A Salt Lake Tribune-Hinckley Institute of Politics poll conducted on Salt Lake City’s hot mayoral race and related policy questions included at least several dozen respondents who live outside of the capital, raising questions about the survey’s integrity.
The poll relied on the Cicero Group to conduct the interviews that took place from July 29-31. The polling contractor and its call center questioned 444 people who were described as “likely voters” in Salt Lake City, through landlines, cellphones and online. But it turns out that at least 57 of those voters actually live in other cities.
This error could skew the poll results, originally published Aug. 8, although Cicero’s team downplayed the severity of the problem. A recalculation of the results Thursday, after stripping out the known 57 non-Salt Lake City residents, showed little difference from the original.
No candidate changed position, with Jim Dabakis remaining in the lead at 21%. In second place, state Sen. Luz Escamilla gained a single percentage point to 18% and City Councilwoman Erin Mendenhall gained one point to 14%.
The poll numbers did not change for David Garbett at 9%, David Ibarra at 6%, Stan Penfold at 5% and Rainer Huck and Richard Goldberger bringing up the rear, each at 0%.
Undecided voters dropped two points to 26%. The margin of error increased from 4.6 percentage points to 4.96 points.
“We’re confident in the results,” Blake Moore, a principal of Cicero, said.
Jason Perry, director of the Hinckley Institute, similarly said he was satisfied with the recalculated survey despite the troubling discovery of some non-Salt Lake City residents in the original results.
He and his team spent much of Thursday working with Cicero “to make sure that these rankings are accurate,” Perry said. “The fact that it’s stayed consistent demonstrates the validity of this poll.”
Tribune Editor Jennifer Napier-Pearce said even though she has received assurances about the reliability of the revised poll, she remains unhappy about the deficiencies in the first poll. Saying the pollster or its call center “really fell short,” she said the newspaper would not work with the Cicero Group on future polls.
“When we make mistakes we own them and try to make it right and that’s what we’re doing here. It’s just unfortunate,” Napier-Pearce said.
The newspaper discovered the problem when reporters called a sampling of the respondents to more deeply explore their answers on a raft of questions focused on Salt Lake City politics and issues, such as air quality. Also polled were views on some candidates proposals for free bus service and a transition to a new model of services for people experiencing homelessness. Six of the individuals contacted told reporters they reside outside the city — five of them in Millcreek and one in South Salt Lake City.
A further look at ZIP codes provided by the pollster also indicated some respondents in areas outside the city — including at least half a dozen outside Salt Lake County.
Initially, Cicero told the Hinckley Institute that these few outliers likely were the result of people who recently moving out of the capital city. Based on a fact-checking second round of interviews Wednesday evening and Thursday, that clearly was not true in some cases.
Five respondents said they have never resided in Salt Lake City.
One woman had lived at her present residence in Millcreek, formerly unincorporated Salt Lake County, for 44 years since moving from Salt Lake City in 1975.
Another has resided in what is now Millcreek for 15 years, while a Sandy woman said she moved out of the capital city nine years ago.
All said they have voted in their own recent local municipal elections, although one man who said he moved from Salt Lake City to West Jordan five years ago, said he did recently receive a Salt Lake City ballot.
One problem with the poll was that it did not ask potential respondents if they currently live in the city, relying on the accuracy of the voting list purchased from the government. Instead it asked a question about the individual’s interest in the race to get to a sample of “likely voters.”
That reliance on the list’s accuracy was a bad call, particularly because Cicero said it was using a 2016 voting database, according to Salt Lake County Clerk Sherrie Swensen.
“That’s so old. I can’t even tell you how outdated that would be,” Swensen said Thursday evening. “That’s horrible. They should have bought a new list.”
Residents nowadays are so mobile that even with county efforts, in compliance with state law, to keep voter lists current by using the national change-of-address database (compiled from U.S. Postal Service notifications), errors crop up.
Salt Lake County mailed out 324,000 countywide ballots July 23 after this rigorous updating and still received 12,000 returned by the Postal Service, Swensen said. Some 3,000 of those had moved within the county and the clerk’s office updates those to reflect the new address.
She said a voter list for the entire county would run $300 and a city voter list would cost less.
Moore, with Cicero, defended the use of the 2016 voter list and said comparing it with a 2018 version produced little change. “To our team, it didn’t seem like it was going to be problematic. We were comfortable with it.”
He acknowledged that it probably would have been an extra safeguard to add a screening question to ascertain whether respondents were Salt Lake City residents, but said self-identified address information is generally less reliable than voter list addresses.
He defended the work of the pollster, which for years operated under the Dan Jones & Associates name, and of its call center, Clear Insights.
“We’ve had a long history of running rigorous and accurate polling,” Moore said.
The primary on Tuesday will narrow the field of eight to the top two vote-getters, who will go head to head in November’s election to replace Mayor Jackie Biskupski, who is not seeking reelection to a second four-year term.
Told of the inclusion of at least some non-Salt Lake City resident in the poll, Mendenhall said: “It just shows it’s anybody’s race.”
Garbett, a former attorney with the Southern Utah Wilderness Society, said he hopes voters take a lesson from this gaffe.
“This is a reminder that you should vote for the candidate that you agree with and want to see [win] rather than putting too much stock in polling,” he said.
Dabakis, Escamilla and Penfold did not offer any comment when told of the issue of some non-resident respondents to the poll.
Ibarra said, “Based on the numbers we have [from internal polling], something is drastically wrong” with the Tribune-Hinckley poll. “I believe I will prevail."
Follow @danattrib