Do issues matter? This article extends recent research on issue voting and campaign agenda-setting to voting decisions in congressional elections. We use a unique data set that includes information from a survey of candidates and campaign aides who competed in the 1998 House elections and a survey of individuals who voted in them. The study assesses the impact of campaign-specific variables on citizens' voting decisions, while controlling for relevant attitudinal and demographic factors. We find that when a candidate and voter agree on what is the most important issue in the election, the voter is more likely to vote for that candidate if that candidates party "owns" the issue. The effects of shared issue priorities are especially strong for independent voters.
Do issues matter? The roles of issues and campaigns in elections have been the subject of a recurrent debate in the political science literature. Traditional democratic theory holds that campaigns inform citizens, offer them clear and distinct choices between candidates on the issues, and motivate them to participate in elections. Voters are expected to respond to the information they receive from campaigns and cast their ballots for the candidate who most directly addressed their concerns (Dewey 1954: 122). Early empirical research on campaigns and elections, however, demonstrated that voters' sociological characteristics and partisan attachments were usually more important than issues, as well as campaigns and other short-term forces in influencing how people vote (Berelson et al. 1954; Campbell et al. 1960). More recent research on presidential elections has shown that issues influence how individuals cast their ballots when voters focus on issues that favor a candidate's party (Petrocik 1996).
The scholarly debate over issue voting is far from resolved. Disagreements exist over theoretical considerations, including whether retrospective, prospective, or agenda-setting approaches better explain a voter's decision. Disagreements also arise over methodological approaches, such as whether issue voting is best researched in the field or in a laboratory setting. Additionally, most research focuses on high-visibility elections, such as presidential and gubernatorial contests. However, issue voting also may take place in contests for the U.S. House of Representatives and other down-ballot elections.
This study advances our knowledge of the impact of campaigning on party-owned issues on voters. It also extends recent research on agenda-setting in experimental settings and presidential elections to congressional elections by linking issue-based campaigning to individuals' voting decisions. Previous research on the role of issues in congressional elections based on spatial voting theories have used measures of issue proximity that evaluate voters' ideology across a range of issue questions rather than identifying voters' issue priorities (see Serra and Moon 1994; Wright 1978). These studies also operationalized candidate issue positions by imputation, using either incumbents' roll call votes or voters' perceptions of candidates' issue positions to determine, indirectly, the candidates' issue priorities. Other studies of issues in House elections have focused exclusively on campaign advertisements or voter issue preferences without integrating the two (Abramowitz and Segal 1990; Spiliotes and Vavreck 2002). We use a more direct approach that relies on candidates and campaign aides to establish what constituted the most important issue in their campaigns and on the voters in their districts for what they believe was the most important issue in the election. The study uses a data set that includes information from a survey of candidates and campaign aides who competed in the 1998 House elections and a survey of individuals who voted in those elections to assess the impact of campaign variables on citizens' voting decisions. We find that when a candidate and voter share issue priorities, the voter is more likely to vote for the candidate if that issue is "owned" by the candidate's party. …