I regret having to say that I’m simply not buying the finding of our own poll of the governor’s race. It has Mike Beebe ahead of Asa Hutchinson by, well, 21 points.
This is the telephone survey conducted last week by the veteran PhDs, Ernie and Zoe Oakleaf of Opinion Research Associates of Little Rock. It was done for Stephens Media’s Arkansas News Bureau, where I work.
You’ll find the poll reported today at arkansasnews.com and in some of the state’s finer newspapers.
Let me be clear about what I’m saying. These are reputable, competent pollsters of a personal and professional integrity that I’d never even remotely doubt. This overall poll, which covers a lot more than the governor’s race, is exhaustive, insightful and credible, as readers will learn as additional findings get reported later in the week. But the governor’s race result — Beebe 52, Hutchinson 31 — simply strikes me as a wild aberration. The vast margin, I mean.
The poll’s demographics make sense. The other findings, many of which will be reported later in the week, are logical.
They reveal that the random sample of 500 registered voters was conservative and moderate and overwhelmingly opposed to gay foster parents. Respondents had predictable, conventional views on illegal immigration, George W. Bush and Iraq.
But when asked to choose between “Mike Beebe, the Democrat; Asa Hutchinson, the Republican; Jim Lendall, the Green Party candidate; and Rod Bryan, the independent,” with the order switched from call to call as is customary, these respondents said something far different from anything we’ve seen.
What we’ve seen are polls showing Beebe ahead from three to 10 points.
Just Monday, Rasmussen Reports, a national automated poll, released findings showing the race tightening, with Hutchinson having pulled to within four points, 45-41.
The Rasmussen poll was conducted in one day, last Tuesday. Opinion Research polled over five days, Tuesday through Saturday.
Opinion Research’s new numbers are what they are. A professional pollster produced them objectively, admits surprise, stands by them and offers possible explanations.
I should acknowledge that there have been a couple of recurring Republican criticisms over the years of Opinion Research’s methodology, neither of which I put any stock in.
One is that Republicans insist the Oakleafs ought to draw a bigger sample from the 3rd Congressional District, which is Republican and has grown much faster than the other three districts since the Census of 2000.
But Ernie Oakleaf, who consistently relies on samples of 125 persons in each of the four districts, can show mathematically that the difference from such weighting would be minuscule. I’ve seen him take his poll, apply a little population/turnout multiplier for the 3rd District and produce a change inside a percentage point.
Republicans also think samples are to be manipulated. They acknowledged to me a couple of weeks ago that they super-screened the sample for their recent internal poll to produce on purpose a set of demographic breakdowns that they insisted more accurately reflected the electorate. The Oakleafs make random calls, screen for registered voters and otherwise let the demographics and chips fall where they may, which is more pure, more objective, more independent.
Opinion Research produced a sample identifying itself as 38 percent conservative, 41 percent moderate and 13 percent liberal. That sounds about right. The sample identified itself as 41 percent Democratic, 20 percent Republican and otherwise independent or unaffiliated. That’s perhaps a little light on Republicans.
The sample strikes me as a tad heavy on retired people. Maybe younger people only have cell phones anymore. But it’s an established fact that retired people vote in higher percentages than younger working ones.
The sample identified itself as 52 percent female and 48 percent male. And get a load of this: Women favored Beebe by 56-27.
Some pollster is way off. Even the famously eccentric Arkansas electorate is not that volatile.