THE BLOG
01/09/2008 08:51 pm ET Updated May 25, 2011

Don't Overplay The "Bradley Effect"

In scrambling to scrape the egg off their faces and find out what happened to the erroneous polling that caused the media to look so foolish for their predictions in the wake of Hillary Clinton's surprise victory in New Hampshire, some are trying to ascribe a race-based explanation. They are citing the "Bradley Effect," a reference to a phenomenon in electoral polling in which some white voters lie to pollsters saying that they will vote for the black candidate on the ballot while actually voting for the opponent, as the most likely explanation for the final result. I'm not saying the Bradley Effect played no role whatsoever in the outcome, but I do believe that those who are drilling this hole are searching in the wrong place. The Bradley Effect is largely a general election, not a primary/caucus phenomenon.

The Bradley Effect has long been a reality in polling for important races featuring Black candidates. Named for former Los Angeles mayor and 1982 California gubernatorial candidate Tom Bradley had a double-digit lead just before his 1982 gubernatorial campaign against Republican George Deukmejian. He lost. Doug Wilder had a double-digit lead in his 1989 race against Republican J. Marshall Coleman. He won by less than one point. Harold Ford, Harvey Gantt, and David Dinkins, among others, all fell victim to this phenomenon to varying degrees.

Let me stipulate here my belief that race can never, ever be underestimated in an America election; we have a long, unfortunate history of racial voting in both parties. However, to argue that it was the reason Obama lost is to do a disservice to Clinton's victory. She won despite the fact that a significant portion of the opinion-based television and radio media were pulling for an Obama victory. Additionally, those who push this explanation seem to be arguing that Iowa Democrats are less racist that New Hampshire Democrats. Arguing that Iowa is more progressive than New Hampshire strikes me as ridiculous, given that Iowa has never elected a woman as Governor or to Congress.

Two other reasons may better explain what happened. First, the polling ended too soon. Most of the significant polls that were conducted concluded on Sunday afternoon. They likely missed the brewing animus some New Hampshire women toward Obama and John Edwards over their Saturday night debate gang up on Clinton. I think that, coupled with her emotional response to a voter question the next day, may have moved some women from the sidelines or other candidates toward Clinton. The polls never saw that.

Second, women stood up for Hillary in a big way. There was a huge gender gap between Clinton and Obama in New Hampshire. Obama beat Clinton in Iowa by five points; Clinton returned the favor in New Hampshire, winning by 10 points.

While the Bradley Effect cannot be discounted, those who cling to it may be guilty of sour grapes, particularly since Obama won Iowa. However, he's taken in more than $40 million in donations, leads in many states, and is the darling of the liberal left. A few bigots can't keep him from the nomination if he's meant to win.

Michael K. Fauntroy is an assistant professor of public policy at George Mason University and author of Republicans and the Black Vote. A registered Independent, he blogs at MichaelFauntroy.com.