Capturing Campaign Effects

The Impact of Campaigns on Discrepancies, Errors, and Biases in Voting Behavior

The Impact of Campaigns on Discrepancies, Errors, and Biases in Voting Behavior Patrick Fournier TH E R E A R E VA R I O U S WAYS B Y which political campaigns can be relevant . One can look at the impact of political campaigns on vote intentions, on opinions about leaders and issues, on perceptions of candidates’ issue positions , and on the determinants of decisions (Bartels 1988, 1992; Franklin 1991; Johnston et al. 1992, 1996b; Johnston, Hagen, and Jamieson 2004; Gelman and King 1993; Norris et al. 1999; Nevitte et al. 2000; Blais et al. 2002). However, one terrain for campaign inBuence remains to be explored: what happens to interpersonal diversity in behavior during a campaign? Previous research has identiAed two forms of heterogeneity in political behavior: diversity in the process of decision making and diversity in the outcome of decision making. First, there is evidence that people of different information or sophistication levels do not reason the same way about politics: they rely on different considerations, or they give different weight to similar considerations (Stimson 1975; Sniderman, Brody, and Tetlock 1991; Johnston et al. 1996b; Fournier 2000). Second, low political information ultimately leads to individual and aggregate choices that deviate from informed decisions: less knowledgeable citizens have opinions that differ from those they would have had they been fully informed (Bartels 1996; Delli Carpini and Keeter 1996; Althaus 1998; Luskin, Fishkin, and Jowell 2002; Sturgis 2003). The dynamics of these two types of interpersonal heterogeneity during political campaigns have not been investigated. We do not know 45 whether the differences between very and less informed citizens are minimized or compounded by the progress of a campaign or whether they remain intact, unaffected by the unfolding of campaigns. Yet, such an investigation is crucial, as it informs us about the impact of the democratic process on the quality of voters’ decisions. We need to ascertain whether campaigns help people make more enlightened choices. Never mind learning and persuasion, this is arguably the most socially important effect that a campaign can have. This study tackles an important but neglected issue by analyzing the impact of campaigns on both diversity in decision making and diversity in decision outcome. Three separate campaigns are examined: the 1988, 1993, and 1997 Canadian federal elections. The goal is to determine whether campaigns increase, decrease, or do not inBuence the differences between information groups in the determinants of vote choice on one hand and the individual and aggregate deviations in vote choice attributable to low information on the other hand. Heterogeneity and Political Campaigns There are two types of interpersonal heterogeneity in models of individual political behavior: heterogeneity in the process of decision making, and heterogeneity in the outcome of decision. Heterogeneity in decision making refers to relationships between independent and dependent variables that are not uniform across the entire population, that differ in strength across segments of the population. For instance, in a model of voting behavior , partisan identiAcation, ideology, issues, and leader evaluations may not have the same impact on the vote of all citizens. Ideology may be more closely linked to the decision of some individuals (perhaps the more informed ), while leader evaluations are stronger determinants of the choice of other citizens (possibly the less informed). Such results are generally understood as a sign of differing decisional or reasoning rules: people with differing levels of expertise think differently about politics, relying on different considerations or giving different weight to similar considerations (Rivers 1988; Sniderman, Brody, and Tetlock 1991).1 Sniderman, Brody, and Tetlock (1991) found interpersonal diversity between individuals of different education levels among a myriad of political decisions. An expanding number of studies have uncovered evidence of this type of heterogeneity between citizens of different levels of political information (Zaller 46 Capturing Campaign Effects 1992; Lupia 1994; Delli Carpini and Keeter 1996; Johnston et al. 1996b; Krause 1997; Fournier 2000).2 There is also evidence of heterogeneity in decision outcome. Studies demonstrate that the information deAciencies of the electorate seriously affect individual choices and collective preferences (Bartels 1996; Delli Carpini and Keeter 1996; Althaus 1998; Luskin, Fishkin, and Jowell 2002; Sturgis 2003). First, at the individual level, lack of information leads to opinions that differ from informed opinions. For instance, uninformed voters do not mimic the choices of informed voters with similar sociodemographic proAles: they do not vote the way they would have had they been fully informed (Bartels 1996...