March 3, 2008
How White Men Got Over Race and Boosted Barack, Big Time
Former Bill Clinton pollster says: Obama campaign has rewritten
by Craig Charney
Former Bill Clinton pollster says: Obama campaign has rewritten the rules
Eight weeks ago, when Barack Obama lost lily-white New Hampshire by 3 points after leading in the polls there by 10, many questioned whether Obama could win among whites.
Pundits asked if the “Bradley effect” noted in the 1980s had returned. Back then, black candidates like Tom Bradley in California and David Dinkins in New York did much worse on Election Day than they had in the polls.
The lesson of the past month’s primaries and caucuses is, resoundingly, this: The importance of race in American voting really has declined. The Bradley effect is history. Whites no longer exaggerate their support for a black candidate.
And the unsung heroes of this historic shift are America’s most maligned minority: white men.
How do we know that whites’ secret reluctance to vote for a black has faded? If the Bradley effect still worked, preelection polls should have continued to overestimate Obama’s chances, as they did in New Hampshire. But my firm’s analysis of 49 election-eve polls published before Super Tuesday in 16 states gives the lie to the “lie factor.”
Few of the final polls before the Feb. 5 vote made pro-Obama errors. Four-fifths got the winners right — and mistakes favored Clinton more often than Obama.
We also checked whether the white candidate’s vote rose more than the black’s from the final poll to election day, as happened in several Bradley effect cases two decades ago. Yet on and since Super Tuesday, Obama’s gains when actual votes are compared to preelection polls exceeded Clinton's in 10 of 13 states where we analyzed poll findings.
In the most recent major test, in overwhelmingly white Wisconsin, Obama’s actual vote jumped 9 points from the average of the election-eve polls, while Clinton’s slumped by 3 points.
Surprisingly, the most accurate polls used robo-calls: people seem likeliest to tell the truth if there's no one at the other end! But the old days, when whites routinely told interviewers they'd vote for blacks but balked at pulling the lever, are over.
Many have noted Obama’s appeal to young whites: Of those 18 to 29, he took 63% in New York, 56% in Georgia, and 72% in Virginia. But they form only a small share of the voters.
The real core of Obama’s white support is white men, who are a quarter to half the electorate in most states. In Connecticut, for instance, he swept them, 57% to 40%. The same happened in California and Massachusetts, although this was offset by Clinton’s big leads among white women (and in California, Latinos), giving her the state.
Even in border and southern states, polls show that after John Edwards withdrew, white men switched in droves to the black candidate. Obama took them by almost 3 to 2 in Virginia.
Exit polls show Obama carried white men nationally on Super Tuesday, and they have been vital to Obama’s victories then and later. He has won eight of the 12 states where he led or tied among white men, but just two of 12 where he lost them.
He won Wisconsin by 17 points because he blew Clinton out among white men, 63% to 34%.
Yes, race still does matter. Obama and Clinton do better among voters of their own race, while some Hispanic voters have hesitated to back the black candidate. Moreover, the conservative white voters dominant in the general election are less likely to vote for a liberal black candidate than those in Democratic primaries. They might be more reluctant to say so, too (though general election races with black candidates in 2006 showed no Bradley effect).
But this year has shown that something has changed in U.S politics. Race is no longer the trump card. It has become one more factor in the mix, along with gender, age and ideology — even in the South. Whatever else happens in this election, this is a big step forward.
Charney is president of Charney Research, a New York polling firm. He was senior analyst on President Bill Clinton's 1996 polling team.
< Return to CHARNEY RESEARCH STAFF PUBLICATIONS >