Matt Rota

As the 2008 U.S. presidential election approaches, tens of millions of voters have to make up their minds. They face the task of sifting through media reports, televised debates, political advertisements, campaign literature and conversations with family and friends to identify a candidate who best reflects their political views.

That just may be too much to ask, though. As political scientists have long lamented, the general public knows depressingly little about politics. Most Americans can identify the president but barely half know the name of even one cabinet member and only one-third correctly identify their two U.S. senators or their congressional representative. In surveys, roughly half of registered voters display little understanding of how government works or of current political issues.

Even if a voter knew enough to evaluate each presidential candidate’s positions on diverse issues, he or she would still need to tally pros and cons on those issues for each candidate and determine who most deserved support. Decision researchers in various fields have long favored this exhaustive, coldly logical approach, even if only as an ideal that less methodical thinkers should strive for.

Yet according to many psychologists, people will never think that way. We shun rationality and seek as little information as possible when making judgments, the experts assert. Instead, individuals use strategic shortcuts, also known as rules of thumb or heuristics, to decide. The latter term, of Greek origin, means “serving to find out or discover.” Heuristics require minimal mental effort but prompt irrational and biased judgments — or at least so say some psychologists.

Political scientists generally assume just the opposite. They regard heuristics as tools for the average citizen to fashion reasonably accurate political judgments out of sparse civic knowledge.

A recent experimental innovation promises to better illuminate heuristics’ strengths and weaknesses. Researchers now can track how volunteers decide whom to vote for during mock presidential election campaigns. Results so far indicate that well-informed voters employ heuristics better than they do extensive information analyses to select a candidate who best reflects their own views. In contrast, poorly informed voters experience problems in picking appropriate candidates, especially when using rules of thumb.

In general, rational folk who seek as much information as possible about candidates’ positions on many different issues tend to make poorer decisions about whom to support in mock campaigns than do those who follow simple heuristics. These rules of thumb include choosing candidates based on what political party they belong to or which organizations endorse them.

“At least in politics, more information does not always result in better decisions,” says political scientist Richard Lau of Rutgers University in New Brunswick, N.J. “In fact, it often results in worse decisions.”

Other new research suggests that heuristics based solely on certain emotional reactions to candidates, such as admiration and contempt, also guide voting decisions surprisingly well.

12-minute campaigns

Although political scientists typically use surveys to examine voters’ attitudes about political issues and candidates, Lau and colleague David Redlawsk of the University of Iowa in Iowa City take a different approach. They use computers to model campaigns and track how people actually decide whom to vote for in mock elections.

Lau and Redlawsk revised the classic “information board” that has long been used in psychology and marketing to study decision making. An information board looks much like the game board for the television show Jeopardy, with a matrix of columns and rows of boxes that conceal information. Columns on the board are headed by various alternatives, such as a series of political candidates. Rows are labeled with different attributes, such as experience and stands on issues.

In the updated version, volunteers uncover information that they want to learn by clicking a box on the screen. Researchers record what information gets examined, the order in which it’s retrieved and how long it’s perused. Over the past seven years, the two investigators’ findings based on the method have stirred much interest at political science conferences.

Lau and Redlawsk’s “dynamic process-tracing” method uses the information board format to mimic the overwhelming flow of information during presidential campaigns. This approach features a mock primary campaign with six candidates, two Democrats and four Republicans or vice versa, followed by a general election campaign between each party’s nominee. Volunteers register with a party, vote in that party’s primary and then cast ballots in the general election.

The primary campaign lasts about 20 minutes. The general election unfolds over 12 minutes.

During a campaign, columns of boxes on what looks like an information board scroll down a computer screen and disappear, replaced by others at the top of the screen. Participants thus have access to only a fraction of the total information pool at any one time. As in real campaigns, some types of candidate information, such as poll results, appear more often than others do, such as endorsements and issue statements.

