Nationalization of the Electorate in the United States

A current debate within the field of U.S. politics revolves around the extent to which the electorate is nationalized, that is, responds to national-level political forces. We provide additional evidence to help resolve this debate. Using the variance-components model of political effects developed by Stokes, we analyze presidential, senatorial, gubernatorial, and congressional races for the period 1962–1984. The results indicate that presidential races are highly nationalized, and congressional races largely localized. Senate and gubernatorial races are found to have a large state-level component to them. The analysis is also split into two segments to ascertain if a change in political effects is seen during this relatively short time frame; no such differences are found. We conclude that nationalization is not a process that can be understood from evidence provided by congressional races alone. Our results suggest that constituencies respond to political forces in a complex way, reacting to factors that are most relevant to a particular electoral contest.