Alternative Models of Issue Voting in the 1978 Congressional Elections.

Abstract

This study challenges the widely held belief that minimal issue voting occurs in congressional elections. That traditional conclusion is based on a conceptualization of the congressional vote decision with respect to issues which is inadequate because it assumes that citizens must compare the positions taken by competing c and idates and vote for the c and idate whose views most agree with their own. Instead, because of factors like party, incumbency and the lower stimulus character of congressional elections, some citizens only pay attention to one c and idate and vote their approval or disapproval of his issue positions; such people would also be issue voters. This suggests that we should study the behavior of three types of voters: those who pay attention to the st and s of zero, one or two c and idates (zero-, single-, and paired-alternative voters). The data source is the 1978 University of Michigan National Election Study, and the questions on which the analysis is based are the seven-point issue scales on which respondents located themselves and their congressional c and idates. The topics considered are the frequency of and differences between the three groups of voters, the relationships between issues, party and incumbency to votes, the impact of various rationalization proceesses, and differences between House and Senate voters. The results show a more complex view of congressional issue voting than has generally been presented. With party and incumbency controlled, a set of traditional partisan issues affected the vote decisions of single-, and paired-alternative voters. Though House issue voters were a clear subset of Senate issue voters and though there were more Senate paired-alternative voters, issues had a similar impact on the decisions of the House and Senate voters who took them into account. While clear nonattitude and projection effects were also present, they appear comparable to such effects in the presidential case, and their combined impact leaves room for genuine congressional issue voting. Finally, the effects of incumbency were more complex than anticipated because they varied across types of election contests, groups of voters and facets of issue voting.