Maine firms pan N.H. poll, Baldacci lead

loading...

AUGUSTA – A controversial poll conducted by a New Hampshire research firm has concluded that Democratic gubernatorial candidate John Baldacci holds a huge 41-point lead over his closest rival and that “the race is [his] to lose.”

But in unusually harsh criticism Tuesday, three of Maine’s major pollsters challenged the veracity of the survey’s findings and dismissed its methodology as “flawed,” “unreliable” and “unsound.”

Rather than casting Baldacci as a winner, Bowdoin College Professor Christopher Potholm, who runs the Command Research polling firm, said the poll’s attempt to assert that Baldacci is supported by 62.5 percent of all Maine voters could ultimately backfire and the Bangor native could be perceived as “a real loser.”

“Baldacci hasn’t got that kind of level and, in fairness to him, he’s not saying that he does,” said Potholm. “But what happens in the next go-around when he only has 55 percent or 52 percent or less? People are going to be saying that John Baldacci is dropping like a stone when, in fact, he may not have moved at all.”

The poll at the center of the dispute was undertaken by RKM Research and Communications of Portsmouth, N.H., on behalf of WCSH News Center 6 of Portland, WLBZ News Center 2 of Bangor and the Bangor Daily News. Ironically, polls conducted for the media, as opposed to a political candidate, generally carry credibility in the public survey industry.

R. Kelly Meyers, director of research at RKM, said the poll was based on a telephone survey of “likely voters” in Maine. Those voters were defined as someone who is at least 18 years old, registered to vote in Maine and described as “definitely” or “probably” voting in the general election.

The survey was based on a random, computer-assisted sampling of 410 voters across the state conducted between May 6 and May 8. It incorporated a margin of error of 4.8 percent.

Among its primary findings, the RKM poll concluded that neither of the two gubernatorial Republican primary candidates Peter Cianchette of South Portland or Jim Libby of Buxton “appeared poised to pull ahead of Baldacci.” Both candidates finished second in competing scenarios against the Democrat with Cianchette picking up 21.5 percent of the vote and Libby, 12.4 percent.

More troubling to the Maine pollsters than Meyers’ results were the methodologies he employed to reach his conclusions. Those criticisms cut across party lines. Potholm, a Republican, was joined by Patrick Murphy, a Democrat, who runs Strategic Marketing Services, of Portland, and also by Mary Ellen Fitzgerald, an independent who is president of Critical Insights of Portland.

The pollsters deplored Meyers’ reliance on survey respondents described as “likely” voters on the basis that they would “probably” be voting in the general election. Potholm said that without more probative inquiries to determine whether the respondents had voted in the most recently held presidential or statewide elections, it would be difficult to conclude that someone who said they would “probably” vote in the fall would actually vote.

“People rarely say ‘of course not, I never vote'” Potholm said. “If someone tells us that they’re ‘probably’ going to vote, we don’t even count those. In our experience, that’s just a nice way of saying ‘I’m not going to vote.’ Strange and mysterious, this poll doesn’t ring right to me. Nonvoters placed with voters tend to skew the results dramatically in favor of the incumbents.”

“I’m sort of speechless,” Fitzgerald said in reaction to the RKM poll. “Really these numbers are pretty surprising to me. I do think there’s a real problem with the way this was done because you are getting presentation bias. We stay away from ‘probably going to vote’ questions.”

Meyers denied the suggestion he had employed “presentation bias” to pump up the number of “likely” voters. In fact, he insisted the contrary was true because his interviewers prefaced their questions by offering a number of reasons why people don’t vote before inquiring whether the respondents would be voting.

“Bias would be, ‘As you know, responsible people turn out to vote on election day, what about you?'” he said. “That’s the opposite of what we do. We acknowledge that many people don’t vote, tell them why and then have them tell us how likely it is that they will vote”

Still, Meyers admitted that five or six months before an election makes it difficult to have a “a very tight likely voter model.”

“We want to ensure the minimum amount of information that can qualify them as a likely voter,” he said. “Different organizations use different techniques, but they’re all designed to do the same thing: which is to get around the dilemma of people over-reporting their intent to vote. A tighter model won’t necessarily produce more accurate results this far in advance of an election.”

Other results in the RKM survey that elicited jaw-dropping responses from the Maine pollsters included findings that:

. Baldacci and Cianchette were tied in a statistical dead head among Republicans.

. 89 percent of all those surveyed had already made up their minds on who would get their vote for governor.

. Jonathan Carter, the Green Independent Party candidate, would poll only 1.6 percent of the statewide vote when he has always polled between 5 and 7.5 percent.

. Independent candidate John Jenkins of Lewiston would get 0.5 percent of the vote, compared to independent candidate David Flanagan’s 0.2 percent of the vote.

Another troubling aspect of the RKM poll was its failure to identify all of the gubernatorial candidates to the respondents. The poll asked two questions on that point: “Would you vote for John Baldacci, Peter Cianchette or someone else?” That question was followed by a second that substituted Libby for Cianchette.

“By not reading all the candidates’ names, they just get farther and farther away from anything that could possibly be used by sensible people,” said Potholm, whose son works on the Cianchette campaign. “It’s a methodological error, piled on top of a methodological error.”

But Meyers defended his decision not to use other candidates’ names in the question and countered that every pollster makes a determination on those considered to be viable candidates and those perceived as less viable.

“There is no rule or right or wrong standard for that decision,” he said.

“Well, I wouldn’t bet 2 cents on a telephone survey that did not read the actual ballot to the voter,” Potholm said.

Murphy was particularly astonished at RKM’s low number of undecided voters, which at 11 percent, is about half to two-thirds less than what is normally expected a month before the June primary.

“I would have to say that at this point, there’s no way that only 11 percent of the voters are undecided,” he said. “Taking everything into account, that number would have to be more like 25 to 30 percent.”

The RKM poll also drew mixed reactions among the candidates and their representatives. Donna Gormley, spokeswoman for Baldacci, said she wasn’t surprised by her boss’s popularity. At the Cianchette camp, Roy Lenardson characterized Carter’s projected 1.6 percent of the vote as “utterly ridiculous.”

The poll also included projections for the U.S. Senate race that pits incumbent GOP Sen. Susan M. Collins against Democratic hopeful Chellie Pingree. Collins was the predicted winner with 63 percent of the vote, compared to Pingree’s 23 percent. Pingree spokeswoman Jennifer Sargent mused “those numbers just don’t add up for us.”

Conclusions for the 1st and 2nd Congressional District races were dismissed because the samples were too small at around 200 voters in each district.

Some of the pollsters, like Fitzgerald, expected their criticisms of RKM to be brushed aside as sour grapes for failing to win the media contract.

“People are going to say that and I don’t want to cast aspersions on that particular business entity at RKM, but these results fly in the face of everything I’ve ever seen,” she said.