Why we vote the way we do

Researchers at the University of Michigan’s Institute for Social Research (ISR) have been trying to understand voter behavior since the presidential election of 1948, when the institute, then the newly founded Survey Research Center, launched what eventually became the American National Election Studies (ANES).

The national survey was carried out through face-to-face interviews with the same sample of respondents just before and after the election. Based on the power and promise of that data, researchers asked the same basic set of questions for the next presidential election, and have continued to do so ever since. "The American Voter," by ISR researchers and ANES survey designers Angus Campbell, Philip Converse, Warren Miller and Donald Stokes, quickly became the classic text on voter behavior. Many of the central claims of their 1960 book, based on survey data from the 1952 and 1956 presidential elections, have guided thinking about voter behavior to this day.

Their analysis produced some unexpected conclusions: independents were not the thoughtful and informed voters most observers assumed them to be, but instead were less interested and engaged than partisans, and the average voter was surprisingly unsophisticated as most citizens didn’t make their voting decisions based on policy questions, nor did they hold consistently liberal or conservative views across issues.

The recently published "The American Voter Revisited," by ISR researcher William Jacoby and colleagues Michael Lewis-Beck, Helmut Norpoth and Herbert Weisberg, updates the context of the original book to current times, catalogues how the field has evolved in the intervening years, and tests whether the theories in play more than 50 years ago still apply in the political environment of the 21st century.

Jacoby says the most interesting finding is the extent of overall agreement with the original, particularly given "the vast changes in the mass media, and the supposed decline of American political parties, and the supposed increase in the sophistication of the American public."

But the book illuminates current politics and reveals some changes. For both the 2000 and 2004 elections, the analysis showed that negative perceptions of Democratic candidates Al Gore and John Kerry were more pivotal in putting a Republican in the White House than were positive perceptions of George Bush. Social groupings long identified with the two main political parties—such as labor with Democrats and business with Republicans—still exist, but are not as pronounced or clear-cut as they were. There are more independents than in the 1950s and more of them are politically active and informed. In addition, while the number of citizens who hold consistent ideological views is still small, it has increased from about 10 percent to almost 20 percent, chiefly due, Jacoby says, to “an unusually polarized period of American politics.”

Jacoby is eagerly anticipating the ANES data for the 2008 election. He predicts that the trend of ideological differentiation between the two parties will continue, showing up in stronger-than-usual policy orientations among voters. As for that voter in the booth and whether he or she will be capable of making an informed decision, Jacoby is quick to stress that "The American Voter" never said that voters are fools, and that both the original and the latest round of analysis allow for some optimism in that regard.

"Voters are not capricious," he said. "Using the limited tools that voters employ, they vote correctly most of the time and make the vote that is relatively consistent with their interests."

Provided by University of Michigan