Maybe race wasn’t the only factor

Decker is a Times staff writer.

It has entered political lore as the “Bradley effect” -- the supposed tendency of some white voters to lie when asked if they support a black candidate, producing a bubble of support that isn’t really there.

Named for the precipitous defeat in the 1982 California governor’s race of Los Angeles Mayor Tom Bradley -- after polls suggested his victory -- the effect has led to churning concern among backers of Barack Obama’s presidential campaign that the lead he holds will vanish come election day.

But the anxiety masks reality: Bradley’s narrow loss stemmed from a convergence of political difficulties for the mayor, who was then seeking to become the nation’s first African American governor, and only one of them was his race. Twenty-six years later, those engaged in that contest still differ on whether there was a Bradley effect.


More to the point for Obama, there is no evidence that one still exists. A recent study by a Harvard political scientist showed no sign since 1996 of an otherwise unexplained election day drop-off in support for African American candidates for governor or U.S. Senate.

That is not to say that race is not an issue, particularly as Obama seeks to become the first black president. Exit polls in primary states demonstrated that for many voters, Obama’s race was a stumbling block. But those voters were open about their views, suggesting that polls may be roughly accurate.

Joe Trippi, the deputy campaign manager for Bradley in 1982, thinks voter discomfort with the Democratic mayor’s race was key to his defeat but that those concerns have eased with time.

“Whatever doubts race caused 26 years ago, it doesn’t create the same level of doubt today,” Trippi said.

“Anyone who thinks it’s zero is kidding themselves,” he cautioned. “But it’s a hell of a lot closer to zero than it was. . . . I just don’t see this election as being close enough [for it] to matter.”

More than this campaign, the 1982 governor’s contest was fraught with the issue of race. It was less than a generation removed from the late-1960s riots in America’s cities, including Los Angeles, that sent fearful white voters scrambling for the suburbs. Bradley, the reserved, patrician mayor, a former police officer and city councilman, was running against George Deukmejian, the state’s Republican tough-on-crime attorney general and a former state legislator from Long Beach.

On the same ballot was a U.S. Senate race between Pete Wilson, then the GOP mayor of San Diego, and Jerry Brown, then the Democratic governor. Also on the ballot -- and this would matter more -- was Proposition 15, a measure that would have imposed statewide handgun registration and a freeze on new handgun sales. The candidates for governor lined up on opposite sides -- Bradley supported it, Deukmejian opposed it.

Intent on brushing back gun controls, opponents of the measure mounted a massive voter registration drive to draw gun enthusiasts to the polls. At the same time, Republicans took advantage of a change in state law that allowed any Californian to vote by absentee ballot. In previous governor’s races, only those with medical needs or travel plans could vote absentee.

The combination of more conservative gun-backing voters and more ways to get them to vote would help seal Bradley’s fate. In some rural, mostly white counties, turnout was up substantially, by 5% to 10%, recalled Bob Mulholland, a senior Democratic Party advisor. Two-thirds more absentee ballots were cast than in the previous governor’s race, and they were cast mostly by the older white voters who favored Deukmejian.

According to statistics collected at the time, Deukmejian beat Bradley among absentees by more than 113,000 votes. Bradley won among votes cast on election day by almost 20,000 ballots, but that left a 93,000-vote margin of victory for Deukmejian.

“Clearly the Democrats and the Bradley campaign did not take vote-by-mail as seriously as the Republicans did,” Mulholland said.

The gun vote was key, according to Los Angeles Times exit polls. Those voting against the gun measure, roughly two-thirds of the voters, voted strongly for Deukmejian. Those supporting gun control sided strongly with Bradley, but they accounted for only a third of the voters.

But that was not all the bad news for Bradley. Black voters, expected to demonstrate overwhelming support, instead cast a below-average number of ballots. That gave more electoral heft to whites, a Times exit poll showed.

Still, it was not the result, but the seeming surprise, that led to suggestions of a Bradley effect. Times preelection polls showed Bradley’s 14-percentage-point lead in September had been cut in half by October, suggesting Deukmejian was gaining ground. The Times did not survey after that point, about three weeks before the election. Internal polls by Deukmejian’s campaign showed that trajectory continuing until election day, when their polls showed the race tied.

But a Field Poll published three days before the election gave Bradley a 7-percentage-point lead, and surveys of voters leaving the polls showed a strong Bradley edge. Pollster Mervin Field publicly projected a Bradley victory as the polls closed. Times exit polls showed a single-digit Bradley victory -- as well as an erroneous lead for Brown in the Senate race.

Some campaign staffers on both sides suggest that polling errors created the impression that voters lied to cover discomfort with Bradley’s race.

“It’s a myth,” said Steve Merksamer, Deukmejian’s campaign director and, later, chief of staff. But, he added, “it’s a lot more compelling media narrative to say that there was latent racism.”

But the notion that voters had lied gained ground when similar slumps occurred in high-profile races by two other black candidates -- David N. Dinkins, running for New York mayor, and L. Douglas Wilder, running for governor of Virginia.

Dan Hopkins, a Harvard political scientist, studied every race involving a black candidate for governor or U.S. Senate from the Wilder year, 1989, until 2006. He concluded that the Bradley effect cost African American candidates 2 or 3 percentage points until the mid-1990s.

Since then, “we’ve really found no evidence for a systematic Bradley effect,” he said. He pointed particularly to the U.S. Senate race in Tennessee in 2006, where African American Democrat Harold Ford Jr. was defeated after a heated contest that ended with a racially inflected ad portraying him as a womanizer.

“If you were going to see a Bradley effect, it would have been there,” he said. “But there was no such effect.”

Hopkins said the shift occurred at the same time that crime ebbed as a voter concern, and as the nation enacted welfare reform -- taking off the table two issues that had been used to press voter buttons about African American candidates.

Charles Henry, a UC Berkeley professor of African American studies, looked at the 1982 election and declared that race was “the major factor” in Bradley’s loss. Henry notes that more liberal Democrats running on the same statewide ballot won -- but they were white.

Henry said that the passage of time raised some questions: Do people no longer lie to pollsters? And if the absence of issues like crime and welfare has helped African American candidates, will they be hurt if the issues reemerge? And then there is the impact of voters who, unlike the Bradley effect balloters, are openly reluctant to vote for an African American.

“You can certainly predict that people will vote against Obama because he is black,” Henry said. “I think there will be an anti-Bradley effect; some will vote for Obama because he is black. But there’s no precedent in history for that overcoming the vote against him.”

In Pennsylvania and Ohio, exit polls this spring showed 1 in 5 Democratic voters felt that the race of the candidate was important in deciding whom to back.

Those who thought race was important went overwhelmingly for Hillary Rodham Clinton.

Frank Gilliam, the dean of the UCLA School of Public Affairs who has studied voting patterns, said one of the frustrating things about the Bradley effect was the difficulty in figuring it out.

“This is like proving a conspiracy theory: You can’t because it’s secret,” said Gilliam, who thinks Obama could lose 2 percentage points to a hidden anti-black vote. But he also believes that worry about the Bradley effect may overshadow a groundbreaking aspect of this presidential race.

“There’s going to be millions of white people who are going to vote for Obama,” he said, “and that’s no small potatoes.”