Kari Lake had lost Arizona’s 2022 governor’s race by thousands of votes, falsely declared that she’d been robbed, and was trying to overturn the election results in court when, finally, she was declared the winner of the race in March.
But not by Arizona’s millions of voters. Rather, well-known pollster Rasmussen Reports announced that, actually, a majority of its survey respondents claimed to have voted for Lake over Gov. Katie Hobbs (D). Not only that, Rasmussen Reports said — it was an eight-point blowout in Lake’s favor.
The news from Rasmussen Reports, which was based on an unprecedented 1,001-person poll taken four months after the election, was the boldest claim yet from the pollster, heralding a miraculous supposed 92% turnout rate in the election — rather than the certified rate of 62.56%.
Mark Mitchell, Rasmussen Reports’ lead pollster, didn’t simply report that Arizonans were wary of voter fraud, another one of the poll’s findings. He leaned into the conspiracy theory that Lake had been robbed of a victory.
“This is a very strong signal that Arizona voters are aware that there are irregularities and cheating,” Mitchell told Steve Bannon in an interview on Bannon’s show, adding later: “What we found is that voters told us that just four months ago they elected Kari Lake as the governor of Arizona by 8 points.”
Lake appeared ecstatic. “NEW Rasmussen Poll Reveals the Red Wave did happen,” she wrote in a tweet that claimed Republican candidates Abe Hamadeh and Mark Finchem had also wrongly been declared losers. “Lake, Hamadeh and Finchem won according to poll and 50% of Arizonans say the 2022 Election issues were done intentionally to hurt Republicans.”
The practice of polling an election four months after the fact, and using the results — an imperfect snapshot of a tiny subset of Arizona voters — as the basis for questioning an actual election result was “totally backwards,” “out of bounds,” and “very inappropriate,” political analysts told HuffPost.
It was the culmination of a bizarre shift for Rasmussen Reports. After two decades of fairly consistent right-leaning polling of elections and political issues, the firm had stopped acting like a normal pollster. Rather than simply asking Americans about their beliefs in conspiracy theories and fringe views, Rasmussen Reports has amplified those beliefs, lending them its legitimacy as a household name in politics — to the delight of right-wing politicians and influencers eager for new material.
Citing its own poll results on Americans who think they know someone killed by the COVID-19 vaccine, for example, Rasmussen Reports tweeted a cartoon showing a cemetery with the word “coincidence” on every headstone. “ICYMI: We’ve all come a long way from ‘safe and effective,’” the firm tweeted.
Months after “Dilbert” creator Scott Adams referenced a Rasmussen Reports poll — asking respondents for their opinion on “It’s okay to be white,” which has become a far-right troll catchphrase — as part of a racist meltdown that ended with him urging white people “to get the hell away from Black people,” Rasmussen Reports is still retweeting Adams as he tries to gain new fans.
Even without corresponding poll data, Rasmussen Reports’ Twitter page has frequently amplified fringe conspiracy theories to its audience of 447,000 followers. Earlier this month, above a poll that simply asked whether voters were concerned about “cheating” in the 2024 election, Rasmussen Reports tweeted, “Printing & mailing real election ballots for phony voters at phony address’s [sic] that get retrieved, voted, mailed back & counted is likely the key to the organized theft of the 2024 presidential election.”
Now, as it amplifies ever-zanier claims, Rasmussen Reports — former President Donald Trump’s favorite pollster, of course — is facing a rare instance of pushback from within the insular community of pollsters and political scientists.
“I don’t consider them a legitimate pollster,” Alan Abramowitz, an emeritus professor of political science at Emory University, told HuffPost. “They have a conservative audience and they’re trying to sell subscriptions.”
Late last month, the editorial director of data analytics at ABC News — and its subsidiary, the popular data-oriented news site FiveThirtyEight — threatened to “formally ban” Rasmussen Reports from its election coverage, including removing it from election forecasting models, unless it answered questions about methodology and its relationship with several right-wing outlets, including Bannon’s “War Room.”
ABC News wouldn’t be alone: CNN doesn’t report on Rasmussen Reports poll results because they don’t meet the network’s editorial standards, analysts there have noted in the past. Referring to those standards, a spokesperson for the network told HuffPost, “We can confirm this still reflects our current reporting standards.”
Faced with the mounting backlash, Rasmussen Reports lashed out, publishing the email from ABC and staging a populist appeal.
“Mainstream and legacy media outlets ignore, slander, and attack us for polling on ‘forbidden’ topics,” the firm tweeted last week, before a long list of polls asking about various right-wing grievance topics. “If they dislike the results, then they should correct us by testing the same questions. Yet we stand alone.”
