Can a woman or a black man win?

As a pollster, I’ve been asked repeatedly: Can an African American or a woman really win the presidency?

Allowing my partisan hopes to outrun my certain knowledge, I have generally answered in the affirmative, almost reflexively.

But is it really true? Our real-world experience, frankly, is less than reassuring. The private sector provides ample evidence that discrimination against blacks and women still exists in education, hiring, advancement and pay. More to the point, few women or blacks hold high public office. More than half the country is female, yet women make up just about 17% of House members, senators and governors.

African Americans represent about 11% of the electorate and occupy about that percentage of seats in the House, but most represent majority black districts. There is only one black governor and a single African American senator -- and he is trying to flee for the White House!

Those statistics are troubling but not necessarily conclusive. A range of barriers -- from the entrenched incumbency of white men to lesser fundraising bases for women and blacks -- could account for these dreary facts. Indeed, when women run in general elections for Congress, they win at least as often as similarly situated men.

In addition, public opinion data indicate Americans at least understand that it is socially unacceptable to voice negative sentiments about blacks or women in the White House. Just 5% told Gallup pollsters that they would not vote for a black, and 11% said they would withhold support from a woman.

Such was not always the case. In 1958, 53% admitted to Gallup that they would be unwilling to support an African American and 41% would refuse to back a woman. Even today, voters appear comfortable confessing certain prejudices -- 24% claimed they would not vote for a Mormon, for instance; 42% would not vote for a 72-year-old, and 53% would oppose an atheist.

Although these poll results seem to reflect a fundamental change in attitudes toward women and blacks, we can’t be sure they herald altered voting behavior. Pointing to the New Hampshire polling fiasco -- in which the polls uniformly failed to predict the correct winner -- many assert that surveys cannot accurately measure the effect of race. In truth, though, Barack Obama got exactly the vote in New Hampshire the polls predicted; what cries out for explanation is the substantial under-prediction of Hillary Clinton’s vote.

Nonetheless, there is clear evidence of gaps between voters’ stated attitudes and their actual behavior across many realms, and that distance can be even wider in matters of race and gender. For example, in a 2000 poll, 84% of white Southerners said they opposed bans on interracial marriage. Yet when the issue was put before Alabamians in a statewide referendum that year, half of whites voted to keep the ban.

Studies have repeatedly caught voters applying gender and racial stereotypes to candidates, with women being seen, for example, as more compassionate; men as tougher and blacks as more liberal. However, these experiments have all been conducted with hypothetical candidates in imaginary elections. Surveys in actual elections reveal that voters often do not apply these stereotypes to real candidates. In real life, many female candidates are seen as better able to deal with issues such as defense and crime, while some African Americans are viewed as more conservative than white opponents.

In the end, people tend to vote for individuals (and parties), not for a race or a gender. Shortly before John F. Kennedy’s election, 24% of American voters reported that they would be unwilling to vote for a Catholic. Yet many of those voters ended up casting ballots for JFK because his Catholicism was just one, less weighty, aspect of who he was and what the contest was about.

This year, the individual running under the Republican banner will be at a serious disadvantage on all the structural fundamentals that, more than anything else, determine the outcome of presidential elections. The desire for a change of parties after eight years of Republican rule, an exceptionally unpopular incumbent, the most unpopular war in polling history and an economy sliding into recession all presage a Democratic victory in November, regardless of the race or gender of the nominee.

Some will want to celebrate that milestone as the end of racism or sexism. In fact, it will be just an important marker on what will remain a long, arduous road to equality.

Mark Mellman is a Democratic pollster whose clients have included Sens. Harry Reid, Al Gore and Barbara Boxer.