Turning Regression Inside Out: How to Make the Most of “General Linear Reality” in the Study of Social Organization
A cogent case has been made for the need to transcend “general linear reality,” referring to the domination of (generalized) regression modeling as a standard but severely limited method for pursuing studies of social organization (see Andrew Abbott, Sociological Theory, 1988). Researchers interested in social networks have long tended to avoid the general linear model, seemingly with good reason (see Paul DiMaggio, “Comments on ‘What Theory Is Not’,” Administrative Science Quarterly, 1995).
Conventional regression modeling pertains to relations among variables. In this talk however I show that such modeling has a dual and may be turned “inside out:” the usual regression coefficients may in fact be usefully defined and computed from a network among the cases. Research on network modeling, and insights from sociological field theory, may be applied to this network, and doing so leads to new discoveries about the organizational and relational underpinnings of regression models and their applications. I review recent work of my research group on these topics, and discuss several different examples involving welfare states, terrorist organizations, and political mobilization. I argue that relationally-oriented students of social organization should transcend “general linear reality” by exploiting it.
Ronald Breiger is a Professor of Sociology and (by courtesy) Government and Public Policy at the University of Arizona. His primary research interests are in social network analysis, adversarial networks, and cultural and institutional studies. He received the Simmel Award of the International Network for Social Network Analysis (2005), with Linton Freeman served as Editor of the journal Social Networks (1998-2006), and was elected Chair of the Section on Mathematical Sociology of the American Sociological Association (2009-10). He has recently published in Poetics, the Annual Review of Sociology, and Sociological Methods & Research.