CHAPTER SIX II2 PAUL S.HERRNSONAND KELLYD.PATTERSON Jacobson, Gary C. 1980. Money in Congressionul Elections. Yale University Press. -. 1990. “The Effects of Congressional Campaigning in House Elections: New Evidence for Old Arguments.” American Journal of Political Science 34 “Basic Rule” Voting: Impact (2): 334-62. Keith, Bruce E., David B. Magleby, Candice J. Nelson, Elizabeth Orr, Mark C. Westlye, and Raymond E. Wolfinger. 1992. The Myth of the Independent Voter. of Campaigns on Party- and University of California Press. Kenny, Christopher, and Michael McBurnett. 1994. “An Individual-Level Approval-Based Voting Multiequation Model of Expenditure Effects in Contested House Elections.” American Political Science Review 88 (3): 699-707. Luntz, Frank I. 1988. Candidates, Consultants, and Campaigns. Oxford: Basil SHANTO IYENGAR Blackwell. MacKuen, Michael. 1984. “Exposure to Information, Belief Integration, and In- JOHN R. PETROCIK dividual Responsiveness to Agenda Change. ” American Political Science Re- view 78 (2): 372-91. McGuire, William. 1998. “The Nature of Attitudes and Attitude Change.” In Gardner Lindzey and Elliot Aronson, eds., Handbook of Social Psychology. Vol. 2. Addison-Wesley. Patterson, Thomas E. 1993. Out of Order. Knopf. AMERICAN POLITICS HAS A “fifth estate” of campaign man- Skaperdas, Stergios, and Bernard Grofman. 1995. “Modeling Negative Campaign- ing.” American Political Science Review 89 (1): 49-61. agers and political consultants who are believed to be able to design cam- Stokes, Donald E. 1966. “Some Dynamic Elements of Contests for the Presidency.” paign strategies and tactics that persuade voters to support their candidate.l American Political Science Review 69 (3): 812-26. This conventional wisdom makes their services almost essential for many West, Darrell M. 1993. Air Wars: Television Advertising in Election Campaigns, campaigns, and the credibility of a candidate’s effort often is judged by the 1952-l 992. Washington, D.C.: Congressional Quarterly Press. reputation and win-loss record (and sometimes they are equivalent) of the Zaller, John R. 1992. The Nature and Origins of Mass Opinion. Cambridge, consultants working for the campaign. The use of consultants is so wide- U.K.: Cambridge University Press. -. 1993. “The Converse-McGuire Model of Attitude Change and the Gulf spread that it is not difficult to find jurisdictions where political operatives War Opinion Rally.” Political Communication 10 (4): 369-88. (albeit minor ones) are plying their trade near the bottom of the ballot in races for school boards, municipal judgeships, and county commissions. The sine qua non of their prominence is a prior belief that campaigns shape the outcome of elections. How much consultants contribute to the effi- cacy of a campaign is a subject of debate and not one that we will attempt to resolve.’ This chapter addresses the efficacy of campaigns themselves. There is evidence of campaign effects. “Attack” campaigning, canvass- ing, voter contact, party organization work, campaign spending, media coverage, candidate appearances, television advertising, and the activities of political consultants have been studied for their effects on turnout, can- 1. See the interesting exchange on the 1980 election between Pat Caddell (1981) and Richard Wirthlin (1981), major advisers to, respectively, Jimmy Carter and Ronald Reagan; see also Bradshaw (1995) and Thurber and Nelson (2000). 2. For studies that do address this question, see Herrnson (2000) and Medvic (2000). 113 “BASIC RULE” VOTING 114 SHANTO IYENGAR AND JOHN R. PETROCIK 11s didate images, fundraising, and election outcomes.3 This work has not, The near-conventional academic wisdom, therefore, is that day-to-day however, produced anything near a consensus that campaigns are as impor- campaign events and tactics are mostly “sound and fury signifying noth- tant as one would expect given the money and attention lavished on them. ing. “7 Even some of the most recent assessments indicate that campaigns There is a strong correlation between the vote and structural variables have no net effect because each strategic move by a campaign is countered (a term used here as a shorthand for national economic conditions, in- by its opponent.8 cumbent approval, domestic and foreign tranquility, the honesty and in- tegrity of officials, and standing predispositions, such as party identification). This correlation has persuaded many political scientists Explaining Weak Campaign Effects (few of whom, it must be admitted, have any experience in campaigns) that election outcomes are shaped by factors largely immune to campaign Researchers have proposed a variety of theories to explain why campaign strategies and the maneuvering of candidates. Some findings-for