Presidential Polls 2012: Even if There is a Liberal Bias in the Election, Obama is Still Up By 3

With the election 36 days away, a new set of polls shows Obama leading Romney by 2% to 3% in the general election. Republicans have admitted a tough couple of weeks for the Romney campaign, but some are turning attention away from the candidates themselves and towards the polling data, claiming the polls have a liberal bias, with even Fox News pollsters joining the liberal media conspiracy. With polls having the potential to demoralize Romney supporters, are Republicans right to criticize?

“Most pollsters are using 2008 party preferences to weight their 2012 survey samples, reflecting a much larger Democratic preference than is now really the case,” claims Dick Morris, political commentator and Fox News contributor. “Pollsters are weighting their data on the assumption that the 2012 electorate will turn out in the same proportion as the 2008 voters did. But polling indicates a distinct lack of enthusiasm for the president among his core constituency.”

The 2008 election was indeed exceptional. A record number of minority voters participated in the general election, with nearly one-in-four votes cast by non-whites, according to the Pew Research Center.

"The 2008 presidential election saw a significant increase in voter turnout among young people, blacks and Hispanics," said Thom File, a Census Bureau voting analyst.

This record turnout helped Obama win in 2008, as the groups with the greatest increase in voter turnout also voted for Obama.

The latest data indicates that Obama continues to poll strongly with young people, minorities, and women. According to the latest Fox News poll, if the election were held today, 51% of women would vote for Obama compared to 40% for Romney. Likewise, 58% of individuals under the age of 35 would vote for the president, compared to 35% for Romney.

But are these estimates oversampling Democratic constituencies?

The mainstream method for collecting polling data involves randomly selecting telephone numbers to ask a set of question about voter preferences. This does not advertently favor any party preference, though it does pose a set of challenges.

To begin with, the percentage of people responding to telephone polls is decreasing, with fewer than 10% responding today compared to 36% in 1997. Part of this may have to do with the fact that fewer Americans have landlines today than in the past. Those without landlines tend to be younger and more Democratic, and many polls account for their underrepresentation in polling data by reweighting results to account for age based upon Census Bureau data. This reweighting is notably based upon census data, not upon 2008 voter turnout, as Morris claims.

Another potential pitfall that could affect polls’ accuracy is whether or not calls are made only to those on registered voter lists. More registered voters identify as Democrat rather than as Republican.

“Pollster Scott Rasmussen, who weights his robocall results by party identification, adjusted monthly, has shown a much closer race than most pollsters who leave party identification numbers unweighted,” wrote Michael Barone for the National Review.

But even Rasmussen has Obama ahead with 50% of the vote compared to Romney’s 47% in the latest poll.

 

“On the whole, it is reasonably impressive how unbiased the polls have been [historically],” wrote Nate Silver in his extensive analysis of historical polling bias for the New York Times. “In both presidential and Senate races, the bias has been less than a full percentage point over the long run, and it has run in opposite directions.”

While current polls do project more Obama supporters than Romney supporters on November 6, it is unlikely they have any intentional bias in aggregate. Organizations like Real Clear Politics attempt to ease the burden placed upon voters to decipher the meaning of contradicting polling data by averaging reputable poll results, and their conclusions indicate the race is much closer than some Republicans fear.

Instead of pointing fingers at the pollsters, attention should return to the major personal qualities and policy issues differentiating the two major candidates in voters’ minds. It is likely the reason for Romney’s poor performance in recent polls lays there.