PolitiFact quote chops Trump down to a "False" rating

 It's weird to see one of the supposed great lights of contemporary journalism engage in conduct that wouldn't fly in an introductory journalism class.

That's PolitiFact, again and again.

This time it's yesterday's "False" rating from PolitiFact North Carolina no, kidding, it's Lou Jacobson and PolitiFact National sticking the "North Carolina" tag on a fact check mentioning a North Carolina politician,  Republican Sen. Thom Tillis.

PolitiFact frames the entire issue as whether Trump considers Tillis a current senator:

President Donald Trump and Sen. Thom Tillis, R-N.C., have been at odds over the next chairman of the Federal Reserve.

They can’t even agree on whether Tillis is still a senator.

And PolitiFact supposedly has the transcript to prove the point:

Trump insisted Tillis, who is not seeking reelection, is no longer in office, even after Bartiromo corrected him.

Bartiromo: "And you think Kevin Warsh can get confirmed? Do you think Thom Tillis is going to give you a vote?"

Trump: "That's why Thom Tillis is no longer a senator."

Bartiromo: "OK. On the on the AI thing —"

Trump: "You know Thom Tillis is no longer a senator, right? He quit."

Bartiromo: "Well, he's on his way out."

Trump: "Well, no, he quit, but he quit."

Bartiromo is right.

Cut and dried, right? Why are we talking about this?

We're glad you asked! PolitiFact tampered with the evidence to try to put a respectable face on yet another case of "gotcha!" journalism.

Search the fact check over and you won't find any evidence Trump was aware Tillis retains his ability to vote on the Warsh nomination, except with the response from the White House (and wouldn't we like to see the wording of PolitiFact's outreach):

The White House referred PolitiFact to April 15 remarks by Treasury Secretary Scott Bessent, who said, "Well, I'm sure Sen. Tillis wants to do the best thing for the Federal Reserve, for the American people. He has publicly said that Kevin Warsh is a great candidate for the chair. So let's get to the hearings and see where we are then."

PolitiFact encourages its audience to believe Trump thinks that because Tillis quit the Senate he'll no longer figure in on the Warsh nomination.

But that's completely at odds with the un-elided transcript (re-formatting to better fit our page design, PolitiFact omissions in bold):

Maria Bartiromo: "And you think Kevin Warsh can get confirmed? Do you think Thom Tillis is going to give you a vote?

Donald Trump: "Well, we're going to have to find out. He might not, but that's why Thom Tillis is no longer a Senator.

Maria Bartiromo: "OK. On the on the AI thing --"

Donald Trump: "You know Thom Tillis is no longer a Senator, right? He quit."

Maria Bartiromo: "Well, he's on his way out."

Donald Trump "Well, no, he quit, but he quit. And I think he doesn't want the legacy of stopping a great -- a person who could be great. I think Kevin Warsh is going to be great. He doesn't want the legacy of having an incompetent guy stay there for longer than is necessary. I know Thom Tillis, he's a good man, I don't think he's going to hurt."

Mind you, we're using the transcript PolitiFact claimed to have used as its source. PolitiFact just cut away the parts that show Trump well aware that Tillis would have a vote on the Warsh nomination.

Given Trump's awareness of Tillis' continued role in that impending vote, how does PolitiFact's conclusion follow, that Trump and Tillis supposedly disagree on whether Tillis continues to serve in the Senate?

Spoiler: It doesn't follow. We're left to reconcile, if we can, Trump's statement that Tillis is no longer in the senate with Trump's acknowledgement that Tillis continues to serve in the senate. We figure it's most likely Trump was making the point that Tillis opposition to Trump is unpopular with his constituents, helping lead to the decision not to run for re-election.

Did Trump express that idea clearly and without ambiguity? Nah. And PolitiFact will likely never glom onto the fact that ambiguities and fuzzy language are poor environments for fact-checking. Or never will at least as long as Trump's in office and they can smell another "gotcha!" in the offing.

Afters:

We didn't emphasize the point above, but it's proper in professional journalism to note the omission of significant parts of a quotation with the use of an ellipsis or ellipses (...). PolitiFact may have fallen into the trap of thinking their method was okay because it's improper to begin a quotation with an ellipsis. But that's in the context of a simple quotation. The rule doesn't give journalists a license to lop off the entire sentences from a transcript as the proper observation of that rule. The journalist can follow the rule by taking just one quotation out of the transcript and starting the sentence without an ellipsis. The reader can infer that the quotation isn't necessarily complete and hopefully the journalist can be trusted not to omit anything essential. But a transcript isn't really a transcript when sentences are omitted without cuing the reader.

PolitiFact led its readers to falsely believe it was providing a real, full transcript of the relevant exchange.

Update April 16, 2026: Belatedly added the URL for the flawed PolitiFact fact check.

HoyStory: Trump's mostly right about Pope Leo IVX (PolitiFact wrong again)

 Fellow PolitiFact critic Matthew Hoy saved us some time by hitting a target we were fixing to home in on.

Hoy presses the point that Trump's point should be easy to understand. When the Pope won't endorse the steps need to keep Iran from obtaining a nuclear weapon it's tantamount to the Pope giving that eventuality his tacit okay.

Short version:

A shorter version of why @PolitiFact's "Pants on Fire" rating for Trump should, at the very least, be "Mostly True." With bonus Swalwell analogy. pic.twitter.com/Fb55U9MPLt

— Matthew Hoy (@hoystory) April 14, 2026

Long version:

PolitiFact hands Trump a Pants on Fire that really should be a Mostly True.https://t.co/UCL5qsovDG

— Matthew Hoy (@hoystory) April 14, 2026

Also, buy Matthew's book about fact checker perfidy, featuring PolitiFact and a fair number of examples explored here at PolitiFact Bias.

