Very little evidence exists on which forms of voter contact have the greatest impact on election outcomes. In this article we attempt to assess systematically the effect of different forms of voter contact on congressional vote choice. By estimating a system of equations for spending and vote choice with three different forms of personal and media contact we are able to investigate how different forms of contact directly affect the vote choice as well as mediate the impact of candidate spending. We find that incumbents and challengers benefit differently from the different forms of contact. Challengers tend to get a bigger direct effect from contact in general probably because as a group they are less well known than incumbents. Incumbents, on the other hand, seem able to use contact through the media to maintain their advantage. Contact through television enhances the effect of spending to a much greater extent than is true for challengers, suggesting that incumbents who reach a large percentage of the electorate are difficult to beat.
Any discussion of elections and campaigns eventually turns to the topic of money. Indeed, it is well documented that the amount of money spent trying to get elected to public office in the US has grown considerably over the last 30 years, although that trend has slowed a bit recently (Alexander 1992). While the trend is easy to see, the reasons for its occurrence are less clear. The availability of new campaign technologies and the everchanging finance laws under which candidates operate are commonly cited reasons. Perhaps the main reason that candidates spend so much is because of the uncertainty surrounding their campaign efforts. Evidence regarding the cost of additional votes and the effectiveness of various campaign techniques is virtually nonexistent (Alexander 1992; Jacobson 1992), as evidenced by the frequently held assertion that "half the money spent on campaigns is wasted. The trouble is, we don't know which half" (Jacobson 1992).
The result of this uncertainty is often seat-of-the-pants decision making in terms of allocating campaign resources. Candidates will often do what has worked in the past, both in terms of their own successes as well as imitating what seems to have worked for others. They will also let tradition guide them, producing yard signs, bumper stickers, and buttons, because campaigns have always produced them (Kayden 1978). As Jacobson (1992: 82) points out, this uncertainty also stimulates innovation. There is an incentive for less advantaged candidates (typically challengers) to try out new tactics. With comparatively few dollars and little hope of winning, they have little to lose and everything to gain.
The goal of a political campaign is straightforward: to contact, convince, and get to the polls, enough voters for victory on election day (Jacobson 1992). It is the purpose of this study to initiate the systematic examination of different ways of contacting voters as they affect congressional vote choice. Several questions guide our research effort. To what extent do the different methods of contact directly affect vote choice? Can we identify certain forms of contact which show the strongest direct effects? Additionally, does the method of contact enhance or inhibit the effect of candidate spending? Or, to put it another way, does the manner in which the voter is reached mediate the effect of the candidate's effort to influence the voter? Finally, are these effects, if any, the same for challengers and incumbents, or do interesting differences emerge?
FORMS OF CAMPAIGN CONTACT
While the figures vary along several different lines, it is usually the case that a large percentage of the campaign budget is spent trying to produce and distribute persuasive messages to the voting public or some targeted segment of it (Goldenberg and Traugott 1984). Indeed, that is the whole purpose of a campaign; to reach potential voters and persuade them to vote for you. …