According to some reports Thursday morning, the Oregon’s governor race was suddenly up for grabs. A new poll, as it was reported anyway, showed voters making a huge swing from the incumbent Democrat to his Republican opponent as scandals engulfed the governor’s fiancee. But there was a big problem: The poll didn’t ask a straight-up head-to-head question. But it was treated as such, anyway, sparking a momentary freakout among political watchers.
“A dramatic shift in poll numbers has taken place since Governor John Kitzhaber’s campaign became riddled with scandal,” declared Portland radio station KXL. “A poll commissioned by KATU has Dennis Richardson with a surprising double-digit lead,”
The radio station’s report was tweeted out by Political Wire:
Wow, the scandal involving Oregon’s first lady has shaken up the OR-Gov race. Kitzhaber now trailing badly.
http://t.co/KpaqpXt0dV
— Taegan Goddard (@politicalwire) October 23, 2014
But this wasn’t a straight-up horse race poll. It was delving into a subset of the respondents to find out how much of an impact the recent scandal was having.
Democratic Gov. John Kitzhaber’s fiancee, Cylvia Hayes, has been the center of controversy in recent weeks, with revelations that she had married an Ethiopian migrant in the 1990s for money and worked at a pot farm around the same time.
So the new Survey USA poll asked about it and the results didn’t look good for Kitzhaber, who has held a consistent and sizable lead over GOP challenger Dennis Richardson. But the pollster had not asked a traditional “Would you vote for…” question.
But that didn’t stop some of the coverage from comparing this poll to other recent horse race polls.
The poll started by asking voters whether they were following the governor’s race and then whether they were following the Hayes scandal. That narrowed the field from 950 contacts to 407 voters. Then it asked them whether the news would change their vote in favor of or against the governor.
That’s where the situation turned worrisome for Kitzhaber. Just 38 percent of those surveyed said they would still vote for Kitzhaber (and another 1 percent said they were switching from Richardson to Kitzhaber). But 18 percent said they would switch from Kitzhaber to Richardson.
Those 18 percent, paired with the 37 percent who said they would continue backing Richardson, would give the challenger a 55 percent to 39 percent lead over the incumbent — at least among this population.
But that’s the limit of the poll. Not only was the wording unusual, at least compared to the typical “Would you vote for…” query, but it was a much smaller sample size. And the poll’s authors acknowledged that.
“The poll’s essential question was asked just of the 407 qualified respondents,” they wrote. “This may or may not directly overlap the universe of Oregon’s likely voters in 2014.”
But that didn’t stop KXL — and some of the national political observers who picked up the poll before delving into its methodology — from portraying the survey as a possible game-changer.
But a review of the poll itself lead to a quick reevaluation of its merits.
In retrospect, this is a pretty dumb way to conduct a poll
http://t.co/KpaqpXt0dV
— Taegan Goddard (@politicalwire) October 23, 2014
