Can we trust the polling in Utah’s U.S. Senate race?

Frank Pignanelli & LaVarr Webb


The U.S. Senate race between Sen. Mike Lee and Evan McMullin is a rough-and-tumble slugfest. Campaign activities are generating arguments among politicos over polling, negative ads, superPACs and all the other campaign stuff hacks like us enjoy. We share the fun.

Both campaigns have revealed results of internal polling to show their candidate ahead. The Utah Debate Commission/Lighthouse Research survey showed Lee ahead 48%-37% with 5% undecided. Last week’s Deseret News/Hinckley Institute of Politics poll added fuel to the fire, showing the race essentially tied (Lee 36%, McMullin 34%, 16% undecided). What does this polling puzzle say about the race?

Pignanelli: “Don’t worry about polls, but if you do, don’t admit it.” — Rosalynn Carter  

Utah’s Senate race, already the most peculiar in an abnormal midterm election season filled with atypical contests, is now more unusual. The disparity in polling is fostering heated quarrels about whether the methodology used is capturing the preference of the general population or those most likely to vote. This is important as the latter category will select the next senator.

Campaign veterans are distilling the distinction. Some survey organizations are beginning with essentially the phonebook to develop a list of respondents, then asking whether they are likely to vote. (Unfortunately, good intentions do not always remain through election day.) Other entities are using databases of citizens with a voting history to construct a call list and asking the likely to vote question. The conjuncture in the former method, by not filtering for truly likely voters, is providing a snapshot of the general population and not an accurate reflection of individuals following the race who will cast a ballot.

Experienced local politicos are suggesting surveys initiated with a list of determined likely voters show Lee ahead with a comfortable margin. Furthermore, the well-respected Nate Silver and his website of FiveThirtyEight.com analyzes surveys conducted in every federal race in the country. Based on a scrutiny of the polls in the Utah race (i.e. quality, recency, sample size and partisan lean) they project Lee has a 93% chance of winning.

The Senate race — a weird outlier — could be influential for future campaigns and how they poll.

Webb: Properly conducted survey research is an accurate reflection of the attitudes of those surveyed at a snapshot in time. But it’s very difficult to predict voting outcomes because it’s hard to survey, in the correct demographic proportions, people who will actually vote.

Polling in the last few elections has underestimated the strength of many Republican candidates. In 2020, for example, President Joe Biden’s victory over Donald Trump was closer than the polls indicated. And the number of House seats picked up by Republicans was a big surprise.

Even The New York Times recently published a lengthy story by its political polling guru admitting that many polls don’t accurately reflect the turnout of conservative “Trump voters.”

Most pollsters try hard to be accurate. They’ve made adjustments to find more Trump voters. But it’s difficult to get conservative Republicans to participate in polling. We’re in a very divisive, skeptical time, and many conservatives don’t trust pollsters or the media. Most people only answer cellphones when the caller is known, and if they do answer and it’s a polling firm, most hang up. It’s also hard to get Trump voters to participate in online survey panels. When pollsters know a demographic group is underrepresented, they “weight” the results to make up for it. But accuracy remains difficult. 

All of this leads me to believe that Republicans, including Lee, will do at least a little better than is indicated by polling. Certainly, if a Republican is 20 points behind, he or she is going to lose. But if the race is close, it might make sense to give the Republican candidate a couple extra percentage points.

Will I be right? We’ll see in a little over five weeks. 

Lee and McMullin are both supported by well-funded super PACs (independent expenditure political committees). These entities are running negative advertising, as are each of the campaigns. Who will the attacks help or hurt?

Pignanelli: This Senate race confirms superPACs spend millions of dollars on awful ads that fail to resonate because of anemic messaging. Issues exist that could be exploited with creative and even humor-oriented commercials. McMullin is especially vulnerable to such, who has yet to establish an identity other than he is not Mike Lee.

Webb: I would ignore the attack ads and vote on the basics of this race: Lee is a conservative Republican who votes accordingly and who has been a reliable Trump supporter. What you see is what you get. McMullin is an independent who is strongly supported by Democrats and who has made a career of attacking Trump and those who support him. On key issues, he is more likely to vote with Democrats than with Republicans and his election could ensure Democratic control of the Senate.

Is there a possibility the trajectory of this race could change before Election Day?

Pignanelli: Deeper legal trouble for Trump could move some undecided for McMullin. Conversely, additional economic worries will bolster a red wave in Utah that helps Lee.

Webb: Anything can happen, but I don’t expect any dramatic October surprise. Lee has more money to spend, so expect his advertising and outreach to pick up significantly in the days ahead. McMullin will try to keep pace. 

Previous
Previous

Do endorsements really matter in the U.S. Senate race?

Next
Next

Opinion: The Great Salt Lake, a new Utah flag, abortion debates — what will affect midterms most?