WikiSummary, the Social Science Summary Database

Popkin: The reasoning voter

From WikiSummary, the Free Social Science Summary Database


Popkin. 1994. The reasoning voter: Communication and persuasion in presidential campaigns. 2d edition. Chicago: University Of Chicago Press.

In Brief


To develop a theory that explains how voters form opinions and make decisions.

Main Argument

Popkin relies on a theory of low information rationality to explain how voters are able to make rational choices between candidates. Voters do this by using information shortcuts that they receive during campaigns, usually using something like a "drunkard's search." Voters use small amounts of personal information to construct a narrative about candidates. Essentially, they ask themselves this: "Based on what I know about the candidate personally, what is the probability that this presidential candidate was a good governor? What is the probability that he will be a good president?"

Place in Literature

Draws on the Columbia school's insights about campaigns, and socialization, Downs's theory of information shortcuts and voters as investors and on modern political psychology and behavior studies to develop his theory. Argues against theorists such as Converse and those who argue that political behavior is irrational. See Lupia and McCubbins for further development of Popkin's arguments.


Develops theory by drawing on previous research. Supports theory with analysis of recent elections.

The Argument

Low information rationality

Popkin's analysis is based on one main premise: voters use low information rationality gained in their daily lives, through the media and through personal interactions, to evaluate candidates and facilitate electoral choices.

Political "Knowledge": Despite a more educated electorate, knowledge of civics has not increased significantly in forty years. According to Popkin, theorists who argue that political competence could be measured by knowledge of "civics book" knowledge and names of specific bills (i.e. the Michigan studies) have missed the larger point that voters do manage to gain an understanding of where candidates stand on important issues. He argues that education has not changed how people think, but it does allow us to better interpret and connect different cues.

Information as a By-Product: Popkin argues that most of the information voters learn about politics is picked up as a by-product of activities they pursue as a part of daily life (homeowners learn about interest rates, shoppers learn about prices and inflation etc.--thus, people know how the economy is doing). Media helps to explain what politicians are doing and the relevance of those actions for individuals, and campaigns help to clarify the issues. Voters develop affinity towards like-minded opinion leaders in media and in personal interactions.

Media and Friends: Interpersonal communication is seen as a way of developing assessments of parties and candidates. Information received from the media is discussed with friends and helps to create opinions. While voters do care about issue proximity, they also focus on candidate competency and sincerity and rely heavily on cues to make these evaluations.

Party ID as a Running Tally

Drawing on Fiorina (1981) and on his own earlier work (Popkin et al. 1976), Popkin views party identification as a running tally of party assessment and looks at party identification of candidates as providing an important default value which voters use to evaluate them. He sees "a sophisticated pattern of transmission from past elections and interactions among and between people in the current election" (p.71).

Creating Narratives about Candidates (ch 4)

Popkin argues that voters often function as clinicians (who gather limited information and infer from it a broader narrative), in contrast to statisticians (who weigh only facts in order to make a decision). He illustrates a few concepts to explain this relationship:

  • Representativeness Heuristic: Voters often compare a candidate to a pre-existing stereotype of how certain people act. For instance, they may compare a presidential candidate to their image of what a president should be like, or compare a candidate to their stereotype of how someone who "does the right thing" would act. Essentially, we take our pre-existing idea of what a president should be, then compare it with personal information about the candidate using a "goodness of fit" test. A few bad votes in a politicians record can dramatically undermine this "goodness of fit," since we like to think that mistakes by others were intentional (whereas our own mistakes were the result of external conditions). Voters use this personal information not only to predict what kind of president you will be, but also (e.g.) what kind of governor Jimmy Carter probably was: Based on who Carter is, how likely is it that he was a good governor?
  • Gresham's Law of Information: A small amount of personal information can drive out a large amount of previous impersonal information, because personal information is much more helpful than political information in constructing narratives. Because personal information is so important (much more than a political record is), even new challengers can rapidly catch up with incumbents in the polls--though only the incumbent has a political record, voters get personal information about both candidates. Voters judge candidates more on how "presidential" they look than on their actual record.
  • Framing: Framing is the way that we look at the president. For example, heavy media coverage of economic problems leads us to not only update our evaluation of the president's handling of the economy, but also to weight this issue-specific evaluation more heavily when making a broader evaluation of the president's overall performance. Popkin discusses five frames that matter.
    1. Candidate vs president. When incumbents speak only from the Rose Garden, they are seen as presidents. When they are seen on the campaign trail, they are mere candidates.
    2. Candidate's personality vs candidate's record (see above). Personality matters more.
    3. Candidate vs nominee. Political conventions can change the way you are viewed (a candidate becomes a nominee).
    4. Domestic vs international issues. Which one matters for this campaign?
    5. Inflation vs unemployment vs poverty. Which economic problem is current affects how candidates are viewed.
  • Pseudocertainty (Calculation Shortcuts): When we can use one of these shortcuts, we are more confident in our evaluations (although, ironically, our evaluations are probably more likely to be incorrect):
    1. When all our information is consistent (i.e. all supports one candidate)
    2. When probabilities are close to 0 or 1. We don't understand finer probabilities well and are uncomfortable with them.
    3. A good sure thing vs a probabilistic better thing. We like the sure thing better, even though the expected value of the gamble is better (see Quattrone and Tversky 1988).
  • Drunkard's Search: The term is based on the image of a drunk looking for his lost car keys where the streetlight is shining, even though that's not where he lost them; he looks there because that's where the light is. People are more likely to use one-dimensional searches, such as focusing on a single attribute about a candidate, or using the front runner's characteristics as a measurement of other candidates. Symbols are often drawn upon to represent issues (e.g. the hostages representing Carter's incompetence in foreign affairs.). Our decision about where to look for information (i.e. which streetlight to use) determines which decicion we make.


Popkin also focuses significantly on the role of the campaign in facilitating choice. He argues that the campaign (1) increases the importance of (some) issues, (2) strengthens the connections between issues and the office, and (3) increases the perceived differences between candidates. Details:

  • Increasing importance of issues: In the Columbia studies, the authors found that voters have varying degrees of attachment to the parties--and that, if the campaign demonstrates the importance of particular issues right now, then crossover voting is most likely if you aren't confident in your party's ability to handle those issues.
  • Strengthen connections between issues and office: Though voters can imagine connections between certain policies and certain offices, it's hard for them to know which offices are responsible for certain things. News and campaigns can remind voters that a particular office can influence a particular policy area.
  • Increase perceived differences: Demonstrate to voters that the candidates are likely to do different things about the issues that are important.

Data: Primaries

Popkin draws on lessons from presidential primaries to illustrate his theory, noting that, since primaries are more complex than general elections, it should be more difficult to apply the theory to them. He notes how candidates who do well in early primaries get large bounces in approval. Popkin argues that this is due to new info people receive about the candidate, as well as to the victories making a candidate a better strategic choice, not to mention having a persuasion effect by themselves. He links the failure of some experienced candidates to their (perceived) incompetency or to their failure to create a strong narrative about themselves in relationship to the presidency. After laying out the theory, Popkin uses it to analyze the Democratic primaries of 1976 and 1984, the Republican primaries of 1980, and the election of 1992.