A series of graphs published by My Pillow CEO Mike Lindell on his social media channels and livestreams appearing to show real-time U.S. midterm election updates do not indicate or prove fraud, experts told Reuters.
Lindell published four graphs on his Instagram account purporting to show real-time vote counts during the elections. Repeatedly during livestreams published via his channel on the platform Frank, Lindell refers to apparent upswings for Democrat candidates as “crime spikes,” inferring that sharp increases in tallies for one candidate is proof of nefarious activity.
But such ‘spikes’ seen in graphs using unofficial election data do not indicate or provide proof of fraud, with races rarely officially called on election night (here).
UNOFFICIAL ELECTION DATA
Lindell claims that the source of the data to create the graphs is Edison Research (viewable 44.58 here).
Edison Research does not publish official results – which often take weeks to certify – but instead provides unofficial tallies as the results are being counted per county and state.
“All vote counts are unofficial until they are certified by each state, which can happen weeks after the elections. Until then, reports of these unofficial vote counts are just that: unofficial,” Sukrit Venkatagiri, a research scientist at the Center for an Informed Public, University of Washington, told Reuters.
Venkatagiri co-authored a piece for the Election Integrity Partnership (EIP), a non-partisan research coalition, on how real-time visualizations of vote counts can lead to unfounded claims of voter fraud (here).
“Visualizing these unofficial vote counts in real-time, where the incoming data may not be uniform in distribution can lead to these ‘spikes.’” Such visualizations can be potentially misleading, Venkatagiri said, adding: “It’s important not to attribute this to malice, but just a quirk of the underlying data, data entry errors, or errors later in the data collection and reporting pipeline.”
COMMON CAUSES OF ‘SPIKES’
Rob Farbman, the executive vice president at Edison Research, told Reuters that the company’s data did not indicate fraud, and that upswings in votes per real-time data do not point to nefarious activity.
“Counties all over the country are tilted heavily to either Democrats or Republicans,” Farbman said. “If any one of them is a large county and reports its vote that can cause a spike in a line graph.”
“Votes are counted at different times depending on the county. When the votes are released, we report them. The statewide vote total will vary by party as the night progresses,” he added.
Walter R. Mebane, Jr., a professor at the departments of political science and of statistics at the University of Michigan, told Reuters: “Different kinds of places and kinds of voting modalities report vote tallies at different times.”
“Voters with diverse preferences live in different places and use different kinds of modalities,” Mebane said, adding: “None of that variation has anything to do with voter fraud or other nefarious activity.”
“Some states can count mail-in ballots before polls close, causing a large number of votes being reported at once initially,” Venkatagiri added (here).
“There can also be delays in counting and tabulating in-person votes,” Venkatagiri said. “This can result in vote count changes later on, and some of these changes may favor a candidate from one party over another,” he told Reuters.
MINNESOTA GOVERNOR’S RACE
Lindell posted one such graph that showed supposed real-time updates of the Minnesota governor race between Democrat Tim Walz and Republican Scott Jensen with a caption that reads: “They went for the complete steal” (here).
In a livestream published Nov. 8, a guest speaking on the livestream claims that the graph shows “almost half a million votes dumped at once” in favor of Walz (viewable from 18:15s here). “Every update tends to mirror each other,” the guest adds, with Lindell suggesting that the graph is “statistically impossible” given that “each line matches the other one.”
Cassondra Knudson, the deputy communications director at the Minnesota Secretary of State, told Reuters that Minnesota's largest county Hennepin reported around 9 p.m. CT that votes leaned toward Walz by a large margin. This tally aligns with the county's history of voting in favor of Democratic candidates (here), (here), (here), (here).
Per the Minnesota Secretary of State website, Hennepin County had 791,682 eligible voters, with Ramsey County following behind with 314,797 voters (here). Walz won Hennepin County with 70.4% of the vote to Jensen’s 26.8% (here).
MICHIGAN ATTORNEY GENERAL RACE
Lindell also posted a graph pertaining to the Michigan Attorney General race between Democrat Dana Nessel and Republican Matt DePerno (here), with a caption claiming that the race was “being stolen in the middle of the night” from the latter. In the livestream, the guest discusses an apparent spike in votes in favor of Nessel shortly after 1:25 a.m. (viewable from 29:45s here).
Edison had updated the Michigan governor’s race much earlier in the night compared to the attorney general race, Farbman said. This was due to Edison prioritizing accessing the governor’s race from the state data feed ahead of the attorney general vote. This would have created a bump in votes in the attorney general race later in the evening compared to the governor’s race.
The jump in votes in the night represented this update, which included Wayne County where Nessel had an over 200,000 lead at that point, Farbman told Reuters.
ILLINOIS SENATE RACE
The My Pillow CEO posted a graph showing apparent real-time counting of the Illinois Senate race between Democrat Tammy Duckworth race and Republican Kathy Salvi, with Lindell claiming via Instagram that there was a “million vote spike and the only explanation is corrupt electronic voting machines” (here).
The graph appears to show an upswing and subsequent decline for Duckworth, with the guest speaking on the livestream claiming that the graph is proof that one “can’t trust the system” (viewable from 34:55s here).
The spike and subsequent dip were caused by a typographical error, Farbman told Reuters. In a data entry, Edison added a nine at the beginning of Duckworth’s number in Rock Island, with the figure fixed minutes later.
GEORGIA SENATE RACE
Through Instagram, Lindell again claimed that a race was being “stolen” from a Republican candidate, publishing a graph appearing to show real-time updates in the Georgia Senate race between Republican Herschel Walker and Democrat Raphael Warnock (here).
A guest on his livestream claimed most races show Democrats having an early lead, likely because “that’s where most of the cheating happens.” The graph appears to show the race narrowing, with a subsequent uptick in favor of Warnock. Lindell claims that the uptick is proof of “real-time crime” (here).
“In Georgia, we call a 9pm upload toward one candidate ‘The Atlanta metro has reported,’” Robert Sinners, the communications director at the Georgia Secretary of State office, told Reuters.
“It stands to reason that when Fulton or DeKalb County reports their vote totals, the count for the Democrat candidates will increase,” Sinners said, due to “our largest counties getting it done early and accurately.”
Farbman similarly told Reuters that the graph correlates with DeKalb reporting (increasing from 22,716 to 210,388) at exactly when the uptick is viewable. “DeKalb is the largest county in Georgia and extremely Democratic,” he added.
Representatives for Lindell did not respond to Reuters’ request for comment.
False. The four graphs appearing to show real-time midterm election data do not point to fraud or nefarious activity. Visualizing unofficial vote counts in real-time where the incoming data may be non-uniform in distribution can lead to apparent “spikes” in graphs, experts told Reuters.
This article was produced by the Reuters Fact Check team. Read more about our fact-checking work here.
Our Standards: The Thomson Reuters Trust Principles.