'Attraction Effect' Helps Voters Pick From the Pack

By Shankar Vedantam
Monday, January 21, 2008

Some years ago, a political scientist conducted an interesting experiment that speaks to the fractured race for the Republican presidential nomination, which now has six candidates, five issues, and four potential front-runners.

Diane Lowenthal said she had two groups of volunteers evaluate two candidates running for public office: One group was told about a politician with a plan to pump $2 million into the local economy, but who was fighting rumors he had been caught driving under the influence of alcohol. The other group heard about a candidate who had a plan to pump $1 million into the local economy, but with no ethical question marks hanging over him.

When Lowenthal compared the evaluations of the volunteers, she found that the group that evaluated the candidate with the clean record had given him a higher rating than the candidate with the bolder economic plan received from the other group. But when a third group of volunteers evaluated both candidates together, it was the candidate with the dubious reputation and the stronger economic platform who won, said Lowenthal, an assistant professor in American University's Washington Semester program and School of Public Affairs. To understand why the experiment is relevant to the current race, you must understand why it turned out the way it did. Regardless of whether volunteers were evaluating candidates one by one or together, the candidates themselves did not change. Why did the preference of the "voters" change? Seeing both candidates together changed how much importance -- how much weight-- people attached to the pieces of information they were given.

In the absence of comparative information, volunteers who were told a candidate would bring $2 million into the local economy did not have a good way to evaluate whether this was bold or mediocre planning. They did not need any comparative information, however, to know a DUI rumor was problematic. So they gave the candidate low marks. But when volunteers were able to compare the candidates, they could now see that a $2 million plan was tangibly better than a measly $1 million plan. This put the DUI concern into a new perspective -- it was just a rumor. The candidates remained identical, but the weight that volunteers assigned to the measures changed -- and this reversed the voters' preferences.

In real-life political races, candidates are always evaluated together. In the GOP race, I count five issues that seem important to Republicans: social conservatism, fiscal conservatism, illegal immigration, national security and the Iraq war. Your list might be different, but the point of research by Lowenthal and many others is that victory is not decided by the characteristics of the candidates, but the relative weight voters assign to the different issues.

Thinking about the race in this way has immediate and counterintuitive consequences. Political pundits often argue that when two candidates share an interest in an issue, they hurt each other because they are competing for the same voters. This is partly true, but it ignores the fact that having multiple candidates interested in an issue increases the weight that voters attach to that issue at the expense of others. An increased focus on one issue can help all the candidates who are strong on that issue. You can see why researchers dub this phenomenon "the attraction effect" -- candidates inadvertently help other candidates with similar views.

Mike Huckabee's win in Iowa and his strong showing among social conservatives in other states, for example, may have helped Mitt Romney defeat Sen. John McCain (Ariz.) in Michigan and Nevada because Huckabee's win raises the weight that voters attach to social conservatism. Huckabee is stronger than Romney on that issue, but a greater weight on social conservatism helps Romney because he is stronger on that issue than McCain -- and because many Republicans prefer Romney to Huckabee on other issues.

"If the attraction effect holds . . . Romney should do better in a three-way race against Huckabee and McCain than in a two-way race with only McCain," Lowenthal said.

In the Democratic race, John Edwards's strength or weakness can help or hurt Sen. Barack Obama (Ill.), because both candidates are campaigning on a mantra of change. The better Edwards does, the better all "change candidates" will do; if Edwards should fade, voters will attach less weight to the issue of change.

Another way to think of the attraction effect is to imagine a line with many hanging baskets. If you increase the weight in any basket, it shifts the center of gravity of the whole line toward it. Scott Highhouse, a psychologist at Bowling Green State University who has also studied the attraction effect, said this is why a Green candidate in a three-way race with a Republican and a Democrat is likely to help the Democrat -- her presence tilts the center of gravity of the race to make the Democrat look like the most reasonable option.

The attraction effect explains why politicians such as Romney try so hard to be all things to all people -- and why the strategy can work, even if voters find it distasteful. In a complicated race with many issues and no clear front-runner, someone who comes in second- or even third-best on most of the issues can appear to be the strongest candidate overall -- because on each issue, he gets to rub off some shine from the person who is No. 1.

© 2008 The Washington Post Company