THOSE POLLSTERS JUST DON’T UNDERSTAND RACE
Friday, January 25, 2008.
By Keith Boykin
Two weeks after the election, analysts are still scratching their heads trying to come up with an explanation for the stunning disparity between the pre-primary polls and the final results in the New Hampshire primary.
The prevailing theory seems to blame the problem on the unusually large number of undecided voters who made up their minds at the last minute. Senator Hillary Rodham Clinton won that group.
There’s another theory that has gotten some consideration in the black community but hasn’t been taken as seriously by the pollsters. It's the longstanding problem of polling when black candidates run for office.
Opinion polls have over-counted support for black candidates in several closely watched elections over the years. This problem first surfaced when Los Angeles Mayor Tom Bradley lost his 1982 bid to become California’s first black governor. The polls showed Bradley, a Democrat, with double-digit leads, but he was defeated by white Republican George Deukmejian. The phenomenon became known as "the Bradley effect."
In 1989 we saw it again. The two most famous examples of the disparity between poll numbers and election results took place with two African American candidates who actually won. Douglas Wilder narrowly defeated white Republican Marshall Coleman in the 1989 Virginia governor’s race after pre-election polls and one exit poll showed him with double-digit leads.
On the same day, David Dinkins narrowly beat white Republican Rudy Giuliani by two points in the New York City mayor’s race after leading by 18 points in a newspaper poll a week before the election.
The polling disparity doesn’t always show up on election day.
In the last major election cycle in 2006, the polls accurately predicted two of the most closely watched races with African American candidates. Black Democrat Rep. Harold Ford, Jr. narrowly lost to white Republican Bob Corker in the Tennessee Senate race, and black Democrat Deval Patrick easily defeated white Republican Kerry Healey, just as the polls had predicted.
And then there’s Iowa.
ABC News polling director Gary Langer pointed to the accuracy of this year’s Iowa polls as evidence that race is not a likely factor in disparities between the polls and results in New Hampshire:
“There have been previous races that misstated support for black candidates in biracial races. But most of those were long ago, and there have been plenty of polls in biracial races that were accurate,” Langer argued. “And there was no overstatement of Obama in Iowa polls,” he said.
Similarly, Frank Newport at Gallup seems to dismiss the possibility of a racial gap and again points to Iowa as proof. “Discrepancies between pre-election polls and the actual vote for black candidates, in fact, are certainly not the norm," he writes. "This year, as an example, the pre-election polls in Iowa were very accurate in relationship to Obama's actual vote in the caucuses.”
Both pollsters make good points, but Iowa is not New Hampshire. The Iowa polls may have been accurate because Iowans vote in public caucuses. The New Hampshire polls may not have been accurate because primary voters cast their ballots in private.
Two weeks ago Jon Stewart interviewed pollster John Zogby on “The Daily Show,” and the issue of race never came up as a possible explanation for the polling failure. Zogby cited the 18 percent of New Hampshire voters who made up their minds at the last minute as a possible reason.
But Andrew Kohut, the president of the Pew Research Center, challenged that logic. Among the voters who said they made their decision on election day , Clinton’s margin of victory was “very small,” Kohut wrote in the New York Times. Clinton won 39 percent of the late deciders while 36 percent went for Obama, a gap which Kohut says is “too narrow to explain the wide lead for Mr. Obama that kept showing up in pre-election polls.”
Kohut is one of the few pollsters who seems willing to acknowledge the possibility of a race-based explanation. The others seem to dismiss the idea before giving it any serious consideration.
Race may be the answer, or maybe it’s not. At this point we just don’t know if race was a factor in the polling disparity. But it’s too early to dismiss the possibility. This is an unprecedented election. With no incumbent from either party, a viable black candidate, a viable woman candidate, a viable Mormon candidate, and incredibly high voter interest and turnout, the 2008 election is taking the country into uncharted territory.
Since Saturday's Nevada election was a caucus, South Carolina will be the first Democratic primary since New Hampshire. Barack Obama is widely expected to do well in South Carolina with the help of high black voter participation.
But going into Super Tuesday on February 5, we will really get to put the polls to the test. Will the polls accurately predict the outcome in states where Obama is predicted to do well, or will the people tell the pollsters what they think they want to hear?
America has made considerable progress in race relations in the past 50 years, and many of us would like to think that race doesn’t matter. We’ve elected black governors and senators, and appointed black cabinet members, diplomats and military officers at the highest levels of government. But the principle of racial equality has yet to be tested at the highest level of public office.
If we really want to know how race affects politics, we need not jump to conclusions either way. Instead, we should observe the upcoming state primaries and carefully watch what voters say to the pollsters.
And then watch what they do.
Keith Boykin is a writer, broadcaster, journalist and political commentator. He appears regularly on CNN and several major television networks. He blogs at Keithboykin.com
Please e-mail comments to comments@thenewblackmagazine.com