ABSTRACT
Social scientists, with ever-greater sources of survey and other data, and using the tools of multivariate analysis, have become adept at thinking of social phenomena as having multiple contributing factors, and at weighing the relative contributions of each factor to given outcomes. An upside of these approaches is a greater appreciation for the complexity of the social claims on individual attachments-through gender, for example, as well as race, and race as well as education. This appreciation, however, still tends to overlook the importance of variation across contexts.
