PRINCETON, NJ -- Gender and education are both strong predictors of Democrats' preferences for their party's presidential nominee. Generally speaking, the more education a Democrat has, the less likely he or she is to support Hillary Clinton, and the more likely to support Barack Obama. Additionally, women are more likely than men to support Clinton, while men are more likely than women to support Obama.
An aggregate of Gallup Poll Daily election tracking interviews with Democrats, conducted from Feb. 1 through Feb. 7, shows that these two variables combine to become a powerful predictor of a Democrat's vote. Among the most highly educated Democrats -- those with postgraduate educations -- both men and women are more likely to support Obama than Clinton. Among the least educated -- those with no college experience -- both men and women are more likely to support Clinton than Obama.
The range of support for the candidates according to education and gender is as follows:




In short, education is a highly significant predictor of Democrats' vote choices, particularly among the two groups at the extreme ends of formal education -- those with postgraduate degrees and those with high school educations or less. Gender, too, is a predictor, but is essentially overwhelmed by the impact of education in the two extreme groups, such that both men and women who have postgraduate educations prefer Obama, and both men and women who have no formal education beyond high school fairly strongly support Clinton.
Survey Methods
These results are based on an aggregate of Gallup Poll Daily tracking interviews conduced Feb. 1-7, 2008. Each night, approximately 1,000 national adults, aged 18 and older, are interviewed by telephone. 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).
The aggregate of the Democratic nomination trial heats includes telephone interviews with 2,871 Democrats and Democratic leaners. One can say with 95% confidence that the maximum margin of sampling error for these results is ±2 percentage points.
Results based on the aggregates of Democratic men, Democratic women, and various subsets of these groups according to level of formal education are associated with higher margins of maximum sampling error.
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.
Click below for more on these topics: