by Andrew Kohut, President, Pew Research Center

Voting Machine

When evaluating Sen. Hillary Clinton’s 2008 prospects, one question remains inescapable: Are voters ready to elect a female president? On one level, the question would appear to be settled. In February, Gallup found 88% saying they would vote for a well-qualified woman for president. Contrast this with public opinion in 1969, the year Clinton graduated from Wellesley College. At that time, just 53% said they would support a well-qualified female presidential candidate.

As to opinion of Clinton herself, surveys by the Pew Research Center for the People & the Press and other national polling organizations show a consistent gender gap in her support. Recent Newsweek surveys show that in a hypothetical contest against Republican frontrunner Rudolph Giuliani, half of female voters, compared with just 44% of male voters, said they would vote for Sen. Clinton.

Table

A gender gap in opinions of Hillary Clinton is not new. It has been apparent since she rose to national prominence as one of the most politically involved first ladies in history. The pattern has been so consistent for such a long time, it raises the question of whether this is strictly a Clinton phenomenon, or whether it is characteristic of how American voters generally respond to female candidates.

Around the world, women are no strangers to contests for the top leadership position – France’s Ségolène Royal being only the most recent, though unsuccessful, contestant. And women have been victorious in such elections in countries as varied as India (Indira Gandhi), Israel (Golda Meier), and Great Britain (Margaret Thatcher). But elections in other countries offer little or no guidance as to likely U.S. voter reactions to a female candidate among either men or women.

A better guide may be voting patterns in statewide elections. While these are not necessarily indicative of voters’ willingness to support a woman in a nationwide presidential contest, they are probably the most relevant indication of the mindset of the current American electorate in this regard.

Table

A new Pew analysis of voting patterns for female Democratic candidates for U.S. senator and state governor in recent years suggests that the gender differences in support for Clinton at this early stage in the campaign are, on average, typical for Democratic women who run for statewide office. Pew examined exit polls from 40 senatorial and gubernatorial elections since 1998 in which a female Democratic candidate faced a male Republican. On average, the female Democratic candidates captured the votes of 55% of women compared with 47% of men; the average eight-point gender gap in these contests is identical to that in the match-up between Clinton and Giuliani in the Newsweek survey.

When the 40 races involving female Democrats are compared with a control sample of elections from the same regions involving male Democratic candidates, the Democratic women performed at least as well as the men. On average, Democratic female candidates for senator and governor won 51% of the vote in their elections; Democratic male candidates in the control group won 47% of the vote.

There are no signs that female Republican candidates receive a similar advantage. In the 15 races observed where a female Republican faced a male Democrat, the Republican women did no better than a comparable sample of male Republican candidates. More specifically, women voted for female Republican candidates about as often as they did for male Republicans.

Patterns of gender support in Senate and gubernatorial races suggest that the strong backing that Mrs. Clinton is getting from women is typical for female Democratic candidates for high office. Whether this will extend to the presidency remains an open question. When voters are asked to resolve the question, 13% said they would be more likely to support a female presidential candidate while 11% said they would be less likely to vote for her. And there was the familiar gender gap in response even to this hypothetical question. By a margin of 19% to 7%, women said they would be more likely, rather than less likely to vote for a woman. Men split evenly on the question.

Methodology

The analysis of statewide election results for male and female candidates combines election outcomes and exit poll results for all races in which one candidate was female in the election years 1998, 2000, 2004 and 2006 (2002 was omitted because of problems with the exit poll data that year). This resulted in 55 Senate or gubernatorial races, 15 with a female Republican vs. a male Democrat, 40 with a female Democrat vs. a male Republican. Races with two female candidates were separated from this analysis. Another 55 races with two male candidates were selected to serve as a control group for the analysis. To ensure comparability, states were stratified regionally, and male vs. male statewide races were randomly within each region in proportions that mirror the share of female vs. male statewide races from each region of the country.