Comment Central: Will race hurt Obama?

There is probably only one person now standing between Barack Obama and the presidency. His name is Tom Bradley; he is the black, Democratic former Mayor of Los Angeles, and he has been dead for ten years.

In 1982 Bradley ran for the governorship of California, and was expected to win by a wide margin. In the run-up to the election, polls gave the African-American candidate a lead of between 9 and 22 percentage points over his white opponent. On election day the first exit polls also predicted that he would win, and one newspaper even declared him the victor on its front page.

Bradley lost by more than 100,000 votes.

The conventional wisdom is that he was the victim of a hidden racial reaction. Many white voters told pollsters that they would vote for the black candidate but, in the anonymity of the voting booth, they did not. The impetus for the deception was not simple racism, but social pressure - white voters, it seems, did not want to appear racist by admitting that they would be voting for the white candidate rather than the black one.

At its simplest, the so-called “Bradley Effect” (which some dispute) holds that polls overestimate support for an African-American candidate because, when race is involved, voters misrepresent their intentions.

The effect is not unique to Bradley: in 1989 David Dinkins, the black Democrat, was expected to romp home in the New York mayoral election, but won by only a narrow margin; in the same year Douglas Wilder, another black candidate, was 11 points ahead the day before the Virginia gubernatorial election, and won by only 0.5.

This, then, is the spectre that haunts the Democrats. Mr Obama is at present up to 14 points ahead in the national polls, a lead that would seem unassailable were it not for the unpredictable Bradley Effect, a nagging fear that defeat could be snatched from the jaws of victory because opinion polls behave differently when a black candidate is running.

A similar phenomenon occurred in the 1992 election in Britain, which pollsters put down to the “shy Tory factor”. Opinion polls put the Tories one percentage point behind Labour; but in the final result, the Conservatives won by nearly eight percentage points.

In certain, rare circumstances voters do not tell pollsters the truth, more out of embarrassment than mendacity.

The hidden effect of race is even harder to predict in the Obama-McCain battle, since we are in uncharted polling waters - there has never been a black versus white competition at a national level in the US before. Complicating matters further, those people most likely to be prejudiced against a black candidate are also those least likely to talk to pollsters and answer surveys: there may be an anti-Obama block out there that is simply not showing up on the radar.

Is it possible that a submerged racial iceberg could still scupper the Obama ship just as it appears to be cruising into harbour?

The latest evidence - anecdotal, statistical and historical - suggests that although the ghost of Bradley still hovers over the election, its effect is diminishing. A Gallup poll in 1958 found that 58 per cent of whites would not vote for a black candidate; by 1989, that figure had dropped to 19 per cent; by 2007, only 5 per cent felt the same way. Even taking into account the taboo surrounding overt expressions of racism, that is a stunning change.

This is not a post-racial election: some whites will still be motivated by old-fashioned prejudice, and others by its queasy residue most often expressed as “doubt”. Yet race has faded, as has the social stigma in stating a preference for the white candidate over the black one. There is less pressure to lie to pollsters than there was even a decade ago.

In the 1980s and early 1990s racially charged issues of crime and welfare were central to the political debate. Today those concerns have faded. This election is principally about the economy, and despite the efforts of some Republican strategists to stoke the fires, race does not have the political purchase it once had.

Michelle Obama made a sound point last week: “If there was going to be a Bradley effect, or it was going to be in play, Barack wouldn't be the nominee.”

Indeed, there could even be a Reverse Bradley Effect in operation: support for the black candidate may, in fact, be higher than polls suggest in some areas, particularly in regions, such as the South, where it is more socially acceptable for whites to voice distrust of blacks. There may be two competing deceptions at work here, equally hard to trace: a black lie (“I say I am going to vote Obama, but secretly I will not.”) and a white lie (“I say I am not going to vote Obama, but secretly I will.”)

In the Democratic primaries, pollsters found the Reverse Bradley Effect (ie, an underestimate of the Obama vote) in 12 states, and the Bradley Effect in only three.

Yet many Democrats remain anxious that the secret ballot on November 4 will reveal, once again, America's secret racial fears. America's attitude to race has changed: how much it has changed will be measured by the Bradley Effect, or lack of it.

If, once again, the polls prove to have overestimated Mr Obama's support by a wide margin, then the conclusion must be that voters have not only lied when asked whether they will vote for a black man, they have lied on a scale never seen before.

But if, as most experts claim, the polls turn out to be an accurate reflection of voters' intentions, then this will finally show, in the words of Mr Obama himself, that America is no longer “irrevocably bound to a tragic past”. The Bradley Effect will have proved ineffectual, and the unquiet ghost of Tom Bradley will finally be laid to rest.

Times Recommends

• Farwell to the stiff upper lip
• It’s about who runs the US, not health
• The Iron Lady and her blue felt-tip sympathy