An organizational perspective to funding science: Collaborator novelty at DARPA

Research funding is known to affect the rate and direction of scientific and inventive activity. Most commonly, this is understood as occurring through the allocation of funds toward certain types of research and by altering the disclosure regime. This paper calls attention to another set of factors that likely also influence scientific and technological outcomes, but has gone largely unexamined: organizational practices at research funding organizations. I explore these factors in a mixed-methods study of the Defense Advanced Research Projects Agency (DARPA). Leveraging 39 interviews of DARPA-funded scientists and program managers, I describe DARPA's agency-driven approach to research program development and intense interpersonal management style. Next, I investigate the impact of these organizational practices on the behaviors of sponsored researchers, in particular the tendency of scientists to form novel collaborations. Using patent disclosure data from the Massachusetts Institute of Technology (MIT) Technology Licensing Office and employing coarsened exact matching to identify a control set of collaborations, I find that DARPA-funded research is at least 6.0 percentage points more likely to involve novel collaborations. A detailed exploration of the novel collaborations presents a picture that is consistent with the paper's description of collaboration-promoting practices at DARPA. By identifying important organizational practices and examining their impact on scientist behavior, this paper aims to highlight the potential influence of organization practices on scientific and inventive activity.