration, research team performance, and scientific & technological capabilities in Colombia---a bottom-up perspective

Recent trends show that Colombian science and technology (S&T) performance is improving rapidly. This is presumably the result of two 'mega trends' characterizing the Colombian S&T system: (1) the rapid professionalization of the R&D enterprise, as reflected by the formation of research teams with the support of the Colombian government and the elite research institutions; (2) the internationalization of its scientific community, especially since the 1990s after the opening of the economy to foreign trade.;This dissertation examines the factors affecting Colombian S&T performance, and particularly the ways international research collaboration affects local scientific and technological capabilities. S&T capabilities are measured by the ability of research teams to produce bibliographic outputs, and to contribute to local knowledge.;Research hypotheses are tested using Zero Inflated Negative Binomial Regression models and logistic regressions to account for the effects of international research collaboration on team output while controlling for team characteristics, partner characteristics, scientific discipline, sector, the characteristics of the teams' home institution, and team location. The study uses control groups and the Propensity Score Matching approach to assess the overall impact of international research collaboration on research team performance while controlling for the effects of endogeneity and selection bias.;Results show that international research collaboration is positively associated with both team output and teams' ability to contribute to local knowledge. The study shows that such effects depend on the type of collaboration chosen and the type of partner involved. Particularly, it shows that while co-authoring with colleagues located overseas or receiving foreign funding increases team output, hosting foreign researchers does not seem to affect a team's productivity once all other variables are held constant. It also finds that collaborating with partners from the South yields greater productivity counts than collaborating with partners from the North, and that funding from southern countries is associated with greater productivity rates than any other combination of collaboration activity and origin of partners.;The study also finds that hosting foreign researchers does not appear to be associated with the probability of teams to involve Colombia in their research process either, and that receiving foreign funding or co-authoring with colleagues located overseas increases a team's probability to contribute to local knowledge. Similarly, the study finds that collaboration with partners from northern countries is strongly associated with a team's ability to contribute to local knowledge, while collaboration with partners from southern countries is not. The study finds that although the number of participating researchers holding doctorates positively affects team output, it negatively affects a team's ability to contribute to local knowledge -- but as team size increases beyond 9 members with a PhD, its effects become positive at an increasing rate. Finally, the study finds curvilinear effects of team size, team age and number of active R&D projects a team manages. Theoretical and policy implications of these and other counterintuitive findings are discussed.