WASHINGTON -- Most election forecasters, based on the generic polls, are predicting an anti-Republican wave in November that will produce big Democratic gains in the House and Senate.
But summer polls, like conventional wisdom, can sometimes be inaccurate -- or at least exaggerated -- often because of the way questions are phrased to get a desired result.
For example: A mainstay of election-year surveys is the right-track/wrong-track poll to measure how people feel about the overall direction of the country. This question has routinely elicited an overwhelmingly wrong-track response in the 60 percent to 70 percent range.
An Associated Press-Ipsos poll of 1,001 Americans, that included 813 registered voters, reported last week that 71 percent believed the country was moving in the wrong direction. Only 26 percent thought otherwise. A Zogby poll, using about the same wording, showed voters saying wrong track by 59 percent to 34 percent.
Such numbers are fed into the analytical hard drive of the top election forecasters in the country, producing a deeply pessimistic, anti-incumbent spreadsheet on their computer screens.
Veteran political forecaster Charlie Cook writes that when this month's primary upsets, in which voters dumped three incumbents (two Democrats and one Republican), are "combined with Congress' abysmal job-approval ratings and extremely high 'wrong track' numbers, (they) indicate a very volatile, turbulent election year, the kind that incumbents hate for good reason."
But CNN discovered earlier this month that when pollsters ask the voters essentially the same question about the mood of the country, only with different wording, they got a dramatically different result.
In a survey of 1,047 Americans conducted by Opinion Research Corp. of Princeton, N.J., on Aug. 2 and Aug. 3, CNN asked people, "How well are things going in the country today?"
A combined 55 percent said things were going "fairly well" (47 percent) or "very well" (8 percent), compared with those who said "pretty badly" (29 percent) or "very badly" (15 percent).
CNN polling director Keating Holland notes the question's wording was fundamentally different from right-track/wrong-track language, which he says he did not use but declined to say why. The question he prefers yields "a measurement of how well Americans think things are going in the country today."
"Different questions get different answers," he says. CNN has used this phrasing before because, presumably, it produces a much more accurate reading of the pulse of the country. Virtually all other polls use the wrong-track question, whose pejorative wording yields more negative results.
There are enough anomalies and contradictions in the polls to suggest the forecasts of steep GOP election losses could be a tad premature.
For one thing, polls tell us that whatever the voters think of Republicans, they aren't that crazy about Democrats, either. Moreover, some polls have found that a large percentage of Democratic voters say they could change their minds.
"Only 41 percent of Americans believe that Democratic leaders in Congress 'would move the country in the right direction,'" writes Mark Preston, CNN political editor, on the cable news network's Web site. That is less "than the 43 percent of Americans who believe Republican leaders in Congress 'would move the county in the right direction.'" What this means, Preston explains, is that "Democrats need to do a better job of convincing voters they are better equipped than Republicans to lead the Congress."
When the AP-Ipsos poll found that 51 percent said they would vote for the unnamed Democratic candidate in the congressional election, another far less reported question was asked that cast that 51 percent in a different light: "Will you definitely vote for that candidate, probably vote for that candidate or do you think you could change your mind before the election?" A sizable 18 percent of Democrats said they could well "change their mind," hardly a strong vote of confidence for the opposing party.
Then there are the independents, the nation's fastest-growing bloc of voters who will likely decide this election. Surprisingly, most of them still do not know how they'll vote in the congressional races.
Zogby found an astonishing 41 percent were undecided, compared to 32 percent who said they'd vote for the Democrats and 20 percent who favored the Republicans.
A recent Newsweek poll gives us one more reason to take the generic vote with a grain of salt: It found that "68 percent of registered voters say they have only given the November elections 'a little' or no attention."
With more than two months to go, there is still enough fluidity and doubt in these numbers to suggest this election is far from settled.