Presidential public opinion polls are often quite evanescent.After the 2008 New Hampshire Democratic primary, which Hillary Clinton won by 3% pursuant to falling behind Senator Barack Obama in the polls a day or two before that primary by 9%, Neil Newhouse of NBC News/Wall Street Journalpolling stated the following:“With Hillary Clinton’s victory last night, any shred of reputation that pollsters have for being accurate barometers of public opinion goes out the window.”
The persuasiveness of that mea culpa lasted about the time it takes for New York cabdrivers to honk after the changing of a traffic light to green, to plagiarize a comparison.
Everyone is quite surprised that Governor Mitt Romney is, as The New York Times wrote yesterday, with President Obama “locked in a tight race, with each gathering 46 percent of the vote.” (CBS Evening News pointed out correctly that such national polls are not the measure to watch, since the president is elected by state electors: typical excellence, parenthetically, by Scott Pelley’s fine newscast).
Of course – more on which later.
And there were the invalid poll conundra: as USA Today reported, two polls released on April 16 — a Gallup poll reporting Romney and Obama in a “statistical tie,” while a CNN poll showed Romney ahead by 9 points – sported findings which are outside the margin of error that pollsters often claim happens to that extent about one of one hundred times (lesser margin-of-error results are claimed to occur about one in twenty times – a significant underestimation).Yes, the sample size was different, and the dates were not precisely coincidental (Gallup’s taken over 5 days with CNN’s taken over 3.But this is quite a discrepancy.
How could sophisticated journalists – the pollsters are not motivated or often asked to reconcile such contradictory findings) — not realize that presidential polls taken during the primaries are almost always , but not always, unstable, and that even intensity, when it is measured, is more unstable before the principals from each party are chosen.
Moreover, here is a bit of news for most: polls indicate that it is unimportant who wins the vice presidential nomination.
The choice of Sarah Palin weighed consciously or unconsciously on voters’ minds in 2008 and, in this pundit’s opinion, lost the race for Senator John McCain.
Did the polls show it?No.But everyone who paid attention knows it to be plausible.
Don’t let the polls convince you that a race is decided; they are not random and usually not off by more that 5 points each way, but they are simply not, as Mr. Newhouse once conceded, “accurate barometers of public opinion.”
Prof. Vatz teaches the rhetoric of public opinion in his Persuasion class at Towson University; he is also the author of The Only Authentic Book of Persuasion (Kendall Hunt, 2012)