The Comparative Method and the Study of Development, Diffusion, and Social Change

Abstract

The research investigates substantive and methodological concerns in comparative research, particularly the confounding influences of within-system developmental or causal dynamics, across-time or within-system diffusion, and across-system or spatial diffusion. The article argues that extra-system influences represent a special case of the unmeasured relevant variable problem resulting in model specification error and that proper model specification can mitigate the problem. Structural equation modeling is advocated as a significant advance in such specifications. The argument is elaborated with time-series data on Argentina from 1908 to 1946 to test the developmental relationship between industrialization and political behavior.