PRINCETON, NJ -- Democrats are at most risk of losing the support of independents, conservative Democrats, and, among Hillary Clinton supporters, less well-educated Democrats if those voters' preferred candidate -- Clinton or Barack Obama -- does not win the party's nomination. Black Democrats appear loyal to the party regardless of who wins the nomination.
The finding that sizable percentages of Democrats say they would vote for Republican John McCain next November if the Democratic nominee is not their preferred candidate raises interesting questions about exactly whom the Democrats are most at risk of losing in the general election.
To answer those questions, Gallup analyzed the basic relationship between Democratic candidate support and the current general election vote within various subgroups of the population.
Clinton Nomination Supporters
The accompanying table displays the percentage of Clinton supporters who say they would vote for McCain if the general election matchup were McCain versus Obama.
There are clear differences by subgroup in self-reported vote for McCain under an Obama-wins-the-nomination scenario. In other words, it appears that if Obama is on the ticket, some groups of Clinton-supporting Democrats are more susceptible to bolting the party and voting for the Republican nominee next fall than are others.
The average "defection rate" of Clinton-supporting Democrats away from Obama and to McCain in the general-election matchup is 28%. The two groups of Clinton-supporting Democrats who are significantly above this average in defection to McCain are independents who lean Democratic and conservative Democrats.
The groups that are below average in potential defection to McCain if Obama were running against him are blacks, liberal Democrats, those with postgraduate educations, and core Democrats (i.e., those who identify as Democrats when asked their party identification).
Clinton-supporting Democrats with the most education are the least likely to say they would vote for McCain if he were matched against Obama next fall. Those with less formal education are most likely to vote for McCain.
Among Clinton supporters, there are only small differences by gender in voting for McCain if Obama is on the ticket. There are no significant age differences.
Obama Nomination Supporters
The percentages of Democratic voters who support Obama for the nomination but who would vote for McCain if Clinton is the nominee are lower across most subgroups than is the case for Clinton supporters (in reference to a McCain-Obama race, as reviewed above). This is a reflection of the basic finding that Obama supporters are less likely to abandon their preferred party and vote for McCain -- even if their candidate does not get the Democratic nomination -- than is the case for Clinton supporters in the reverse scenario (19% of Obama supporters would vote for McCain if Clinton is the nominee, compared to 28% of Clinton supporters who would vote for McCain if Obama is the nominee).
But there are differences across subgroups of Obama supporters in their intentions to vote for McCain.
Obama supporters who would be most likely to support McCain in a McCain-Clinton race are independents who lean Democratic, conservative Democrats, moderate Democrats, and non-Hispanic whites.
Those least likely to bolt the party and vote for McCain are blacks, liberal Democrats, and core Democrats.
There are no significant age differences.
Across the board, the data show that Democratic support in the general election is more at risk among some subgroups of voters than among others. In particular, independent voters who lean Democratic are more likely than any other subgroup tested to say they would vote for McCain if their candidate does not gain the nomination. Additionally, conservative Democrats appear to be less attached to the party than are liberal Democrats, and more willing to say they would vote for McCain if their candidate is not the nominee.
Almost 4 out of 10 voters in these two groups who support Clinton say they would vote for McCain if Obama is the nominee. The percentages are still high, but about 10 points lower, for voters in these groups who support Obama when asked about a McCain-Clinton contest.
These findings are not necessarily surprising, but underscore Democrats' vulnerability with voters who are positioned somewhat more in the middle of the political or ideological spectrum. This may also reflect McCain's strong appeal to independent voters, who may not need much nudging to shift their vote from a Democratic candidate to McCain.
Black Democratic voters, regardless of whom they support, seem prepared to remain quite loyal to the Democratic Party. Fifteen percent of blacks who support Clinton would vote for McCain if Obama is the nominee, and only 10% of blacks who support Obama would vote for McCain if Clinton is the nominee. In other words, there is little apparent risk of losing a substantial proportion of black voters regardless of who the nominee is.
This last finding is significant. Obama has the overwhelming support of black Democratic voters at this point, and there has been discussion of the backlash that could occur if Obama were to lose the nomination to Clinton. But these data suggest that Clinton could still expect to receive the vote of most black Obama supporters were she to win and face McCain in the fall. (The data do not address the issue of motivation or turnout, which could be lower among blacks if Obama is not the nominee, nor do the data address the implications of the precise way in which Clinton might win the nomination. If Clinton were to win by the vote of superdelegates, for example, the blowback from black Obama supporters might be greater than if she were to win by gaining the highest percentage of the popular vote cast in primaries and caucuses.)
The data show an inverse relationship between education and Clinton voters defecting to McCain if Obama is the nominee. Thus, an Obama win of the Democratic nomination runs a risk of the Democratic Party losing the November support from less well-educated Democrats who support Clinton.
There is no significant gender difference evident in the data. Both men and women tend to mirror the overall sample patterns in terms of projected vote for McCain, with only minor differences. This may be surprising to some who might expect that women who support Clinton (who would be the first female president in U.S. history if she is nominated and wins the election) would be less loyal to the party if their candidate did not win the nomination. But this is not the case.
Although there are no significant differences in willingness to vote for McCain across age groups, a more detailed analysis suggests a greater hesitation to vote for the Democratic candidate as age increases -- in a situation in which the voter's candidate is not the nominee. This occurs because the "undecided" vote goes up with age. Older Democrats are no more likely to vote for McCain if their candidate does not win than are those who are younger, as noted in the previous sections of this analysis. However, it appears that older voters are less likely to vote for the Democratic candidate when he or she is not the one they support. In other words, instead of declaring a vote for McCain in these situations, these older voters are more likely to be undecided.
The basic difference between Clinton and Obama supporters is observed across all three age groups. In each instance, Clinton supporters are less likely to support Obama against McCain than are Obama supporters to support Clinton against McCain.
Results are based on telephone interviews with 6,657 national adults, aged 18 and older, conducted March 7-22, 2008. For results based on the total sample of national adults, one can say with 95% confidence that the maximum margin of sampling error is ±2 percentage points.
Interviews are conducted with respondents on land-line telephones (for respondents with a land-line telephone) and cellular phones (for respondents who are cell-phone only).
In addition to sampling error, question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of public opinion polls.