Eye on Boise: Behind the numbers, what the polls tell us about Idaho’s race for governor

Two major statewide polls in the Idaho governor’s race provide some interesting insights into where the race stands, but they may not be the same conclusions drawn by the group that commissioned the polls.

A Democratic political action committee called Idaho Voices for Change Now commissioned two polls, one in July and one in August, and plans at least one more in the fall. The PAC just formed and filed with the state on July 17; under Idaho’s campaign finance laws, it won’t be required to disclose its funders until Oct. 10.

“We were frustrated by the lack of attention put on this race and wanted to gauge how the race was shaping up,” said Tim Lim, a Washington, D.C., consultant and spokesman for the Idaho PAC. Lim, who said he’s “just a hired hand,” said the PAC consists of “a group of Idaho voters and activists who really care about this election and want to make a difference. They see this as a really unique opportunity.”

They’ve not yet been identified, as Lim said they’re “wary of attention or reprisals in a state that’s incredibly Republican. So for now they’d like to be in the background and just really let the numbers and the content that we produce speak for itself.”

He did say the PAC is neither tied to nor funded by the national Democratic Party. “We’re an independent state effort; we’re not connected to any national groups,” he said.

The two Idaho polls were conducted July 12-15 and Aug. 2-5; the sample sizes were 1,061 likely general election Idaho voters in the first poll, and 826 in the second one. The margin of error is plus or minus 2.77 percent for the first poll, due to the large sample size; and 3.19 percent for the second poll.

In July, the poll found 38 percent of respondents said they’d vote for Republican Brad Little if the election were held today, while 28 percent chose Democrat Paulette Jordan. Seven percent favored a third-party candidate, and 26 percent were “unsure.”

In August, the poll found Little’s support down two points at 36 percent, with Jordan’s holding at 28 percent. Third-party candidates had dropped to 5 percent, while “unsure” was up to 31 percent.

“Since last month, Lt. Gov. Brad Little’s lead is down to single digits,” Lim wrote in an Idaho Voices for Change Now news release about the August poll.

But Boise State University political scientist Jeffrey Lyons, whose expertise is in public opinion research and who is the new director of survey research for the Idaho Policy Institute, said those numbers don’t indicate a tightening race.

“To me, the story is it’s very similar over time – the differences are within the margins of error,” Lyons said. “Across both polls, Brad Little has a lead over Paulette Jordan outside the margin of error.

“It’s not a huge lead,” he added. But he noted that the poll also queried voters who said they were “unsure,” asking them if they had to decide now, which way they’d lean. In July, 24 percent of the unsure voters said they’d lean toward Little, 8 percent said Jordan, and 59 percent stayed unsure.

In August, on the same question, 23 percent said they’d lean toward Little, 14 percent said Jordan, and 61 percent stayed unsure.

“So he’s got a lead amongst people who say they know how they would vote, and he’s also got a lead among that unsure category,” Lyons said. “If we distribute those numbers up and into the prior question, he’s looking at a little bit of a cushier lead.

“That’s a judgment call, whether we should take those unsure voters and allocate them, or just look at them as unsure. Either way you slice it, Little appears to have the lead.”

Even so, a sizable number of Idaho voters, according to the poll, remain undecided about the race, “which is not uncommon in a race that is this early on,” Lyons said. “The bulk of the campaigning is still yet to come, the bulk of the money is yet to be spent. … There’s probably still a sizable number of voters who probably just don’t know a ton about either candidate.”

When voters are undecided, Lyons said, “The partisanship does end up mattering so much at the end of the day. So a lot of those unsure voters are probably going to break along partisan lines.”

Little and Jordan emerged from contested primaries in the spring, with Little’s a bitter three-way race in which he defeated current GOP Congressman Raul Labrador and Boise physician and businessman Tommy Ahlquist with 37.3 percent of the vote. Jordan notched a more comfortable margin over A.J. Balukoff, the party’s 2014 gubernatorial nominee, in the Democratic primary, winning with 58.4 percent of the vote.

“There probably is, in that unsure category, some Labrador and maybe some Ahlquist voters who are upset, right?” Lyons said. “Who don’t perhaps like how Brad Little’s campaign attacked their candidate or something like that.”

The Democratic primary race wasn’t as bitter, though many in the party’s establishment, including Jordan’s colleagues in the Idaho Legislature, lined up behind Balukoff, rather than Jordan. “Your average voter probably is not super-aware of that sort of distinction,” Lyons said. “If that’s the case, the hunch is probably that you’re getting more Republicans here who are saying they’re unsure, and will probably come home to Brad Little in November.”

That’s what happened in the 2016 GOP presidential primary election, Lyons noted. “There was a ton of support for other candidates. Most primary voters … were not casting votes for Donald Trump. He was winning pluralities, but not majorities.”

That led to speculation that Trump would suffer with the Republican base in the general election – speculation that proved untrue.

“We saw basically 85 to 90 percent of Republicans circle the wagons and come back to Donald Trump and vote for him,” Lyons said. “And that’s basically the same percentage of Republicans who voted for Mitt Romney and previous candidates.

“Voters still come home because of that partisanship.”

One other factor that can affect the accuracy of a poll, Lyons said, is how the pollster defines a “likely” general election voter.

“That’s really hard to do,” he said. “That, to me, is one of the unsung challenges of polling.”

In 2016, major polls in Michigan and other parts of the Midwest assumed that the general electorate would be similar to those who voted four years earlier, in 2012, Lyons said. But in 2016, African-American voter turnout in those areas dropped sharply from its 2012 levels, when Barack Obama was on the ballot.

“That made it look like Hillary Clinton was doing better,” Lyons said.

The company that conducted the Idaho poll, Clarity Campaign Labs, is a reputable Democratic pollster, Lyons said.

“They do pretty good polling,” he said. “I wouldn’t look at these numbers and think that they have done some kind of scheming to inflate Jordan’s numbers or anything like that.”

Clarity earned a B-minus grade in fivethirtyeight.com’s rating of pollsters, based on 10 polls it’s conducted. It was also rated as showing a 0.9 percent Democratic lean on average, which Lyons said is very small.

“So if everything else looks right, I take the survey seriously, and then kind of say, maybe there’s a slight bump that would favor Jordan here,” he said. “But I’m not going to discredit the whole poll.”

Lyons said one thing the Clarity polls did well was their approach to measuring the partisanship of their samples. First, respondents were asked, “Do you think of yourself as a Republican, a Democrat or an independent?” In July those numbers came out 44 percent Republican, 18 percent Democrat, and 38 percent independent. In August, it was 38-21-41.

Then those who said they think of themselves as independents were asked, “Would you say you generally lean Republican, lean Democratic, or are purely an independent?” In July, 33 percent of that subgroup said it leaned Republican, 22 percent Democratic, and 45 percent purely independent. In August, those numbers were 36-19-46.

“We know that those ‘leaners’ … are just as partisan in their voting behavior as the Republicans and Democrats who self-identify,” Lyons said. “Partisanship is kind of a dirty word. People are hesitant to slap a label on themselves.”

The July poll also included favorability questions about Little and Jordan (Little got 25 percent favorable, 22 percent unfavorable, 40 percent neutral and 13 percent unsure; Jordan got 26 percent favorable, 16 percent unfavorable, 31 percent neutral and 26 percent unsure).

The August poll included a series of questions about issues, including one on the Medicaid expansion initiative (45 percent said they’re planning to vote yes, 19 percent no, and 36 percent unsure). Lyons said the wording of the question appeared “reasonable,” and said, “Medicaid expansion across the country is generally polling pretty well, even in conservative states like Idaho. Those numbers are not surprising to me at all.” He added, “These are encouraging numbers for the folks who are trying to expand Medicaid.”

In Lim’s news release about the August poll, he said, “Jordan’s path to victory is clear: She needs to consolidate the ‘yes’ vote on the Medicaid expansion initiative and Little must be tied to the ‘no’ vote.”

Among the issue questions in the August poll were two clearly designed to be negative toward Little, one saying he “has proven he will not protect Idaho public land,” and the other saying he “doesn’t take positions on the issues that affect Idaho voters the most,” including the Medicaid expansion initiative, on which Little has said he’ll stay neutral and leave the outcome to voters.

“To their credit, they put those questions afterwards on the survey,” Lyons said – after respondents already had been asked who they favor in the race. That means the suggestive wording of those questions didn’t taint the original answers on candidate choice.

Lyons called Clarity’s B-minus grade from fivethirtyeight.com “fairly respectable,” saying, “There certainly are plenty of pollsters that they give F grades to. So that means it’s not bad.”