Capturing Campaign Effects

Informational Rhythms of Incumbent-Dominated Congressional Elections

Informational Rhythms of Incumbent-Dominated Congressional Elections Laurel Elms and Paul M. Sniderman IN C U M B E N C Y A DVA N TAG E A N D informational asymmetries go together in campaigns for the U.S. House of Representatives. It is an uncontroversial proposition that challengers do less well on Election Day in large measure because fewer citizens know who they are and because those citizens that do so know less about the challenger than the incumbent (Mann and WolAnger 1980; Jacobson 1997b). The information voters have about candidates obviously matters. But when must the information, if it is to matter, be acquired? And how is what is learned about the challenger and the incumbent, and, not least, when it is learned in the course of a campaign , tied to what candidates actually do? By way of an analytic strategy we shall parse a question proAtably posed in quite a different context—who knew what and when? In particular , we want to explore the notion of knowledge and the rhythms of its acquisition . We will examine candidate knowledge including and beyond simple name recognition and recall. More fundamentally, we shall argue that the signiAcance of knowledge in congressional elections hinges on when it was acquired. By applying the Arst rolling cross-sectional survey design to House elections, we will explore the informational dynamics of congressional campaigns. 221 Informational Rhythms: Arguments and Hypotheses Our primary objective is to take seriously the dynamics of congressional campaigns. This requires exploring not only the consequences of an election campaign but also the timing of those consequences in the course of the campaign. Consider the classic studies of congressional elections. With few exceptions (Mann 1978; Abramowitz 1975), they assess what citizens know at a single point in time, usually after the campaign is over. But it is indispensable to establish when in the course of the campaign citizens learned whatever they wound up learning about the candidates. It is indispensable because when voters learn about candidates and how much they learn about them are connected. Consider the quite different situations of two citizens who are attempting to decide whether to support the incumbent or the challenger. One of them has followed both candidates throughout the campaign; the other stumbled across the name of the challenger while preparing a practice ballot. When interviewed after the election, both voters may recall the challenger’s name and appear equally informed. But the challenger had a month to make a case to the Arst voter and no more than a week to the second voter. The second voter ends up with the same information as the Arst voter but experienced the campaign quite differently. Because how much voters know about candidates tends to be tied to how long they have known about them, it is necessary to establish not only what voters know about the candidates but also when they learned what they know. The public opinion survey we analyzed, focusing on the congressional elections in Missouri in 1994, is the Arst to have utilized a daily rolling cross-sectional design to study House elections. Until 1996, almost all measures of candidate knowledge in congressional elections were obtained only after the election. The 1996 National Election Study (NES) was the Arst to include a limited number of questions about the congressional candidates in the preelection wave. The 1996 NES has four random quarter samples, while the Missouri Election Study consists of daily random samples. Although restricted to a single state, the 1994 Missouri Election Study included respondents from all nine congressional districts, exposed to campaigns conducted at a variety of levels of intensity. Although no challenger won, a restriction that should be kept in mind in evaluating the generality of our results, a number of them put up good Aghts. The Ane 222 Capturing Campaign Effects granulation of the daily random samples, taken together with the variation in campaign intensity, offers a unique opportunity to observe how congressional campaigns, when they are actively contested, can change the informational landscape. Informational gains, if any are realized, are our principal concern. But, supposing for argument’s sake that citizens do learn about candidates in the course of a campaign, what do they learn? Awareness of the candidates, initially, was equated with an ability to recall who was running. But it became clear quickly that knowledge comes in more than one form. As Mann and WolAnger...