To obtain print-quality JPEGs, contact the Office of Public Affairs at (815) 753-1681 or e-mail email@example.com.
Contact: Tom Parisi, NIU Office of Public Affairs
January 22, 2007
DeKalb, Ill. — Polls consistently show that about nine of 10 Americans say they would vote for a qualified female presidential candidate. But new research indicates a significant percentage of survey respondents are hiding their true feelings to avoid the appearance of being sexist.
A team of political scientists led by Matthew Streb of Northern Illinois University found that slightly more than one in four Americans would likely be unsupportive of a female candidate for president. The findings were consistent among both male and female respondents across several demographic groups.
“Our results indicate that a significant percentage of people are hiding their true feelings on questions related to female candidates for the presidency,” Streb said. “Consequently, public opinion polls tend to exaggerate support for a female president.
“We believe our findings have major implications, both with regard to the future of women in political office and to the accuracy of public opinion polls on certain issues,” he added. “While women candidates seem to be making some strides in races for many offices, including executive positions such as governor, the office of the presidency may be difficult to reach.”
The new research could have implications in the 2008 presidential election, as U.S. Sen. Hillary Clinton (D-N.Y.) is seen as a frontrunner among likely contenders for the Democratic nomination.
“It should be a concern,” said NIU political scientist Barbara Burrell, a member of the research team. Burrell, former associate director of NIU’s Public Opinion Lab, is working on her second book on women in politics and is president elect of the Women’s Caucus in Political Science.
“It’s going to be even harder than we might have imagined for a woman running for president to overcome certain stereotypes,” Burrell said. “Of course, Hillary Clinton could still win, but she may have a steeper hill to climb than some observers believe. The study shows gender is a more significant factor than we might have suspected.”
The team of researchers also included NIU graduate student Brian Frederick and Professor Michael Genovese of Loyola Marymount University. The study will be published later this spring in the journal Public Opinion Quarterly.
The team examined the effects of “social desirability” in polls seeking to determine public support for a female American president. Social desirability is the tendency of poll respondents to lie when they believe their true answers go against perceived societal norms. The effect has been previously documented on sensitive polling issues involving religion, race and gender. For example, it’s common for poll respondents to overreport church attendance.
The 1989 Virginia governor’s race featured African American Douglas Wilder and his white opponent Marshall Coleman. Although polls showed Wilder with a commanding lead, he won the race by a narrow margin (50.1 percent to 49.8 percent). The 1989 New York City mayoral election between David Dinkins and Rudolph Giuliani followed a similar pattern. Studies attributed the inaccurate pre-election polls to problems of social desirability.
“Polling is usually fairly accurate, but there is a long line of research supporting the influence of social desirability, or the tendency of respondents to provide socially desirable responses to questions dealing with controversial issues,” Burrell said. “Most public opinion polls don’t tease out whether people are hiding their true feelings.”
In order to do just that, the researchers employed a well-established public-opinion technique known as “the list experiment” to measure public sentiment on female candidates for president.
A baseline group of survey respondents was asked how many of the following four statements make them “angry or upset:”
A fifth statement was added to the list given to the test group:
The researchers attributed an increase in the mean number of angry items in the test group to the “woman serving as president” statement. The list experiment question was added to an Opinion Research Corporation CARAVAN telephone poll, a national survey of Americans conducted in March 2006. Of the 2,056 respondents, roughly half were given the test question.
“Traditional polls find anywhere from 5 percent to 15 percent of the public say they will not vote for a female presidential candidate,” Streb said. “We found that roughly 26 percent of respondents expressed anger over a female president. It is extremely likely that those who are angered by the prospect of a female president are not going to vote for a woman.”
The researchers were surprised at the consistency of their findings across most demographic groups. Virtually equal percentages of male and female respondents were upset by the prospect of a female president, and nearly equal percentages also were found among respondents with different levels of education.
“You would think educated people, younger people and females would be less upset about the prospect of a female president,” Streb said. “That doesn’t appear to be the case.”
Streb and Burrell suspect social desirability might also be at play in polls regarding the race of presidential candidates, but they did not address that issue in the study. Many political observers believe U.S. Sen. Barack Obama (D-Ill.) could become the nation’s first African American president.
“We can’t quantify it,” Streb said, “but social desirability is almost certainly going to be a factor in horserace polls regarding Obama.”