Voting Behavior and Emotional HeuristicsEmotions, specifically admiration and contempt for a politician, mirrored voters' stated preference better than political party affiliation in a study of 70 voters before the 2004 election, suggesting that these emotions are often used as shortcuts when choosing a candidate. Click on the picture to view a larger version.J. Korenblat (source: Wang, et al., J. Behavioral Decision Making, 2008

At regular intervals, a 20-second political advertisement from one of the candidates takes over the computer screen.

Similar to most voters, no one in the study can read and consider every bit of information presented during these mock campaigns, much less compare candidates on every political attribute.

If a participant employs a particular heuristic, such as paying special attention to which groups endorse different candidates, then Lau and Redlawsk can see whether that person consistently clicks on endorsements during a campaign.

Before the mock campaign, researchers survey each volunteer’s political attitudes to determine the candidate that most closely aligns with each volunteer’s views — thus the best voting choice.

In a groundbreaking 2001 study that launched the real-time analysis of how people make voting choices, Lau and Redlawsk found that nearly all of 657 eligible voters, ages 18 to 84, used heuristics at least some of the time in determining which mock candidate to support. Available shortcuts included relying on a candidate’s party affiliation, making assumptions about a candidate’s ideology based on party affiliation, checking candidate endorsements, tracking poll leaders and judging candidates based on their physical appearance in photographs.

Using shortcuts — especially the tracking of endorsements — allowed most politically sophisticated volunteers, as determined in a survey, to choose and vote for the candidate who best represented their views. That proportion dipped to a bare majority among those who didn’t use heuristics.

Unlike informed voters, politically naïve volunteers usually failed to vote in their own best interests if they used heuristics. Uninformed participants did better when they avoided using rules of thumb, identifying the best-suited candidate about half the time.

“Heuristics aren’t a saving grace for apathetic voters,” Redlawsk says. “But voters who understand the political environment can use these shortcuts to their advantage.”

Even political sophisticates sometimes mess up, however. When presented with a choice between a stereotypical candidate from their own party, say a moderately liberal Democrat, and a free-thinking candidate of the other party, such as a Republican with a mix of conservative, liberal and moderate views, well-informed voters chose the wrong candidate almost half the time.

In this situation, the political environment suddenly became unfamiliar, Redlawsk holds. Decision making shortcuts that typically had worked now fizzled out.

In another study, described in their 2006 book How Voters Decide: Information Processing During Election Campaigns, Lau and Redlawsk find that voters get superior guidance from simple heuristics than from valiant attempts to account for lots of information.

For instance, volunteers who compared candidates on one or a few key attributes — such as the competitors’ stands on abortion and tax policy — frequently chose the politician who best matched their own overall preferences. Accurate choices steadily declined as participants considered more and more political material.

The latter, read-everything strategy overwhelms people’s limited capacity to remember and consciously manipulate pieces of information, Lau and Redlawsk theorize. Voters end up confused rather than enlightened.

That conclusion echoes the findings of psychologist Gerd Gigerenzer at the Max Planck Institute for Human Development in Berlin. Gigerenzer and his colleagues find that “fast and frugal” heuristics aid all sorts of uncertainty-drenched decisions, such as designing a stock portfolio based on choosing familiar versus unfamiliar companies. These rules of thumb tap into meaningful cues for the task at hand.

“Helpful intuitions can largely be explained by the use of simple heuristics,” Gigerenzer says.

Feel the vote

Simple but helpful heuristics may sometimes travel from the gut to the mind. New research suggests that gut-level emotional reactions to political candidates effectively guide voting decisions.

“Although emotional reactions to public events are rich and voluminous, voting preferences may be determined by only one or a few critical emotions,” says psychologist X. T. Wang of the University of South Dakota in Vermillion.

