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.
Poll: Only 4% of U.S. Adults are Newly Insured, Half Choose Obamacare Alternative | Sarah Jean Seman