The Walmart Of Pollsters
Founded in 2003 by Scott Rasmussen — who years earlier had co-founded ESPN with his father, Bill — Rasmussen Reports has long been a reliably conservative-slanted pollster of the American public, generally performing better in years when Republicans had an advantage (like 2016), and worse in years when Democrats did better (like 2018). (Scott Rasmussen left the firm in 2013.)
The firm is distinguished by a few crucial factors: For one thing, Rasmussen Reports uses an automated digital voice to poll people on the phone rather than live operators, which critics note excludes people who only use cell phones (thanks to a federal law prohibiting robocalls to cell phones). Because Rasmussen Reports doesn’t use human operators, it can get a lot of polling done. And this means it can test lots of questions, inputting tons of new material into aggregators like FiveThirtyEight — it’s the second-most-prolific pollster analyzed by the site — as well as providing red meat issue polls for subscribers. Ever since Gallup ended its daily updates on the president’s approval rating in 2018, Rasmussen Reports is also the only pollster still publishing a “daily presidential tracking poll.”
The sheer volume of polls Rasmussen Reports produces can sway polling aggregators, in addition to keeping the firm’s name in headlines, no matter when the next election is.
That level of influence is “sort of like Walmart, in a way,” said Barry Burden, director of the Elections Research Center and a political science professor at the University of Wisconsin–Madison.
“They’re just big players in the market, so they get a lot of attention and can move that market on their own,” Burden said.
“They’re just big players in the market, so they get a lot of attention and can move that market on their own.”
– Barry Burden
Plenty of polling companies do private work for campaigns, interest groups and others, but some of Rasmussen Reports’ sponsors raise eyebrows: The sponsor of the March Kari Lake poll was “College Republicans United,” which counts among its values “opposition to immigration and multiculturalism.” Other sponsors have included Jack Posobiec, the far-right operative who popularized the “Pizzagate” conspiracy theory, and the New York Post’s Miranda Devine, who’s described her relationship with Rasmussen Reports as “symbiotic” — ”They sometimes give me previews of the polls, an early embargo, so I get to break stuff,” she told Politico last year. (According to Rasmussen Reports, “Co-sponsors can suggest questions and other matters, but we will retain the final say on all wording.”)
Rasmussen Reports is also somewhat secretive about its online polling procedures, as well as the unlisted “factors” they use to weight their survey respondent data in proportion to the American public. Earlier this month, the firm referred to its polling practices as “proprietary” and retweeted a post that claimed ABC News’ questions on that topic were “None of your damn business.”
“There’s something in their methodology, and we don’t know what it is, that tilts them toward Republicans all the time,” said Natalie Jackson, an independent research consultant and former director of research at the Public Religion Research Institute. (She also previously worked as HuffPost’s senior polling editor.) “It’s the not really knowing what it is that causes problems.”
Another problem: Leading questions, or the practice of polling questions where “an initial phrase leads the respondent by suggesting the position or stance of an authority with which it might be difficult for the respondent to disagree,” according to The American Association of Public Opinion Research. The AAPOR discourages using such questions — and urges pollsters to avoid “language that pushes respondents to respond in a certain way or that presents only one side of an issue” — but they appear in several of Rasmussen Reports’ polls.
After Rep. Marjorie Taylor Greene (R-Ga.) tweeted in 2021 that a whopping 40% of Democrats said President Joe Biden deserved to be impeached, The Washington Post’s Philip Bump pointed out the wording of the actual poll question from Rasmussen Reports: “Do you agree or disagree with this statement: ‘I think Joe Biden deserves to be impeached because he’s abandoned thousands of Afghans who fought with us and he’s going to abandon some American citizens because he capitulated to the Taliban to a 31 August deadline’?”
On Twitter, just like Greene, Rasmussen Reports claimed simply that 40% of Democrats said “Biden deserves to be impeached.”
Other polls have offered different impeachment logic: “Some Republicans in Congress have endorsed articles of impeachment against President Joe Biden, citing his immigration policy and his failure in Afghanistan, among other reasons. Do you support or oppose impeaching President Biden?” Rasmussen Reports asked last August. The firm’s headline read simply, “Impeach Biden? Most GOP, Independent Voters Say ‘Yes.’”
Other polls present false dichotomies, like when Rasmussen Reports asked likely voters in April, “Which was more harmful to democracy in America: Russian agents interfering in the 2016 presidential election or U.S. intelligence officials interfering in the 2020 presidential election?”
