Measuring and Evaluating Changes in State Opinion Across Eight Issues

Abstract

Is state opinion stable or dynamic? Do states exhibit similar or divergent trends? I answer these questions by creating dynamic measures of state opinion using multilevel regression, imputation, and post-stratification (MRP) techniques across eight issues: two global indicators, partisanship and ideology; and six specific issues, including spending on education and welfare, abortion, the death penalty, presidential approval and consumer sentiment. I find that patterns of dynamism and heterogeneity vary. State opinion on some issues, such as consumer sentiment, is highly dynamic but follow closely with national trends. On other issues, such as education spending, states exhibit more stability with heterogeneous trends. And still other issues, such as abortion, exhibit both stability and homogeneity. The results have implications for longstanding debates concerning the nature and origins of mass opinion, the way that state opinions are translated into policies, and how to measure state opinion. Finally, the extensive data set of longitudinal state opinion is publicly available.