Rather than surveying likely voters, Gallup polled an unweighted sample of Americans 18 and older — which, as experienced consumers of polling data know quite well, invariably skews results to the left, in comparison to actual voters. (The most recent Rassmussen survey of likely voters shows 54% opposed to ObamaCare.)
Liberals who exult over USA Today‘s “turnaround” spin are only slightly more foolish than conservatives who despair. Even the normally pessimistic Allahpundit of Hot Air resists Gallup-induced gloom:
That’s quite a swing among indies from two days ago to Gallup today. Color me skeptical too that the number of Republicans who support the bill – 14 percent, allegedly – now exceeds the number of Democrats who oppose it. . . . The biggest problem . . . is the nationwide sample: It’s fine as a general barometer of public sentiment, but of course the left will want you to believe that this vindicates their theory O-Care is a political winner in November. Problem is, that doesn’t depend on national polling – it depends on swing districts, and at last check those districts were looking grim.
Repeat after me: A congressional mid-term is not a presidential election.
Political forecasting is always a tricky business, and pollsters have long lamented the tendency of some journalists to hype the latest poll numbers as forecasting future developments. Public opinion is a moving target, a single poll is merely a snapshot of opinion at a given time, and results can be distorted by any number of factors.
I’ll have more on this general subject on The American Spectator main page Wednesday morning.