example, effects may be so limited.9 One of the most prominent is the theory of that a president’s pre-campaign popularity accounts for more than two- offsetting effects, which accepts that campaigns can influence voters but thirds of the variance of the vote of the incumbent party-are especially that they almost always have a minimal net impact because the competi- severe blows to the notion that campaigns move the vote.4 It is not sur- tors neutralize each other because of a relative parity of resources, includ- prising, therefore, that forecasting models based on structural indicators ing both funding and know-how.1° have been used extensively to predict election results and that none has A second set of explanations for minimal campaign effects asserts that incorporated campaign-specific strategies, tactics, or events. Their predic- their seeming irrelevance is actually illusory. This account focuses on tive success has shaped a dominant theory of retrospective voting that methodological considerations, particularly the limitations of survey re- allows very little room for candidate maneuvering because the structural search. The founding fathers of campaign research (Paul Lazarsfeld and variables that candidates cannot manipulate are in place before a cam- Bernard Berelson and their successors at the University of Michigan’s In- paign begins.5 In this theory, elections are referenda on the leadership of stitute for Social Research) pioneered the use of sample surveys on the the incumbent president: voters pass judgment on the incumbent’s overall premise that survey respondents’ self-reports are reliable and accurate and record, but particularly on the economic performance of the current ad- that the standard test of a campaign effect-the scale of differences in vote ministration. Many general reviews and applications of these forecasting choice between respondents who self-report high or low levels of expo- models have been done.6 Corroborating individual-level studies have only sure to the campaign-accurately tests campaign effects.l’ reinforced the notion that campaigns have limited and maybe insignifi- If self-reports are not accurate, then the absence of observable cam- cant net effects. paign effects can be attributed to measurement error. In most ways and at most times self-reports seem trustworthy, but there is good reason to be- lieve that reports of campaign exposure are seriously erroneous, because 3. On “attack” campaigning, see Pfau and Kensky (1990); Ansolabehere et al. (1994). the frailties of human memory create considerable slippage between what On canvassing and voter contact, see Huckfeldt and Sprague (1992). On party organiza- respondents say they saw or heard &d what actually transpired. The avail- tion work, see Herrnson (1988). On campaign spending, see Jacobson (1992). On media coverage, see Hershey (1989); Bartels (1993). On candidate appearances and television advertising, see Shaw (1999). And on the activities of political consultants, see Herrnson (1992,200O); Medvic and Lenart (1997); Medvic (2000). 7. See Markus (1988); Bartels (1992, 1997a); Gelman and King (1993). 4. Hibbs (1987), ch. 6. 8. Shaw (1999). 5. The theory of retrospective voting dates to at least V. 0. Key’s classic treatise The 9. For a review of this literature, see Iyengar (1996). % 10. See Markus (1988); Gelman and King (1993). One would expect that these condi- Responsible Voter (Key 1966), but also see Fiorina (1981). $ 6. See Abramowitz (1988, 1996); Lewis-Beck and Rice (1992); Alvarez and Nagler tions are satisfied for presidential campaigns, but not for most other campaigns. (1995); Markus (1988); Kinder (1997); Campbell and Mann (1992); Greene (1993); Brody 11. The NES interview schedule, for instance, typically includes an extensive battery of and Sigelman (1983); Erikson (1989); Rosenstone (1983). questions concerning media exposure (frequency of television news viewing). “BASIC RULE” VOTING 116 SHANTO IYENGAR AND JOHN R. PETROCIK 117 able evidence indicates that self-reported and actual exposure are only In this view, voters are guided by their party affiliation and their assess- weakly correlated.12 ments of the performance of the incumbent exactly because exposure to Adding even further to the measurement error is the fact that self-re- the campaign serves to make these factors even more tightly bound up ported exposure to campaign messages often is related to political atti- with candidate preferences. We suggest that, “Campaigns matter because tudes, including candidate preference, because those who choose to tune they tend to produce congruence between fundamental political condi- in to the campaign may differ systematically (in ways that matter to their tions and predispositions, on the one hand, and vote