In March this year, Rasmussen Reports headlined its Lake poll with another major announcement: “Most Arizona Voters Believe Election ‘Irregularities’ Affected Outcome.”
The “crosstabs” of the poll revealed that the question asked of voters was explicitly partisan: “How likely is it that 2022 election irregularities in heavily Republican areas of Arizona happened intentionally to suppress Republican votes?” The question includes the implicit assumption that “irregularities” — in this case, printer issues affecting various Maricopa County polling places on Election Day — did in fact occur solely or disproportionately in “heavily Republican areas of Arizona.”
But that’s not actually true: “Voting centers in both liberal and conservative parts of Maricopa County were impacted by the printing issues,” according to an Associated Press fact check from four months earlier that cited a spokesperson for the Maricopa County Elections Department.
“There are certainly some Republican areas impacted, but there are a significant number of Democratic-leaning areas that were impacted, as well as a number of swing areas or very competitive areas,” a Republican pollster, Paul Bentz, told AP. A separate Washington Post analysis found that the proportion of registered Republicans in precincts affected by printer issues was “virtually the same” as the share of registered Republicans county-wide — 37% and 35%, respectively.
Perhaps unsurprisingly, Republican respondents overwhelmingly agreed with the poll question, while majorities of both Democrats and unaffiliated voters disagreed with the prompt or were unsure.
![Demonstrators protest outside the Maricopa County Tabulation and Election Center on Nov. 12, 2022, in Phoenix, Arizona.](https://img.huffingtonpost.com/asset/64b061172600006100752056.jpeg?ops=scalefit_720_noupscale)
The Washington Post via Getty Images
‘Stick With The Fundamentals’
Rasmussen Reports isn’t the only firm to use proprietary methods and ask leading questions, but it stands alone in what it does with its data — taking up the mantle of conservative crusader and appearing often on conservative media to push talking points.
“Election fraud deniers are a rapidly shrinking minority,” the company tweeted a few days ago. “That’s good.”
Rather than simply report that, according to a June poll, 55% of Democrats are at least somewhat concerned about cheating affecting the 2024 race, Rasmussen Reports went a step further, tweeting that the majority was “probably because they are waking up to coordinated suppression of an important topic.”
The same leaps of logic apply to the firm’s COVID-19 vaccine polling. In March, Rasmussen Reports tweeted “Our polling indicates @CDCgov is lying when it claims vax side effects and deaths are ‘rare,’” but its evidence for the claim was absurd: One poll it cited merely asked people if they were concerned “about the potential of harmful side effects of the COVID-19 vaccine,” and another asked respondents whether they personally knew anyone “whose death you think may have been caused by side effects of COVID-19 vaccines?”
Rasmussen Reports’ write-up of the latter poll cited the conspiracy theory film “Died Suddenly,” which as HuffPost reported, was loaded with examples of supposed vaccine-related deaths and injuries that were, in fact, totally unrelated, with several happening years before the existence of COVID-19. Rasmussen Reports acknowledged that the film had been criticized as promoting “debunked” science — but added, as if it were meaningful, that it had “been seen by some 15 million people.”
“They’re letting themselves be used — or are intentionally using their products — to provide support for dangerous, fringe ideas in American politics that aren’t grounded in fact, that are about innuendo or rumors, and the polls themselves are measuring innuendo and rumor at times,” Burden said. “And so it brings a kind of vicious cycle that lets kooky ideas have longer lives than they should.”
It’s not clear what has fueled the far-right drift at Rasmussen Reports in recent years. Scott Rasmussen left a few years after the private equity firm Noson Lawen Partners made a “major growth capital investment” in the firm. A 2013 press release on Rasmussen’s departure from his namesake cited only “disagreements over company business strategies.” (“I have never commented on my firm since the day I left and will never do so,” Rasmussen told HuffPost.)
Still, even then, complaints about the company hewed to its polling practices. “[T]here’s not yet evidence that post-Rasmussen Rasmussen Reports won’t tolerate the same flawed polls that adulterated the polling averages for most of 2012,” writer Nate Cohn, now The New York Times’ chief political analyst, lamented in The New Republic.
The firm’s tone seems to have changed with the 2020 election cycle. In late 2019, the Rasmussen Reports account was griping about “Google search bias” and Hunter Biden’s alleged Ukraine-related corruption. When The New York Times’ Twitter account posted on Election Day that “The role of declaring the winner of a presidential election in the U.S. falls to the news media,” Rasmussen Reports compared the comment to George Orwell’s “1984.”
