Wednesday, April 23, 2014
By Jason Singer firstname.lastname@example.org
Assistant City Editor / Online
PORTLAND — The first poll of the 2011 mayor’s race showed former state senator Michael Brennan leading his 14 competitors Tuesday, but other candidates and a polling expert quickly criticized the poll for its methodology, specifically for using automated phone calls and only mentioning nine candidates to its respondents.
According to the poll, Brennan would have nabbed 27.4 percent of first-place votes if the race ended Oct. 30, ahead of former state senator Ethan Strimling, who would have received 21.6 percent of the first-place votes and finished second.
But the poll, conducted by Maine People’s Resource Center, left out six mayoral candidates. Charles Bragdon, Peter Bryant, Hamza Haadoow, Jodie Lapchick, Richard Dodge and Chris Vail were lumped into an “other/none” category, and never mentioned by name to respondents.
David Loughran, campaign manager for City Councilor Nick Mavodones, one of the candidates, blasted the poll’s methodology. He said the calls listed the candidates in alphabetical order, which would make respondents more likely to pick Brennan.
Loughran also said the poll’s demographics don’t seem to mesh with typical Portland off-year elections, and the results don’t reflect Mavodones’ experiences campaigning around the city.
“There are plenty of reasons this doesn’t accurately reflect the electorate,” Loughran said. “We’ve identified thousands and thousands of supporters for Nick. It isn’t consistent with what we’ve been seeing and hearing.”
Mavodones presently serves as mayor but was chosen by other members of the City Council. On Tuesday, Portland voters will elect the mayor at the polls for the first time in 88 years.
The MPRC poll, conducted Oct. 28-30 by MPRC employees, surveyed 477 likely voters using automated phone calls, also known as interactive-voice response, or IVR polling.
The poll had a margin of error of 4.4 percent at a 95 percent confidence level. That means that if the poll were repeated 100 times, in 95 cases the results would be within 4.4 percent of those reported.
The calls were conducted at different times of the day to increase diversity, and weighted by gender, age and precinct to reflect Portland’s usual turnout in off-year elections, said Mike Tipping of MPRC.
But Patrick Murphy, a local polling expert and president of Pan Atlantic SMS Group said “bullet polling,” another name for IVR polling, is “widely inconsistent,” and can be inaccurate as often as it is accurate.
Murphy said he respected Tipping’s work, but person-to-person communication is better, albeit more expensive, because real callers can seek out more diversity and can ensure the calls don’t get answered by children or businesses.
Murphy and Bragdon also said MPRC shouldn’t have eliminated six candidates, since no polling had been done to determine which candidates had realistic chances of winning.
According to the poll’s results, Mavodones would have grabbed the third most first-place votes with 12.6 percent, followed by City Councilor David Marshall (7.2 percent), Jed Rathband (6.7 percent) and Markos Miller (5.7 percent).
The MPRC also asked respondents to name their second- and third-place choices. It then used computer software to simulate an instant ranked-choiced voting runoff, like computers will do for the city’s Nov. 8 election.
The software predicted that Brennan would have defeated Strimling in the final round, 61 to 39 percent, after all the other candidates had been eliminated. Brennan said Tuesday the results pleased him, but they wouldn’t make him complacent.
“I’m glad the results show I’m competitive and I’m in the race, but I’m not going to let up or do anything different until next Tuesday,” Brennan said. “I’m going out right now to knock on doors, meet as many voters as possible and broaden our base as much as we can.”
MPRC says it’s a nonpartisan, nonprofit organization and doesn’t take political stances. But it does share employees with the Maine People’s Alliance, which candidates described as a left-leaning organization.
None of MPRC’s board members have contributed to Brennan or Strimling’s campaigns, according to campaign finance reports.
Tipping said MPRC was not paid for the poll, but Down East magazine sponsored it. He said MPRC is funded by individual donations and grants.
The MPRC said it didn’t include certain candidates because of their low fundraising numbers, lack of endorsements and media coverage and a lack of political experience.
Nonetheless, about 11 percent of the respondents selected “other/none” as their first-place choice. Those respondents still had their second- and third-place choices factored into the final tally.
In response to Loughran’s criticism, Tipping also said research is inconclusive about the order in which candidates are listed in IVR polling and the effect that has on the polling results. He said the candidates will be placed in alphabetical order on the ballot, so MPRC wanted to best simulate the voters’ experience on Election Day.
Tipping also said if the six unnamed candidates had more support, it would have shown up in the “other” category.
Strimling, whose team has commissioned Public Policy Polling, one of the largest national polling companies, to conduct polling of the race, said PPP has found “very different numbers.”
“(Today’s poll) is inconsistent with what we’re seeing,” Strimling said, noting PPP has made thousands of calls, a much larger sampling size. “We see this as a three-way race and certainly don’t see anyone winning by 20 points.
“We’ve had a great week. We’re certainly excited about from momentum from last week’s finance report to (former state representative) John Eder’s endorsement Monday. Even within this poll, we’re clearly within striking distance of winning this race and we feel very good about our chances.”
MPRC noted in a poll memo that the polling took place before Eder, one of the 15 mayoral candidates, endorsed Strimling.
Murphy and numerous candidates also expressed shock that Mavodones had such few first-place votes, which they said hurt the credibility of the poll, given Mavodones’ track record of success in city elections.
IVR polling, however, is often used by SurveyUSA, PPP and Rasmussen Reports, three of the country’s largest public-polling companies. PPP has received particular praise in recent years for its accuracy.