Week of 10.17.08
Prejudice, Polling, and the Election
The "Bradley Effect": Myth or Maybe?
In the 1982 California gubernatorial election, Democratic candidate Tom Bradley lost to the Republican candidate George Deukmejian, even though several public polls showed Bradley with a clear lead. This led to speculation that voters purposely misrepresented themselves and their biases against an African American candidate, such as Bradley, leading to wildly inaccurate poll results. The theory, often referred to as the Bradley Effect, suggests that some fear that they will open themselves to criticism of racial motivation if they reveal their true choice.
The Bradley Effect is named after Tom Bradley, who lost his bid to become America's first black governor, despite being ahead in the polls.
With Barack Obama vying to be the first-ever African American President, speculation about another instance of the Bradley Effect has been widespread. But, behind all the talk and punditry, have we dug deep enough for the truth?
Dan Walters, a political columnist for the Sacramento Bee, covered California's gubernatorial campaign in 1982 and provides unique perspectives on why he believes the Bradley Effect is a political myth.
Dan Walters, political columnist for the Sacramento Bee.
NOW: As someone who covered California's 1982 gubernatorial election, do you believe the so-called Bradley Effect is a myth or a true phenomenon?
In any election poll you're going to have people who don't tell the truth, but my contention is that the so-called Bradley Effect did not decide that election. If you look at the actual results of the election, and you look at the polling before the election, you see that they're not really in conflict. The Bradley Effect is, in fact, a myth.
NOW: But what about polls before the election that showed Tom Bradley winning?
"Bradley was so confident that he had won he actually stopped campaigning about ten days before the election."
Ahead of the election, the polling showed that Tom Bradley, the mayor of Los Angeles, had a modest, not overwhelming, lead in the election, something like 5 or 6 percentage points. The polling that was being done at that time was based on an understandable assumption that all of the votes would be cast on Election Day. Based on the polling, Bradley's campaign was convinced that they had already won the election. So much so that Team Bradley were literally dividing up office space in the governor's suite, and figuring out who would get what job in Bradley's new administration.
NOW: So what happened?
Bradley was so confident that he had won he actually stopped campaigning about ten days before the election. It was so widely believed that he was going to win that many in the California news media stopped traveling with his campaign. Reporters were even writing up their articles in advance on how Bradley won and the historic nature of having America's first black governor.
While Bradley stopped campaigning, his opponent, Republican Attorney General George Deukmejian, kept on going. Before the election, I was traveling with Deukmejian and his campaign. On the Saturday before the election, his team showed me some private polling data, which indicated that Deukmejian was within three or four points of Bradley, which is more or less the margin of error in most polling. In fact, V. Lance Tarrance, Jr., who was Deukmejian's pollster, recently wrote an article debunking the Bradley Effect, citing the daily tracking polls he was conducting—the data that Deukmejian's staff shared with me a few days before the election—indicating that the race was narrowing.
NOW: So they thought there was a chance they could win?
"In California political circles it is well known that the Bradley Effect was dreamed up after the fact to explain away polling and campaign strategy errors."
Based on that polling, Deukmejian's campaign believed they had a real chance of winning. And one reason was because of the absentee vote. A few years earlier, the legislature had liberalized voting rules, making it much easier to vote absentee, thinking it would help the Democrats. But the Republicans saw this and decided to use it to their advantage. They had lists of people who were gun owners and other conservatives and organized a vote-by-mail turnout campaign to mobilize them to vote against a gun control measure also on the ballot and vote for Deukmejian at the same time, believing that the mailed ballots could be decisive in an otherwise close election.
NOW: What happened on election night?
I was at the Biltmore hotel in LA, which was the election night headquarters for Tom Bradley. As the polls closed at 8 p.m, the state's preeminent pollster, Mervin Field, declared—based on exit polling of Election Day voters—that Bradley had won the election. Everyone was crazy with excitement. And Tom Bradley did win that night. But once they counted the absentee ballots, he had lost by about 94,000 votes.
NOW: So you're saying that the voters didn't lie to pollsters in significant numbers?
I'm sure some people lied to the pollsters, but there's no reason to believe that the so-called Bradley Effect had any material impact on the outcome. I would point out that Bradley's fellow Democrat, then-Gov. Jerry Brown, lost a bid for the U.S. Senate that day by about a half-million votes, so Bradley actually did much better than Brown, another indication that there was no appreciable anti-black sentiment working that day. A lot of moderate Democrats and independents voted for Bradley, but did not vote for Brown, shifting to his Republican rival, Pete Wilson.
"People like MSNBC pundit Chris Matthews talk about the Bradley Effect as if it's an established fact."
After the election, Mervin Field promulgated the theory that the pre-election polling was wrong because voters had lied to poll-takers about their intention to vote against a black candidate, even though there was never any statistical evidence, as Tarrance has observed. And the Bradley folks jumped on the theory as a rationalization for the simple fact that they had blown the election by taking victory for granted, stopping campaigning more than a week before Election Day, and being "outhustled" by the Republicans on absentee voters. It was their way of sidestepping recriminations within the Democratic Party for such a narrow loss. In California political circles it is well known that the Bradley Effect was dreamed up after the fact to explain away polling and campaign strategy errors.
NOW: If the Bradley Effect is not true, as you claim, why is it getting so much attention?
It got attention back then because it was such a high profile race and Bradley had been expected to win, based on what turned out to be incomplete polling. The national media parachuted in, looking for a quick and dirty explanation, got it from those involved and the Bradley Effect became an assumed fact. Now, people like MSNBC pundit Chris Matthews talk about the Bradley Effect as if it's an established fact, and it's not. It's just a myth.
NOW: Do you believe the issue of race will cause any problems in polling this election?
I don't know. I'm sure some people will vote against Obama because he's African-American. But it's one of many factors. They may vote against him because he has an Arab name. Some might vote against McCain because he's a white man or because he's 72 years old. I don't know if you can ever determine these things.
THIS WEEK ON NOW
A Better Bailout?
How to Win Virginia
Virginia's Changing Political Face
Prejudice, Polling, and the Election
Buy a DVD
Topics search results will display here.