How Much do Voters Know?

Richard Johnston — Thu, 04 Oct 2007 13:30:00 GMT

Ask the expert: How much do voters know?

Richard Johnston

University of Pennsylvania

How much do voters know? How much can they learn in campaigns? Do they use what they learn when they make choices among candidates? The answers are mixed. Many voters are woefully ignorant but others learn a lot from campaigns. Well-informed voters can be decisive in elections—but not always—and they rarely are representative of their less-informed counterparts.

Voters’ knowledge

In general, voters do not know much, and most are quite oblivious to the discourse among the elites who debate and make policy. The most important statement to this effect was made in 1964 by Philip Converse [4]. He found that when respondents in U.S. National Election Study samples are asked what they like or dislike about parties and candidates, they cannot volunteer much content. When asked pre-formulated questions about policy choices, these same respondents do not link their answers to different questions on similar topics. They give conservative answers to some questions and liberal answers to others, and in no particular order. When asked the same question two or more times, respondents commonly answer in a way that suggests they are drawing answers, so to speak, out of a hat. Particularly disturbing is what this might imply about stability and change: survey respondents with real opinions rarely change them; the change that does appear is produced by a random process.

Converse’s finding dovetails with others. For example, few voters can recall the name of their member of the House of Representatives, for example. Indeed, only about half the electorate knows which party controls the House. Stories like this can be told ad nauseam.[5]

Voters’ attitudes and opinions

Not everyone buys the claim that voters’ opinions are as impoverished as this. Converse assumes that any slippage across questions or over time is the fault of respondents: they give inconsistent answers because they do not have real attitudes. One could as easily assert that the fault lies with the measures. Survey questions commonly carry some ambiguity. Even more to the point, the answers respondents are allowed to give are crude: “agree,” “disagree,” and the like. It is not hard to see that a respondent whose opinions are basically stable might vary his or her responses over an interval of years, or even days. This kind of variation makes connections among questions and among answers to the same question over time look feeble. Correction for it makes respondents’ opinions look quite robust.[2]

For all this, there is still plenty of evidence that many, perhaps most, citizens’ opinions are weakly held and not supported by concrete information. Indeed, much of what citizens think they know is misinformation. When misinformed survey respondents are supplied with correct information, they tend to resist it.[7]

Voters’ ballot choices

But does it matter? Some argue that the debate is largely irrelevant to the choices that voters actually make and to the circumstances in which they make them. On this view, the landscape, although seemingly bleak, has clear signposts that rarely play citizens false. Voters may not always have opinions, but they do commonly have feelings, about parties and social groups. These feelings align, at least crudely, with voters’ interests, and reliance on feelings is a reasonable starting point.[11] Prominent among the groups that citizens might identify with are political parties and their candidates. Candidates have a powerful incentive to simplify voters’ choices. Often merely knowing which side of a question the party is on helps voters connect means to ends.[12] When both Republicans and Democrats in Congress support a measure, the citizenry at large tends to follow, with the most educated and informed citizens leading the way. When the parties divide, their supporters tend to divide as well, and voters’ sophistication tends not to predict the direction of opinion.[15] The critical question, then, is not so much the amount of background knowledge individual voters possess as whether critically placed actors—especially the major parties—structure the choice. When they do, they make it easy for voters to deploy what little information they have. This is especially telling when voters get to legislate directly, as in an initiative or referendum. The state election authorities will circulate a detailed account of each measure, and this may help voters. For most, however, the pamphlet will just be information overload. It helps if voters can figure out who put the measure on the ballot in the first place. If they cannot, they tend to vote No, a pretty sensible choice under the circumstances. If opponents as well as supporters materialize, voters can triangulate the result. If the pattern of support and opposition follows partisan lines, the voters’ task may be quite simple.[8]

Of course, parties and other actors do not always structure the choice. There is no guarantee that the usual suspects of the political world will line up on opposite sides of a question. And when they do, the result is not always enlightening. To the extent that voters use feelings about groups as a proxy for ideas about policy, some of the feelings may be little more than ordinary prejudice and appeals to such feeling often are crude demagoguery.