In a pair of experiments conducted two months before the 2004 presidential election between George W. Bush and John Kerry, Wang studied a total of 210 eligible voters. The first experiment required volunteers to list and prioritize the political issues that most concerned them. Each participant then rated how much candidates’ policies on each issue agreed with their own views, ranked whether they felt positively or negatively toward Bush’s and Kerry’s views on each issue and estimated the likelihood that each candidate would implement policies deemed critical by the voter.

Volunteers then voted for a candidate, and they revealed their own and their parents’ political parties.

Divvying up votes either according to each participant’s party affiliation or ratio of pros to cons for Bush and Kerry closely predicted the final vote breakdown, Wang reports in the January Journal of Behavioral Decision Making.

In the two trials, Kerry received about 60 percent of the vote. In contrast, a mathematical formula that accounted for all information obtained from participants incorrectly tagged Bush as the winner of the first experimental election.

After prioritizing their policy concerns, the 70 volunteers in the second experiment rated four types of emotional reactions to Bush and Kerry on a five-point scale. These “interpersonal” emotions consisted of admiration, contempt, envy and pity or sympathy.

Simply by noting whether Bush or Kerry received a higher admiration rating from each voter, Wang closely predicted the final voting breakdown of study participants. A prediction based on which candidate received a lower contempt rating worked almost as well. In fact, participants’ levels of admiration and contempt for candidates substantially outperformed their party affiliation in predicting their final vote.

Wang is now exploring whether people eligible to vote in this year’s presidential primaries used heuristics and favored certain emotions, such as ranking candidates according to admiration levels.

Political scientists have long used surveys to study whether voters like or dislike candidates, an approach that roughly corresponds to Wang’s focus on admiration and contempt. Analyses of survey data paint a darker portrait of voter decision making than Wang does, however. Voters typically overlook even sharp differences between their own views and those of favored candidates and political groups, contends political scientist Larry Bartels of Princeton University.

“Most of the time, voters merely reaffirm their partisan and group identities at the polls,” Bartels and his Princeton colleague Christopher Achen concluded in a 2006 paper. “They do not reason very much or very often. What they do is rationalize.”

People indeed find it hard to change their long-held opinions about a candidate, even as information that challenges those opinions comes to light, Redlawsk says. During mock election campaigns, he finds that volunteers actually become more likely to vote for an initially liked candidate who suddenly starts to express opinions that differ from their own. However, policy conflicts eventually become so great — usually when 80 percent of information about a preferred candidate clashes with a supporter’s views — that people switch their allegiance to another candidate.

The generous leeway granted to initially favored candidates may, at least for political sophisticates, reflect the intuitive strength of heuristics that they used in the first place, in Redlawsk’s view.

Survey says

As early as 1960, four political scientists concluded that most voters use little knowledge to anoint a political candidate as their favorite. After analyzing national surveys conducted before and after the 1952 and 1956 elections, the researchers concluded that a person’s political party and socioeconomic background powerfully shaped voting preferences. Media reports, political discussions and other factors had noticeable but less pronounced effects on voting decisions, they wrote in The American Voter.

A follow-up to that book, titled The American Voter Revisited, reaches much the same conclusion. Political scientist William Jacoby of Michigan State University in East Lansing and his colleagues probed national surveys conducted before and after the 2000 and 2004 elections.

About 80 percent of the electorate reports only peripheral concerns with politics, the researchers find. Personal identification with one or the other political party remains a prime influence on voters today.

Typical voters use party affiliation to pick a candidate much as consumers use brand loyalty as a convenient way to make purchases, Jacoby says.

However, he adds, one-time surveys may not tap into the broad array of voting heuristics illuminated by Lau and Redlawsk’s “very creative” research method. Although still regarded by many political scientists as tools for ignorant voters to make adequate decisions, Jacoby suspects that heuristics may actually play to the advantage of political sophisticates, as Lau and Redlawsk find.

If that’s true, it suggests an intriguing voting strategy: Stay consistently informed about the political sphere so that you can bypass much of the information thrown at voters during election campaigns and cast a simply effective vote.

• Print
• |
• Comment