Team Size and Geographic Dispersion in Scientific Research:

Team size in scientific research and its geographic dispersion are important because research collaboration indicates the division of labor, and because collaboration is one channel of by which knowledge spills over. For both reasons, economic efficiency of the knowledge-creating industries is intertwined with team size. This paper examines cross sectional patterns and time trends in the size of scientific teams, in collaboration between scientific institutions, and in the internationalization of research over the period 1981-1999. The data consist of 2.4 million scientific papers, or team-piecework observations, that cover the 110 top universities in the United States during 1981-1999. The number of authors per paper is our measure of team size. We find that size of scientific teams increases by 50% over the sample period. An unusually large proportion of the increase takes place since 1990, when the Internet underwent rapid commercial expansion. More striking still is the evidence on geographic dispersion of teams. Three explanations likely account for the rising institutional collaboration and internationalization of science: a) an increasing range of complementary skills is required to do cutting-edge research; b) the costs of team coordination at a distance have decreased; and c) the expansion of research outside the U.S. has made international collaboration more attractive. Given the speed of the adjustment, though, we conclude that declining cost and expansion of research elsewhere are the most likely explanations for growth in institutional collaboration, including abroad.