Can You Trust What Polls Say about Obama's Electoral Prospects? Two Important Trends Suggest Americans May Now Be Ready to Elect an African American President
by Scott Keeter and Nilanthi Samaranayake February 7, 2007
The strong showing of Democrat Barack Obama in early trial heat polls for the 2008 presidential election raises anew the question of whether the American public is ready to support an African American candidate for president. Recent polling points to two significant shifts on this question.
The first is that an ever larger majority of the public indeed says that they are willing to vote for an African American for the nation's highest office. The second is that polls conducted in campaigns pitting white and black candidates against each other are doing a better job of accurately predicting the outcome of the election now than in the past, suggesting that hidden biases that confounded polling in biracial elections in the 1980s and early 1990s are no longer a serious problem.
With Obama poised to declare his candidacy for president this weekend, recent national polling finds that although he trails Hillary Clinton for the Democratic nomination, he does nearly as well as Clinton in general election matchups against the frontrunning Republicans, narrowly leading John McCain and running roughly even with Rudy Giuliani.
More generally, the vast majority of Americans tell pollsters that they are willing to vote for a qualified African American candidate for president. In a Newsweek poll conducted last year, just 3% said they would not do so. This was not always the case. Gallup has asked a version of this question since 1958. Most recently, in 2003, 92% said they would vote for a black candidate for president while just 6% said they would not. But in 1958, a majority of 53% said they would not vote for a black candidate; even as recently as 1984, 16% told Gallup they would not do so.
Can polls that show the public willing to vote for a black candidate be taken at face value? It is undoubtedly true that racial attitudes in the U.S. have become more tolerant over the past five decades, and African American candidates have won high office in many states. But it is also true that the expression of racist attitudes is less socially acceptable now than in the past. This may lead some people to tell pollsters that they are more tolerant than they actually are.
Election Polls in the 1980s and 1990s Missed the Mark in Biracial Elections
Problems with pre-election polls in several high-profile biracial elections in the 1980s and early 1990s raised the question of whether covert racism remained an impediment to black candidates. White candidates in most of these races generally did better on Election Day than they were doing in the polls, while their black opponents tended to end up with about the same level of support as the polls indicated they had.
This phenomenon was first noticed in the 1982 race for governor of California, where Los Angeles Mayor Tom Bradley, a black Democrat, narrowly lost to Republican George Deukmejian, despite polls showing him with a lead ranging from 9 to 22 points. The next year, African American Democrat Harold Washington barely won his race for mayor of Chicago against Republican Bernard Epton. Pre-election polls taken within the last two weeks of the campaign showed Washington with a 14- point lead.
Three highly visible races in 1989 and 1990 also followed this pattern, though in two instances at least one late poll signaled a close race. Virginia Democrat and African American Douglas Wilder edged white Republican Marshall Coleman by less than one percentage point to become the nation's first elected black governor. But two of three polls conducted just days before the election showed Wilder leading by double-digits; a third poll had him 4 points ahead.
Even an exit poll conducted on Election Day showed Wilder winning by 10 points, while accurately tallying the vote in the other two statewide races. Unlike most exit polls that use an anonymous written ballot to collect voters' responses, this one had interviewers asking voters face-to-face how they voted, a situation that might increase the pressure to provide a socially desirable response.
Also in 1989, Democrat David Dinkins, an African American, won victory over Republican Rudy Giuliani in the race for mayor of New York by a slight two points, despite leading by 18 points in a poll conducted by the New York Observer a week before the election.
In the following year, another prominent election featured African American Democrat Harvey Gantt in a bitter race against Republican Sen. Jesse Helms of North Carolina. Two of three independent polls conducted just before the election showed Gantt leading Helms, but Helms prevailed by six percentage points on Election Day. Race was an issue in the campaign as evidenced by a Helms campaign television advertisement featuring a fictional white job seeker who lost out to a minority candidate because of a racial quota. The ad charged that Gantt supported hiring quotas.
