Political scientists generally believe that public opinion affects public policy within the American states. While such a connection may seem obvious, firm empirical evidence for the existence of this relationship has only been presented quite recently. Early studies of state policymaking emphasized economic factors as the predominant determinants of governmental activity (Dawson and Robinson 1963;Dye 1966;Sharkansky and Hofferbert 1969). Once variables like per capita income, gross state product, and employment levels were taken into account, political characteristics like party composition of state legislatures, gubernatorial partisanship, and political culture seemed to have no effect. These results were replicated a number of times and widely accepted by state politics scholars (Hofferbert 1974).Of course, political scientists were extremely uncomfortable about the sizable body of empirical evidence that showed politics was irrelevant to governmental decision-making within the states. Therefore, research continued on this topic with dogged determination to show otherwise. The major breakthrough occurred with the pioneering work by Wright, Erikson, and McIver (1985). They used survey data, aggregated to the state level, in order to develop measures of citizen partisanship and ideology within each state. They provide compelling evidence that these measures are effective gauges of mass orientations, and, more important, they show that these aspects of public opinion have a strong influence on state policy liberalism.Indeed, once public opinion is taken into account, economic factors play only a secondary role as determinants of policy (Erikson, Wright, McIver 1989 We wholeheartedly agree that the recent work represents a highly significant step to understanding the state policymaking process. But, we also believe that further work is necessary in at least two directions. First, it is important to identify where and how public opinion impinges upon the policymaking process. Second, the specification of influences on policies must be expanded to incorporate another set of critical actors in the political process-interest groups.This paper constitutes a first step toward those objectives.
MEASURING STATE POLICYA perusal of the literature reveals that there are three dominant approaches to measuring public policy within the states. Many researchers use program expenditures. Another set of scholars relies upon specific policy adoptions. A third group employs composite variables constructed from multiple indicators. We believe that all of these approaches are problematic, for different reasons.
Previous ApproachesFirst, let us consider the research that use policy expenditures as the dependent variable.Studies that take this approach generally focus on spending (usually adjusted by state population, Despite their ubiquity in the field, there is a serious drawback to the use of policy expenditures. Spending within any given policy area is invariably affected by state spending across other policies, which are not under...