By Albert Hunt
The fear of Democrats this year is that presidential candidate Barack Obama's advantage in the polls is exaggerated.
The "bradley effect'' is the murphy's law of US politics, more accepted than demonstrated. The Bradley effect stems from the 1982 California gubernatorial race when the late Tom Bradley, an African- American who was then mayor of Los Angeles, was ahead in the pre-Election Day polls, only to narrowly lose to a bland Republican.
The conclusion: there were a number of whites who, in the privacy of the voting booth, just wouldn't cast their ballots for a person of colour.
The fear of more than a few Democrats this year, and the private hope of some Republicans, is that, based on this axiom, Democratic presidential candidate Barack Obama's advantage in the polls is exaggerated; without a big lead in pre-election surveys, the outcome will be in doubt.
It's a dubious proposition on two grounds.
More recent contests involving prominent African-Americans, including Obama in the primaries, show little gap between what people say in polls and how they vote. Moreover, any small Bradley effect is likely to be offset by the enthusiasm effect working to Obama's advantage on Nov 4.
"There is going to be an exceptionally large turnout, and that will disproportionately help Obama,'' predicts Democratic pollster Peter Hart. He believes a larger turnout of African-American and younger voters than polls measure will offset any private bias against a black candidate.
In the now-infamous 1982 California race, surveys by the Los Angeles Times and by Mervin Field showed Bradley running 7 points ahead of George Deukmejian less than a week before the election. He lost by fewer than 100,000 votes out of 7.5 million cast.
The Bradley effect was also evident in 1989 when Virginia's Douglas Wilder became the first African-American governor in the US in more than a century. Newspaper polls showed Wilder running about 10 points ahead on the eve of the election; he won by less than a point.
It may be that as Americans have become more racially tolerant and exposed to more black candidates, the Bradley effect has dissipated. That seems to be the message of two elections, with different outcomes, in 2006.
Harold Ford, a young African-American running for the US Senate in Tennessee, was trailing by an average of almost 5 points in the four statewide polls taken right before the election. He lost by less than 3 points, closer than any of the polls.
In Massachusetts, black Democratic gubernatorial candidate Deval Patrick ran more than 20 points ahead of his opponents in the polls, and won by almost that much.
Moreover, Obama's performance in the Democratic primaries belied any Bradley effect, reinforcing the argument, as one analyst declared, that this is a "theory in search of data.''
The Illinois senator did run well behind the polls in New Hampshire, when he was upset by Hillary Clinton, although few people attribute that to racial reasons.
This happened in a few other states such as California. At the same time, Obama ran much better than the polls in the Wisconsin, North Carolina and Virginia primaries, among other races. There was no clear pattern.
Hart thinks there still could be a small Bradley effect, representing perhaps 2 or 3 percentage points. The offsets, he figures, are at least that, maybe more.
"There is unprecedented pride'' among African-Americans over Obama's candidacy, Hart says. "He will get more than 95 percent of that vote, and the turnout will be massive. There won't be many African-Americans who say this election doesn't matter.''
This intensity is evident in last week's Bloomberg/Los Angeles Times national poll. Overall, Obama commands more passion, with 84 percent of his voters describing themselves as enthusiastic versus only 63 percent among Republican nominee John McCain's backers. 89 percent of Obama's black supporters say they're enthusiastic.
In the key battleground states, the well-oiled Obama organisation has to deliver these voters to the polls.
Yet even in states where the Democratic candidate isn't contesting, like Mississippi, Republicans worry that a huge black turnout might enable the Democratic challenger to upset a Republican incumbent senator.
Hart, who has done surveys for MTV on 18 to 34-year-old voters, believes there will also be "an exceptionally large'' turnout among younger Americans.
Polls do capture the strong support these voters express for the Democratic ticket, yet he says that, based on past voting behaviour, these numbers are underestimated: "Unlike before, young voters this time recognise the importance of this election, both in their dislike of George W Bush and their support for Obama," he says.
Ann Selzer, a pollster whose survey last December presaged both the critical Obama victory in the Iowa caucuses and the huge increase in turnout, agrees that many polls are underestimating turnout among younger voters: "They are harder to reach, often not home and may have cell phones which aren't listed,'' and thus aren't included in the random calls of a survey.
The reason Selzer thinks reliable Republican states like Indiana are in play this time is precisely because younger voters and African-Americans will turn out in much greater numbers than expected.
Some polls - Gallup is most notorious - screen for likely voters in a fashion that almost assuredly undercounts blacks and young voters.
Overall, it's a good bet that turnout on Nov 4 will exceed 140 million, up more than 15 percent from the relatively high number of four years ago. For the first time since 18-year-olds were given the vote almost four decades ago, turnout might exceed 60 percent of eligible voters.
If so, this election will be seen as the one where the Bradley effect was replaced by the Obama effect.
Albert R Hunt is the executive editor for Washington at Bloomberg News. The opinions expressed are his own.
"The conclusion: there were a number of whites who, in the privacy of the voting booth, just wouldn't cast their ballots for a person of colour." Or maybe there were a number of whites who objected to the Democrat's policies but were afraid of telling pollsters that they might not vote for the black candidate. Why is it so hare for you leftists to believe some white people vote against Democratic party black candidates because the candidates are socialists not because they are black? Why is a refusal to support communism the same as supporting Nazism in your minds? Refusing to vote Democrat is not the same as white supremacism.