Why the Obama-Romney Race May Not Be the Dead Heat Polls Describe

US News and World Report ^ | September 4, 2012 | Clark S. Judge

[2ndDivisionVet]

As the Republican convention closed on Thursday night, a new poll pointed to the potential for a major surprise when Americans go to the voting booths in November.

Pollsters will tell you that they cannot always count on voters to level with them about which candidates they favor. For example, Gov. Scott Walker's victory in the Wisconsin recall election several months ago was predicted by the polls, but his margin was not. It was a larger margin than the surveys had forecast. Similarly in 2004, Sen. John Kerry went to bed on election night believing he would be the next president. He woke to a surprise, as did Democrats in 1994 and, to a lesser degree, in 1980. Note, in all these elections it was the Democratic who got the unwelcome surprise.

There are any number of hypotheses as to why. For example, on the 2004 election, one of the most knowledgeable men on political statistics in the United States, Michael Barone, has noted that when the media consortium that sponsored the exit polling delved in depth as to why their results were so far off the actual results, they found a remarkable coincidence. In many of the precincts where their samples diverged most markedly from the actual balloting, the interviewers—those who approached voters and asked for whom they had cast their ballots—were attractive, female graduate students. These exemplars of a heavily Democratic cohort may have telegraphed their personal preferences to respondents (particularly male respondents), who had fed them back the answer they so clearly wanted to hear. Makes you wonder whom the exit pollsters will hire for this year interviewing.

A less, shall we say, "exotic" explanation for the GOP polls-to-voting gap is that in big years when a lot of non-Republicans have planned to cast their ballots for the Republican ticket, a sizeable number of union members haven't wanted to take the chance of telling anyone that they planned to buck their leadership. Another explanation is that the distrust many conservatives harbor toward the mainstream media spills over to pollsters. They become less candid when elections are most fiercely contested.

In any event, whatever the cause of the gap, pollsters have asked, is there a question that gets around the candor deficits? And some believe they have an answer. People may not tell you their candidate preference directly. But in close races where there is no credible consensus about the likely outcome, they typically believe that their candidate will win. In other words, ask them which side is likely to prevail in the next election and their answer will tell you their real opinion.

Which brings us to this year. A quick glance at the presidential poll summary at RealClearPolitics.com shows an amazing confluence of the eight national polls the Web site tracks. As of this weekend, all but two find a 1-point difference between the president and Mitt Romney. The remaining two have a two-point spread. In other words, everyone is finding the national vote tied, that is, falling within the margin of statistical noise that goes hand-in-hand with all survey research.

This is new. Only a month ago, the spread between the most pro-Obama and the most pro-Romney poll on the RealClearPolitics list was 11 points. Both polls advertised a 3-point margin of error. Someone was wrong. Since then all the opinion samples have clustered around what the most pro-Romney polls of a month or so ago were reporting: a neck-and-neck race. But given the traditional GOP gap, is a dead heat the real story?

Maybe not. In a poll released on Friday, Rasmussen Reports found that "60 percent of likely U.S. voters believe the next president will be a Republican." Only a quarter thought it would be a Democrat.

Eighty-nine percent of Republicans said the president after Obama would be from their camp, and 55 percent of independents agreed. Surprisingly 38 percent of Democrats had the same view (versus 47 percent who said it would be a Democrat). The telephone survey was taken during the last two nights of the GOP convention and got answers from 1,000 voters, meaning it was a typical sample.

The Rasmussen team was quick to point out that they did not specify whether the "next president" would be elected in 2012 or 2016. But still, with Michigan, Wisconsin, and Colorado having moved from at least leaning Democrat to "in play" in recent assessments, something may be going on below the surface.

Here is one certainty: The Romney and Ryan presentations in Tampa put on display a pair of principled, compassionate, capable men, prepared to lead in a direction consistent with America's deepest values. Maybe, quietly, a much larger slice of the American people agreed than we have suspected.

November could be interesting.

[MNJohnnie]

You might of stumbled over an old story. You got to be careful and find a date. Once this stuff gets on the net it is out there for a long time.

[tanknetter]

ok, found the story via google. The poll and story were from early August. Not sure why they were linked to from the CNN front page this morning ...

[Pharmboy]

NY, CA, IL, DC, MD for sure...the rest are not so sure for Obummer.

[calex59]

Here’s my reasoning for why the polls are way off. Romney draws bigger crowds than Obama and also raises more funds than Bozo. If they were as close as the polls suggest then their crowds and fund raising would also be close. They are not, therefore Romney is ahead by a good bit and will win the election.