Skip to Content

Support MinnPost

Minnesota Public Radio abandons political polling

For the first time since at least 1996, MPR is out of the political polling business.

"No polls from us," says Chris Worthington, MPR's managing director of regional news. "Other obligations."

Worthington didn't further spell out his rationale, but MPR's exit comes two years after their sponsorship of a University of Minnesota Humphrey School of Public Aaffairs poll erupted in controversy. The U's survey, conducted 7-10 days before the election, had Mark Dayton up 12 percentage points — rising from a month earlier, while other polls showed a tightening. Republicans screamed "bias"; Dayton wound up winning the contest by 0.4 points.

MPR later published the U's internal and Gallup's external reviews — accountable, yet embarrassing, for the "no rant, no slant" network. At the time, Worthington said, "We look forward to more conversation about the recommendations."

Cost may also be a factor; MPR's original polling partners, the St. Paul Pioneer Press and KARE11, stopped sponsoring polls years ago. These days, only the Star Tribune and KSTP-TV sponsor full polls, though non-affiliated pollsters such as Public Policy Polling and Rasmussen Reports also play in Minnesota.

The bottom line is that there are plenty of numbers these days; especially for taxpayer-supported MPR, the reputational risk isn't worth it. MPR News has improved its digging and depth on Worthington's watch, and they're better off unearthing stories than paying an outside contractor to become the story.

Get MinnPost's top stories in your inbox

Related Tags:

About the Author:

Comments (8)

Polling history

The Star-Tribune's Oct. 24, 2010 poll showed similar results to MPR's: Dayton 41%; Emmer 34; Horner, 13%. That difference is not much better than MPR’s. Will the Star Tribune discontinue the Minnesota Poll? One polling firm that got the Dayton-Emmer race right was SurveyUSA. As Andy Birkey reported in the Minnesota Independent Oct. 29, 2010 that poll showed Dayton up by one point. Interestingly, the Minnesota Poll also has the marriage amendment as a dead heat. On Sept. 11, SurveyUSA had it passing by seven points. Which pollster will be proven accurate Nov. 6?

A little more context needed here

John, if you actually clink through to my Emmer story below this post, you'll see the Strib poll was taken before the MPR poll, and had Emmer 7 points down -- not far off Emmer's own 5-points-down a few days earlier.

The Strib's Minnesota Poll and MPR's Humphrey Institute poll differed in three significant ways:

1. The Strib poll's gubernatorial margin was declining, while MPR/HHH was rising.

2. Emmer's poll and the Strib poll never diverged by more than a couple of points, and in one case, the Strib actually had the race closer than Emmer's poll. MPR's last two polls had gaps of 5 and 12 points.

3. The final MPR/HHH poll was taken days later than the Strib poll, when even Emmer's poll showed the race tightening dramatically. The Strib wimped out of a final week poll, because (they said) they didn't want to influence the race close to the election. Like it or not, they didn't have a last-minute snapshot that could be cross-checked against Election Day. I know the bias crowd will spin that as part of the conspiracy, but given their hatred of the Strib, maybe they should say thanks.

By the way, Public Policy Polling was also very close on the final - +3 Dayton. I wouldn't get too strung out over a couple of points difference, since no one does a final day or Election Day poll and late-deciders can't really be captured. 

And one other important thing ...

... the Strib changed pollsters this year, from Princeton Survey Research Associates to Mason-Dixon, which did the PiPress-KARE-MPR polls before MPR went with the U.

Could you help me out here, David...

You write:

"(The MPR sponsored poll by the U of M) conducted 7-10 days before the election, had Mark Dayton up 12 percentage points — rising from a month earlier, while other polls showed a tightening."

...and that those MPR/U of M polling results "erupted in controversy." I looked at the internal review document you linked to and I see this, which indicates that the MPR's poll results were inline with polling that the MinnPost/St. Cloud State's conducted for that same race in the same month:

"Differences between polls may not be substantively significant as illustrated by the case of MinnPost’s poll with St. Cloud State, which showed Dayton with a 10 point lead, and the MPR/HHH poll, which reported a 12 point lead."

I understand that the final result of in the election was much closer than MPR polling and MinnPost/St. Cloud polling indicated. And I know that you object to the authors identifying MinnPost/St. Cloud poll as "the MinnPost/St.Cloud poll" (you think that the authors should have identified it as "the St. Cloud State poll" or at least as the "St. Cloud State/MinnPost poll.") And I understand your finding that the MPR poll showed Dayton support rising, where other polls showed the race tightening.

But were the authors correct when they stated that the MinnPost/St. Cloud survey (or whatever you want to call it) reported Dayton having a 10 point lead over Emmer in their final poll before the election? Is that what happened?

St. Cloud poll

Yes. You can see the poll results here:

http://www.realclearpolitics.com/epolls/2010/governor/mn/minnesota_governor_emmer_vs_dayton_vs_horner-1393.html#polls

The SCSU poll is an odd duck. It includes a lot of questions, and a very lengthy 11-day survey time. (They've been criticized for that; this year they cut it to a week.)

The methodological problem is it means very old results are included in what can be a fast-moving target. If you look at the link above, the poll includes results from Oct. 10, three weeks before Election Day. That's far earlier than the other final pre-election polls, including the U's. So it's a bit of apples-to-oranges, but something the U clung to in a fairly weak response.

Also, SCSU publishes only one pre-election survey, so there's no way to map the arc of their polls in a single campaign.

Thanks, David.

I myself wouldn't characterize the U as "clinging" to the St.Cloud/MinnPost "Dayton has a 10 point lead in October" result. I think that their decision to report that result (in support of their own poll methodology) was relevant.

And I don't think that their response was "fairly weak," as you say--given the fact of other Minnesota pollster findings in the middle of October. But these characterizations are matters of opinion.

You are right that polling in October is often an attempt to report on "a fast moving target." But Minnesota resources for conducting week-to-week October polling on Minnesota races don't seem to be there. The MRP/U poll didn't purport to be a picture of the race in late October (as did the SurveyUSA poll.)

So I don't see why MPR/U was singled out for "eruption of controversy" about their polling methods, given that their picture in the middle of October was very similar to the picture presented by two other Minnesota polls (The Star Tribune's poll and the poll from St. Cloud State/MinnPost.)

Thanks again for the info and the link.

Good news

I think they made the right decision. Media sponsored polling always struck me as unethical in journalistic terms. It's a form of paying for the news.

Here come the ads

Now that we may be a swing state, get ready for the volume of garbage ads.
I hope Amy's strong position will end up carrying in some votes for Obama and other Democrats