Reporting bias. Were voters in the 1980 U.S. presidential election being myopic?

Reporting bias. Were voters in the 1980 U.S. presidential election being myopic?

Reagan Library Archives/Wikimedia Commons

Voter Shortsightedness May Skew Elections

“Are you better off than you were 4 years ago?” Ronald Reagan’s famous question in the U.S. presidential election of 1980 is generally a good yardstick for picking a candidate, or at least for judging a leader’s economic policies. But few voters follow it. Instead, they are swayed by economic swings in the months leading up to the election, often ignoring the larger trends. Why are we so shortsighted? A psychological study of voting behavior suggests an answer and points to a simple fix.

Politicians are all too familiar with voter fickleness. As a consequence, says Gabriel Lenz, a political scientist at the University of California, Berkeley, “governments increase spending and reduce taxes irresponsibly just before elections to boost their reelection chances, leading to long-term inflation and other problems.” 

To Lenz, voters’ obsession with recent events sounded similar to “peak-end” theory, a bias long known to psychologists in which people tend to judge an experience by its extremes and its ending. Do individuals judge politicians in a similar way? To find out, Lenz teamed up with Andrew Healy, a political scientist at Loyola Marymount University in Los Angeles, California, and tested the simulated voting behaviors of 7000 people with a series of 29 surveys and experiments. The data gathering alone, which involved online and lab studies, took 2 years.

The first thing the duo explored was voters’ intentions. Are people aware of how much they are influenced by recent economic trends? The answer was surprising. “Voters don't realize what they are doing,” Lenz says. Even for U.S. presidential elections, where voters are clearly influenced by the election-year economy, U.S. voters in a representative survey sample were adamant that they give no extra weight to trends in an election year.

Healy and Lenz then created hypothetical governments with various economic trends and asked people in the lab to judge the hypothetical leaders’ policies. For example, one hypothetical governor was up for reelection after 16 years in office. Overall, the governor increased income, but with a dip at the end of his current term. “Given that participants had 16 years of economic data to evaluate, we thought they would focus less on the end,” Lenz says. And yet the subjects showed the same overemphasis on the most recent years.

The bias was clearly real. But could it be corrected? Healy and Lenz challenged their subjects to evaluate hypothetical governments based on slightly varying information. For example, some received information expressed as yearly income while others received the same information expressed as a yearly growth rate. The same information in a plot of steadily increasing average personal income over 3 years—$32,400, $33,100, $33,800—can also be expressed as a steadily decreasing rate of growth—3%, 2.3%, 2.1%.

That did the trick. Just changing the units of the data was enough to cure voter fickleness. When economic trends were expressed as yearly income rather than rates of change, the subjects made accurate judgments. But if the same information was expressed as a change over time—the bias reappeared, the team reports in the January issue of the American Journal of Political Science.

Aside from clarifying why the end of a politician’s term has such a powerful effect on elections, the study offers a simple corrective, Lenz says. “People find growth rates confusing,” he says. “Journalists could help voters by reporting yearly average family earnings rather than the percent change in yearly earnings.” Or even better, “stories could place long-term growth in historical perspective, telling voters whether it was better or worse than average.”

“I find the conclusions very convincing,” says Adam Berinsky, a political scientist at the Massachusetts Institute of Technology in Cambridge. And unlike most studies of voter behavior, which “only diagnose problems in democratic competence,” Healy and Lenz have found a cure, Berinsky says. “If you present economic growth information in a way that voters can understand … [they] can properly use that information to make informed decisions.”