A few days after Election Day, Rasmussen Reports retweeted a post describing “how they did it” — presumably, how Democrats supposedly stole the election — that falsely accused seven states of “electronic flipping of totals,” and four states of “ballot dumps.”
Later that day, when pro-Trump attorney Lin Wood suggested Georgia Gov. Brian Kemp (R) “go to prison” alongside Georgia’s secretary of state, Rasmussen Reports retweeted the post, referring to it as “an interesting turn.”
In late December, after the news networks had called the race for Biden and after Trump had summoned his supporters to Washington, D.C., on Jan. 6, Rasmussen Reports’ Twitter account posted a quote falsely attributed to Stalin — ”Those who cast the votes decide nothing. Those who count the votes decide everything” — then appeared to urge Vice President Mike Pence to ignore the votes of seven “disputed” states.
“If the votes of all 7 contested states are registered as zero, President Trump will have 232 votes, Joe Biden will have 222,” the account tweeted, echoing Trump’s call to overturn the election results.
Rasmussen Reports, according to a frequently asked questions page, is run by three people.
Mitchell, the chief pollster, is in many ways the company’s public face, appearing often on Bannon’s show and others, and being the talking head for YouTube videos like “Rasmussen Polls: COVID vs. Vaccine. Americans Tell us Which is the Biggest Killer.” It’s not clear what polling experience Mitchell had before joining Rasmussen Reports— according to a LinkedIn account, he was head of order operations at Walmart’s e-commerce department before taking the job in 2021.
Robert Stacy McCain, listed as Rasmussen Reports’ managing editor, is hardly discussed at all on its website and social media channels. But he has a long history as a conservative writer, primarily through his blog “The Other McCain.” McCain, a former Washington Times reporter and editor, drew scrutiny in the early days of the blogosphere after the Southern Poverty Law Center reported that he’d been a member of the neo-Confederate organization “League of the South,” which the SPLC designates as a hate group, while he was working for the conservative outlet.
Among other things, McCain has been widely quoted as writing, years ago, “the media now force interracial images into the public mind and a number of perfectly rational people react to these images with an altogether natural revulsion. The white person who does not mind transacting business with a black bank clerk may yet be averse to accepting the clerk as his sister-in-law, and THIS IS NOT RACISM, no matter what Madison Avenue, Hollywood and Washington tell us.”
During an interview in 2006, he acknowledged having previously written on message boards under the name “BurkeCalhounDabney” — leading to speculation that he’d written under the name Burke C. Dabney for American Renaissance, a white supremacist website. An article under that name argued that attempting to lower birth rates for white women was actually a form of “racial suicide.”
McCain, who didn’t respond to HuffPost’s request for comment, has denied being a racist for years. And he wrote in 2009 that his “acquaintance” with the League of the South began as a newspaper assignment.
“[T]he point is that I was pursuing my professional duty when I first came into contact with the League of the South, and of my subsequent involvement, there are many things that people think they know ― on the basis of SPLC reports ― which are not necessarily true,” he wrote. “And there are many, many thinks [sic] that people do not know.”
When McCain was suspended from Twitter in 2016 for reportedly violating the site’s “targeted abuse” rules — his account is still suspended — he told BuzzFeed News, “The idea that my personal voice will never be heard again on Twitter. Ha! We’ll see.”
McCain and Mitchell run Rasmussen Reports alongside Ted Carroll, president and co-founder of Noson Lawen Partners, which now owns Rasmussen Reports.
In a 2018 story republished on Rasmussen Reports’ website, Carroll said that despite the firm’s “growing megaphone,” its pollsters “need to stick with the fundamentals of proper surveying and ask clear unbiased questions to the public. It’s not what we think; it’s what the public thinks that is important.”
Five years later, when HuffPost reached out earlier this week to request an interview for this story, Carroll declined, citing our claim that Rasmussen Reports had endorsed and amplified conspiracy theories, rather than simply relaying public sentiment.
“Official state sponsored investigations have uncovered significant improper and illegal election practices in multiple American states and those official findings underlie both our polling and story focus — as you might know by reading our Twitter feed,” Carroll wrote, though he didn’t respond to a follow-up question asking for specifics. “You can ignore official evidence if you want but 62% of U.S. likely voters tell us they now believe 2020 was affected by cheating and a majority of voters (including Democrats) now believe 2024 will be affected by cheating also.”
Then, he seemed to embody the voice of Rasmussen Reports’ Twitter page.
“We’re not in the education or debating business,” he added, “so please go ahead and publish your silly partisan hit piece and we will respond if and as we see fit.”