Mapping the Network of North American Colleges and Universities: A New Approach to Empirically Derived Classifications

This research paper aims to better understand the network structure of higher education in North America. It draws on a relationally networked dataset of 1,292 degree-granting colleges and universities in North America to develop a modularity class approach to categorizing colleges and universities based on their own self-defined peer networks and assesses the utility of the modularity class approach as well as several measures of network centrality for predicting offerings of new curricular fields. Results show that not all measures of network centrality equally predict organizational change outcomes, with hub/authority position being most important. Additionally, results show that an empirically derived modularity class approach to categorizing organizations has important strengths in relation to more typical approaches based on prestige or perceived organizational characteristics. The approaches detailed in this paper will be useful for future analysts seeking to explain the spread of innovations and behavior across the higher education institutional field, as well as those seeking to understand clustering and organizational divergence.