Evan McMullin’s campaign says they have clawed out a lead over Mike Lee in Utah’s U.S. Senate race, pointing to a new internal poll that gives them a one-point advantage. On the other hand, Lee’s campaign has internal polling showing him with a comfortable 18-point lead. Should you believe either of these?
Election polls are great for tracking the “horse race” aspect of a political campaign to track who may be ahead or behind at a certain point in time. They can also provide a snapshot of public opinion.
When it comes to internal polling, the best approach is to have a high level of skepticism about the results.
“My general rule is I take all polls with a grain of salt, and I take all internal polls with a heaping tablespoon of salt,” John Cluverius, director of survey research at UMass Lowell’s Center for Public Opinion, says.
Consider that campaigns don’t often release polls that are unfavorable to them. When made public, they invariably show a big lead or the race starting to turn in their favor. That can be used to boost fundraising or generate positive media coverage.
Shortly after McMullin’s surprising internal poll was published in the Washington Post, his campaign sent out a fundraising appeal with the attention-grabbing subject line, “It’s official: We’ve taken the lead.” It didn’t take long for Lee’s camp to counter the McMullin poll, pushing out their latest internal numbers to show that they still held a commanding lead.
Cluverius says the biggest problem with internal polls is a lack of transparency about the inner workings of the polls. Instead of detailed results, most internal polls are released as a polling memo highlighting only the head-to-head results. Internal polling often includes questions that campaigns will use to test possible messages, which can impact other results.
“I don’t want to say all internal polls are trying to deceive the public or that they’re bad. Campaigns make strategy decisions based on their internal polling results, so they want the best data possible,” Cluverius says. “But, historically, internal polls create conditions that are more advantageous for their clients.”
The pair of internal polls released this week are likely outliers. McMullin’s is the only survey we’ve seen where he is ahead. Lee’s poll gives him a larger lead than any other survey. Other polling in the race shows Lee ahead anywhere from five to 14%. The FiveThirtyEight.com polling average for the contest puts Lee ahead by just over 8%.
It also is challenging to get an accurate picture of the state of the contest by looking at independent polling. Surveys by the Center Street PAC consistently show Lee ahead by double digits, while other polls suggest a much closer race.
The most recent poll conducted for the Deseret News and Hinckley Institute of Politics shows a five-point gap in favor of Lee. However, the survey questioned registered voters instead of likely voters. A sampling of likely voters is usually considered to be more accurate. For reference, a pre-GOP primary poll showed Lee with 49% while his two opponents in that election, Becky Edwards and Ally Isom, with 19% and 6%, respectively. Lee won the primary with 62%, 13 points better than the poll suggested. Edwards was at 30%, and Isom got 8% of the vote.