Diversity is a policy objective pursued by most higher education systems; at the same time these systems are also concerned about equity of access and the quality of educational opportunity. For a variety of reasons, individual institutions attempt to benchmark themselves against other institutions. Both activities involve measurement, classification, and the selection of peer. Although often addressed apart from each other, diversity and peer selection can be conceptually linked within single scales of similarity and dissimilarity, although existing paradigms that explain diversity may be too simple for reliable peer selection and comparison. A case study of the University of Toronto (Canada) is used to discover the connections between diversity and peer selection, test existing paradigms, and develop a modified methodology that can be used for selecting peers and measuring diversity. Among the study's conclusions are: (1) program cost structures affect institutional cost structures to a large enough extent to be detected in rankings of similarity and dissimilarity and in the measurement of diversity; and (2) of the four principal paradigms--resource dependence, natural selection, competition, and social organization--resource dependence appears to be the most robust in measuring differences in diversity; natural selection and social organization provide better explanations of how diversity develops. (Contains 29 references.) (CH)