Just days after a poll showed Amanda Janoo with a slight lead over Gov. Phil Scott, a new Vermont Public survey released on Monday suggests the Republican governor is in fact leading his Democratic/Progressive opponent by 10 points.
The starkly different results just five weeks before Election Day add uncertainty around where things stand in the race for the state’s top job.
Taken together, however, the polls show that Scott, who is often cited as one of the most popular governors in the nation, finds himself in his tightest campaign since he was first elected in 2016.
The Green Mountain State Poll conducted by the University of New Hampshire Survey Center found that 49 percent of likely general election voters backed Janoo, compared with 43 for Scott. That poll had 6 percent undecided, and 2 percent favored a different candidate. The survey, which was released on Thursday, had a margin of error of 3.4 percent.
The new poll conducted for Vermont Public showed a different state of the race, with Scott up by 10 points. It found 44 percent of likely voters favored Scott and 34 percent preferred Janoo.
The new poll had a larger number of undecideds, at 16 percent — one possible explanation for the very different results. It was conducted by private New Jersey data collection firm Braun Research under the direction of Vermont State University political science professor Richard Clark. It also claimed a margin of error of 3.4 percent.
Bert Johnson, a political science professor at Middlebury College, told Seven Days that having two divergent poll results does not mean either was necessarily flawed.
“There are ways in which polls can differ, even though they are honestly conducted,” Johnson said. “I have no doubt that both of these polls are doing their best to assess the status of the Vermont electorate.”
Following the latest poll’s release, Janoo’s campaign noted that one in five voters is still undecided.
“The polls vary, but they agree on one thing: this race will be decided in the final weeks,” the release stated.
Scott’s campaign, which called the UNH poll results inaccurate and misleading, said much the same about the Vermont Public survey.
“Public polling has always undervalued the governor’s support and this poll, while much better than the UNH survey, continues to do the same,” the campaign wrote.
The two polls were of similar size. The University of New Hampshire poll asked questions of 835 people between September 17 and 21, while the Vermont Public poll based its findings on 817 calls between September 10 and September 21, with 732 respondents considered likely voters.
The methodology of the two polls was very different, however. The UNH poll asked questions of a panel that had agreed in advance to participate, while the Vermont Public survey randomly reached people via text, cellphone and landline.
Clark said he intentionally included landline users (7 percent of the total) because they represent an older demographic with higher numbers in the northern and southern parts of the state.
“They also tend to be more likely conservative than cellphone users,” he said.
Public opinion could have changed during the longer time frame for the Vermont Public poll, though Johnson said he couldn’t think of a news event that might account for such a significant swing.
Instead, the University of New Hampshire poll may have done a better job of determining who is a “likely” voter, which can be difficult during low-turnout midterm elections.
“It’s possible that the UNH poll is picking up more on that Democratic enthusiasm in their likely voter calculation,” Johnson said.
The Vermont Public poll’s results for the lieutenant governor’s race also showed significantly stronger support for the incumbent, Republican John Rodgers, though he still trailed his Democratic rival, Molly Gray.
The University of New Hampshire poll showed Gray leading Rodgers by 19 points, or a margin of 54-35 percent, with 11 percent undecided.
The new poll still showed Gray ahead but by just 8 points. It had 43 percent of likely voters preferring Gray to Rodgers’ 35 percent, with 19 percent undecided.
The pollsters acknowledged an error in how they conducted this question, however. They noted that about half of the respondents, 363 of them, were asked about Gray in a question that referred to her as a “Democrat/Progressive.” She is actually running as just a Democrat, which was properly noted in the questions asked to the other 454 respondents.
Gray’s support dropped from 46 percent to 36 when the “Democrat/Progressive” label was used. Since she is not actually a Progressive, her support is likely even higher than the 8 percent, Johnson noted.