Election results

Whatever the general level of ignorance, some people know the score, and these may be the voters who produce the overall result, Yes or No, Democrat or Republican. Poorly informed citizens choose sides at random. In doing so, they negate each other. On this view, Converse’s findings, described above, are actually reassuring! It is true that, even though individuals’ opinions tend to be unstable, aggregate distributions of opinion are quite stable. Movement in opinion tends to be gradual and is related in commonsensical ways to external events and to tides in congressional policy-making.[9,13] Although most citizens cannot recall key facts about Congress and indeed about their own representative, when large numbers of incumbents go down to defeat, the pattern is not random. It reflects their behavior, on roll calls [1,14] or in acts of malfeasance [10]. When survey respondents are asked about candidates’ positions on issues, they are more likely to give the correct answer at the end of an election campaign than at the beginning.[6] This reassuring characterization is true much of the time.

But it is not true all of the time, and it is not the whole story about elections in the aggregate. For one thing, it is not always true that less informed persons choose sides of a question at random. On certain kinds of questions, they are systematically drawn to one side or the other. One side may reflect a widely held popular norm or a persistent piece of misinformation. On such questions, the more informed folks are the ones who divide evenly. In general, the result will be determined by whichever group exhibits the most unbalanced distribution. Moreover, more informed persons commonly have different interests and different opinions from the others. When they lead opinion, they do not necessarily take the electorate as a whole where it wants to go.[3] The result does not feel very democratic.

The summary picture is not as bleak as the early studies suggested. Opinions held by the typical individual are probably not as unstable and inconsistent as originally believed. The typical citizen does not know much, but much of the time does not need to be an expert to navigate the landscape. Even if many individuals do not have real opinions, many others do, and the electorate as a whole has meaningful opinions. In short, candidates for office do not waltz before a blind audience. Their policy commitments eventually become known to a critical fraction of the electorate. But the actual result of elections is not always what Americans would collectively wish for themselves were they fully informed.

1. Abramowitz, Alan I. “Name Familiarity, Reputation, and the Incumbency Effect in a Congressional Election.” Western Political Quarterly 28 (1975): 668-84.

2. Achen, Christopher H. “Mass Political Attitudes and the Survey Response.” American Political Science Review 69 (1975): 1218-1231.

3. Althaus, Scott. Collective Preferences in Democratic Politics: Opinion Surveys and the Will of the People Cambridge: Cambridge University Press, 2003.

4. Converse, Philip E. “The Nature of Belief Systems in Mass Publics,” in David Apter, ed. Ideology and Discontent. New York: Free Press, 1964, pp. 206-61.

5. Delli Carpini, Michael X., and Scott Keeter. What Americans Know About Politics and Why it Matters. New Haven: Yale University Press, 1996.

6. Johnston, Richard, Michael G. Hagen, and Kathleen Hall Jamieson. The 2000 Presidential Campaign and the Foundations of Party Politics. Cambridge, UK: Cambridge University Press, 2004.

7. Kuklinski, James H., Paul J. Quirk, Jennifer Jerit, David Schwieder, and Robert F. Rich “Misinformation and the Currency of Democratic Citizenship.” Journal of Politics 62 (2000): 790-816.

8. Lupia, Arthur. “Busy Voters, Agenda Control, and the Power of Information.” American Political Science Review 86 (1992): 390-403.

9. Page, Benjamin I., and Robert Y. Shapiro. The Rational Public: Fifty Years of Trends in Americans’ Policy Preferences. Chicago: University of Chicago Press, 1992.

10. Peters, John G., and Susan Welch. “The Effects of Charges of Corruption on Voting Behavior in Congressional Elections.” American Political Science Review 74 (1980): 697-708.

11. Sniderman, Paul M., Richard A. Brody, and Philip Tetlock. Reasoning and Choice: Explorations in Political Psychology. Cambridge, UK: Cambridge University Press, 1991.

12. Sniderman, Paul M. “Taking Sides: A Fixed Choice Theory of Political Reasoning,” in Arthur Lupia, Mathew D. McCubbins, and Samuel L. Popkin, eds. Elements of Reason: Cognition, Choice, and the Bounds of Rationality. Cambridge, UK: Cambridge University Press, 2000, pp. 67-84.

13. Stimson, James A. Tides of Consent: How Public Opinion Shapes American Politics. Cambridge, UK: Cambridge University Press, 2004.

14. Wright, Gerald C., Jr. “Constituency Response to Congressional Behavior: The Impact of the House Judiciary Committee Impeachment Votes.” Western Political Quarterly 30 (1977): 401-410.

15. Zaller, John. The Nature and Origins of Mass Opinion. Cambridge, UK: Cambridge University Press, 1992.