In 1992, black Democrat Carol Moseley Braun won a 10-point victory over Republican Richard Williamson in a race for a U.S. Senate seat from Illinois. But polls taken just days before the election showed her with a lead ranging from 17 to 20 points. Interestingly, in the Democratic primary, Moseley Braun trailed incumbent Sen. Alan Dixon by a wide margin in two polls taken just a week before she won the election.
2006 Polls More Accurate
Last year's midterm elections featured several important races that pitted black and white candidates against each other. Unlike the experience of the 1980s and 1990s, pre-election polls in most of these campaigns performed well, and there was little evidence of a "hidden" vote for the white candidate.
Although African American candidates lost four of the five statewide races that featured black vs. white candidates, the late pre-election polling tended to mirror the final outcome.1 Black Republican candidates for governor lost by wide margins in Ohio (by 23 points) and Pennsylvania (20 points), but the average of the final independent polls in each state showed similar margins (21 and 23 points, respectively). An African American Democrat, Deval Patrick, won the Massachusetts governor's race by a landslide (56% to 35%) over a white Republican Kerry Healey. Two pre-election polls slightly underestimated Healey's support, but these were conducted about two weeks before the election.
Black candidates also lost in two key Senate races -- Maryland and Tennessee -- but there was no clear evidence of a hidden vote for the white candidate in either state. The more complicated case was in Maryland, where Republican Lt. Gov. Michael Steele, who is African American, lost his Senate bid to Democrat Benjamin Cardin by 10 points, about the same margin as in a Washington Post poll conducted 10 days before the election. Two other polls, however, showed the race to be much closer. But these polls also underestimated the Democratic vote in the race for governor in which both the Democrat and Republican candidates were white. Both polls showed the two candidates running neck-and-neck, but on Election Day the Democrat, Baltimore Mayor Martin O'Malley, beat incumbent Republican Gov. Robert Ehrlich by seven points.
The race for the U.S. Senate in Tennessee was perhaps the most closely watched of all the biracial elections last year. Black Democrat Harold Ford, Jr. narrowly lost to white Republican Bob Corker, 51% to 48%. There were many pre-election polls in this race, and three of the final four polls actually overstated Corker's lead.
Taken together, the accuracy of the polling in these five biracial elections suggests that the problems that bedeviled polling in the 1980s and early 1990s may no longer be so serious. This change is not a result of broader improvements in the methodology of election polling; most election polls in the earlier period were competently done and generally performed well in predicting election outcomes.
The experience of the 2006 elections indicates that racism may be less of a factor in public judgments about African American candidates than it was 10 or 20 years ago. It is true that the African American candidate lost in four of the five statewide races examined, but three of these were Republicans running in a bad year for Republicans. In each of these three cases, the other major statewide race pitted two white candidates (for the U.S. Senate in Ohio and Pennsylvania and for governor in Maryland), and the Democratic candidates' margin of victory was similar to those in races involving a black Republican.
It is also the case that black candidates in these races tended to do as well among whites of their party as white candidates did in other states. For example, the National Election Pool exit poll -- which is conducted with an anonymous ballot and thus less likely than a telephone or face-to-face interview to elicit a socially desirable but erroneous response -- found that 91% of white Democratic voters in Tennessee chose Harold Ford, about the same level of support that white Democrats in Virginia gave Jim Webb (92%). This is also about the level of support that white Republicans in Maryland gave African American Michael Steele (94%).
No one would deny that race still matters in U.S. politics. For the past half century, the political parties have been increasingly divided in their positions on racial issues, and that, in turn, has affected voters' decisions to call themselves Republicans or Democrats. But this review of exit polls and electoral outcomes in several recent elections suggests that fewer people are making judgments about candidates based solely, or even mostly, on race itself, and that relatively few people are now unwilling to tell pollsters how they honestly feel about particular candidates. In such an environment, the high standing of Barack Obama in presidential polling -- or, for that matter, of Colin Powell prior to the 1996 presidential election -- represents a significant change in American politics.
-----
Notes
1 The Senate contest in Mississippi matched a black Democrat (Fleming) and a white Republican (Lott), but almost no public polling was conducted in the race.