Democratic U.S. Rep. Jim Oberstar's campaign is questioning results of a poll released this week by his Republican challenger Chip Cravaack. Oberstar has represented Minnesota's 8th Congressional District for 17 terms, and typically wins re-election handily.
But the new poll, paid for by Cravaack's campaign, shows the challenger running neck and neck with Oberstar.
This is Cravaack's first run for office. His poll was conducted by Public Opinion Strategies for the Cravaack campaign. Poll results are based on phone interviews Sept. 28-29, of 300 likely voters in the district.
The poll shows Oberstar leading Cravaack by just three percentage points, 45-42 percent, which is within the poll's margin of error.
But the Oberstar camp is crying foul. Campaign manager Blake Chaffee calls the survey a push poll -- a poll with wording designed to prompt answers favorable to the sponsor.
"Over the weekend we were contacted by a number of constituents that were respondents on the poll, and they expressed some concern to us about the calls -- basically that the calls were not aimed at getting responses," said Chaffee. "They were aimed at shaping the respondent's opinion with outright falsehoods."
For example, one of the questions had to do with health care reform legislation and abortion, Chaffee said.
The poll is legitimate, said Neil Newhouse, a spokesman for Public Opinion Strategies. Newhouse said the results are based on an initial series of questions with neutral wording. After that came what he calls message questions, which challenge Oberstar's record and test which messages resonate with voters.
He said the poll results are not based on the message questions.
"We certainly ask a couple of issue questions about the congressman based on his record, but we certainly did not ask any questions before the ballot test that would bias in any way the numbers that I've given you," said Newhouse.
Longtime pollster Rob Daves describes a push poll as unethical telemarketing. A legitimate sample survey with a legitimate tally is not a push poll. But he says testing messages can be done without biasing the results.
"Surveys that test messages often present positive and negative messages about candidates and see how respondents react to them, so that campaigns can take these messages and craft strategies around them," said Daves. "Those are legitimate surveys, if done in a methodologically acceptable way."
Public Opinion Strategies' Neil Newhouse paraphrased several of the survey's non-message questions for us. One asks, "Is it more important to elect a Republican who would be a check and balance to President Obama's programs and policies, or to elect a Democrat who would support President's Obama's programs and policies?"
The pollster then asks several more questions before asking the respondent if they prefer Cravaack or Oberstar. Rob Daves says questions like that can influence the polls measure of support for candidates.
"The questions that come ahead of the support questions for the two candidates are likely to influence those questions," he said.
Daves says he can't judge the poll's accuracy without seeing the full set of questions and the methodology. But he says you have to consider who sponsored the poll, and says a good poll releases not only the results, but the questions as well.
"So that people, readers, viewers, citizens, voters, can make an informed decision about its veracity," said Daves. "Unless they do that, then they're holding something back, and you have to wonder about why they're holding something back."
Pollster Neil Newhouse says Public Opinion Strategies can't release the survey questions without permission from the Cravaack campaign.
Cravaack campaign spokesman Kyler Nerison says the Oberstar campaign wants to know what the Cravaack campaign knows; and strategically, Cravaack's campaign doesn't want them to.
For his part, Cravaack says the poll results reflect what he's hearing on the campaign trail.
"The men and women of the 8th district, they say they've had enough," said Cravaack. "They truly believed in Jim Oberstar at one time, but they don't believe in him anymore. I see a lot of Democrats are saying, 'I've always voted for Jim, but no more."
But history may be on Oberstar's side. He's won his last five elections with an average of 67 percent of the vote.