Facebook
Twitter
RSS
Pew Research Center for the People & the Press
The South Carolina Democratic Primary in Black and White
The results in Saturday's Democratic primary in South Carolina offer important evidence -- if not yet answers -- to three big questions in this campaign: Can Barack Obama solidify all segments of the black vote behind him? Can he be competitive among white voters, especially in the South? And can we trust pre-election polls in races that involve Obama?
The Black Vote
There was no hint of any class division regarding Barack Obama among the South Carolina black electorate. He won by similar, overwhelming margins in every demographic segment. Old, young, male, female, well educated, poorly educated, affluent, struggling -- Obama got 70%-80% of the black vote across the board. Clinton took about one-in-five black votes (19%). Edwards had virtually no support among African Americans.
Black voters acknowledged the polarizing nature of the campaign. Nearly half said both candidates attacked each other unfairly. Even among Obama's black voters, nearly half said he attacked unfairly -- but 72% of his supporters said Clinton did so. Three-fourths of Clinton's black voters said Obama attacked her unfairly, but an identical number of them agreed that Clinton behaved the same way.
Most African American voters (83%) said they thought the country was ready to elect a black president, but just 56% said the country was "definitely ready." Somewhat fewer (70%) said the country is ready for a female president -- and just 31% said the country was definitely ready.
The White Vote
While Obama nearly matched his rivals among affluent and better educated whites, he struggled among the less affluent and less educated, as well as those ages 30 and older. In future primary states where whites will constitute a much larger share of the Democratic electorate -- even where Edwards may be weaker than he was in his native South Carolina -- Obama faces substantial challenges in matching Hillary Clinton for the votes of working class and older white voters.
Obama ran reasonably well among college-educated whites, getting about one third of the vote (32%, vs. 35% for Edwards and 33% for Clinton). In contrast, he did poorly among low-education whites, winning just one in 10 (10%) among those with a high school education or less.
Clinton won among white women (42%, vs. 36% for Edwards and 22% for Obama), but trailed among white men. Obama again scored well among the youngest voters -- in this case getting half (50%) of the vote of whites under 30. But he trailed both Clinton and Edwards among voters 30 and older, and received just 16% of the vote among those ages 60 and older.
As pre-election polls suggested, John Edwards benefited from a late swing of voters his way. He won 52% among the roughly 1-in-3 white voters who said they made up their mind in the last week of the campaign. Clinton, by contrast, got 51% among white voters who decided more than a month ago.
There was mixed evidence that Edwards's populist appeal connected disproportionately with downscale white voters. He led by over 60 percentage points among the white voters who said they picked the candidate who "cares about people like me." But he did not do notably better among less affluent whites, nor among those who said the economy is doing poorly.
The Pre-Election Polls
There is at least a hint from South Carolina that the kind of polling failure seen in the New Hampshire Democratic primary may have been unique to that state, and may not portend a return of the so-called "Bradley effect" in which the likely performance of black candidates relative to white ones is overestimated.
But the polling in South Carolina encountered a different kind of problem. The final public polls substantially underestimated Obama's margin, largely because they significantly understated his support among black voters. Some may also have underestimated the size of the black turnout. The size of the error in forecasting Obama's victory margin was actually greater than the size of the error in New Hampshire, but in the latter case, of course, the polls also forecast the wrong winner. Primary polls often have difficulty gauging the enthusiasm of the supporters of different candidates, and South Carolina may have presented a particularly difficult test, given the large black Democratic constituency and the racial dynamics of the campaign in its last week.
From the point of view of those worried about the possibility that polls might be missing white respondents with more negative racial attitudes, or not accurately capturing such attitudes among the respondents that they do interview, the South Carolina result is reassuring. None of the South Carolina pre-election polls overstated Obama's support among white voters or understated white support for his opponents (estimates of Obama's share of the white vote ranged from 10%-21% - he got 24%).
However, as with New Hampshire, South Carolina represents just one election, and a more definitive test is coming on February 5.