Models of Comparative State Politics: A Comparison of Cross-Sectional and Time Series Analyses

Abstract

The purpose of this study is to determine if cross-sectional analyses of interstate policy variation are adequate for explaining the dynamic policy process. Two models, competitive threat and economic resources, are developed, and each is then evaluated across time and across states by multiple regression analysis. The policy dependent variables are measured at the state level: per capita welfare expenditures, welfare expenditures as a proportion of the budget, welfare effort, number of welfare recipients, percent of the budget devoted to education, education effort, and education expenditures per pupil in average daily attendance. The models usually explain more variance in the time series data than in the cross-sectional data. Time series analysis produces several differences of substantive interest, where the inference from time series analysis is the opposite of the inference from cross-section analysis. It is suggested that time series analysis may be theoretically more appropriate in most cases than cross-state analysis.