Elizabeth Warren, PolitiFact, and the postmodern approach to the modern era

PolitiFact Bias completed a research project last year looking at PolitiFact's "True" and "Mostly True" ratings. And though it wasn't the original aim of the study we found that "True" and "Mostly True" count as dying breeds in the "Truth-O-Meter" universe. For example, in 2025 PolitiFact meted out only 10 "Truth-O-Meter" ratings to partisan political figures. There were over 60 in PolitiFact's first year.

Our experience from that study led us to track new "True" ratings a bit more closely this year. Sen. Elizabeth Warren (D-Mass.) cracked the ice with the initial "True" rating of 2026.

I thought the term "modern era" was the key to this fact check. I asked Grok AI to evaluate the fact check. Grok's been tweaked at some point in the past year to elevate its deference to "trusted sources" like PolitiFact, so Grok produced a very trusting evaluation, finding the fact check perfectly solid. I followed up asking whether Axios had defined "modern era" as PolitiFact claimed.

A March 2 Axios article with an opening that nearly matches Warren’s statement: "No president in the modern era has ordered more military strikes against as many different countries as Donald Trump." The Axios story defines "modern era" as presidents post-9/11: George W. Bush, Barack Obama, Trump and Joe Biden.

Short answer? No.

The Axios story mentions "modern era" in the lead, which was pretty much the same thing Warren said. And after that doesn't mention the term again. Axios backs the claim in the lead with 21st century examples Bush, Obama, Trump and Biden. From there, PolitiFact infers that "modern era" was defined in terms of 21st century presidents.

But is that how the average reader understands "modern era"?

From a contemporary perspective, “the modern era” usually means the present day and the recent past, but historians often use it more specifically for the period from about the 1500s to 1945, with the time after 1945 often called the contemporary period.

I put the question to Perplexity AI. What is the "modern era" from a contemporary perspective?

Obviously the term carries some ambiguity. How many times have we seen PolitiFact appeal to a specialized definition backed by expert testimony to downgrade a similarly ambiguous claim uttered by a Republican? It's routine by now.

Well, the "modern era" as understood by historians runs up through 1945, encompassing the whole of the Franklin D. Roosevelt administration. The number of nations receiving military strikes from that administration would outnumber those from the Trump administration.

Not only does PolitiFact use specialized definitions to downgrade "Truth-O-Meter" ratings, it also appeals to the impression a claim leaves on the audience irrespective of whether that understanding finds support in the context of a claim. Yet Sen. Warren's claim failed to raise that warning flag for PolitiFact.

PolitiFact even introduced its own complementary claim about Mr. Trump:

After reviewing these sources and additional reporting, we found that Trump has authorized the highest number of strikes and targeted the most countries compared with other 21st century presidents.

That's nice, but Warren didn't mention the 21st century and neither did Axios. Axios simply limited its comparisons to other 21st century presidents. How was Warren's audience supposed to know that?

Were PolitiFact a neutral and competent fact checker, it would find, at minimum, that Warren left out the 21st century context when that context determines the truth value of her claim. Sounds a bit like the way PolitiFact defines "Half True" or even "Mostly False" on its subjective "Truth-O-Meter":

HALF TRUE – The statement is partially accurate but leaves out important details or takes things out of context.

MOSTLY FALSE – The statement contains an element of truth but ignores critical facts that would give a different impression.

The important detail is the operative definition of "modern era." And "modern era" is the ignored critical fact that would give a different impression (FDR, not DJT).

What might have been: "Experts told us the "modern era" runs from the 1500s through 1945. That span includes president Franklin D. Roosevelt, who launched attacks on more countries than has Donald Trump. We rate Warren's statement "Mostly False."

Instead, PolitiFact drew a target around Warren's point with her point at the center and declared it a bullseye. That's the Texas sharpshooter fallacy.

Ten years of PoF bias 2016-2025

Before we move on to the PoF bias in 2026, we'll mark down the cumulative totals since 2007 and for just the past 10 years. Note that the PoF Bias trends higher over the past 10 years, 2.24 compared to 1.35 over the full history.

The "selection bias" number likewise increases over the past 10 years, measuring 3.95 compared to 2.81 overall. Remember, when we hedge for our own bias in the data it underplays the degree of bias but at the cost of overplaying the "selection bias" number.

We put "selection bias" inside scare quotes because we don't really know exactly what causes the disparity. It could be selection bias, fact checker bias (ideological) or maybe Republicans just lie more and rightly get fact-checked as a result.

Of course we have dozens of examples of biased ratings, so that evidence points toward the first two explanations more than the third explanation.

Next year when the 10-year frame shifts, the Democratic Party will lose out on 13 "Pants on Fire" ratings, so we're anticipating that the 10-year totals 2016-2025 will contrast even more sharply with the cumulative totals. So far in 2026 it looks like the trends continue: lower raw numbers for the GOP and the Democrats, with the Democrats getting ever closer to disappearing from the data. After three full months of 2026 the Democrats have avoided "Pants on Fire" and "False" ratings. In 2026 the Democrats had two "False" ratings heading into July before a late-season flurry rocketed them up to a grand total of six for the year. Their previous low for one year was 2007, PolitiFact's partial first year, when they registered nine. And five of those were "Pants on Fire" ratings!

Did Democrats change that much in 20 years?

PolitiFact's 'Pants on Fire' bias in 2025 WOWZA

Since 2011 PolitiFact Bias has tracked PolitiFact's "False" and "Pants on Fire" ratings, using the rate at which false claims are subjectively deemed "Pants on Fire" to measure ideological bias.

Last year proved a landmark year for the bias study. That's despite the fact that PolitiFact stymied any attempt to calculate a meaningful "PoF bias number" for 2025. That number represents the likelihood PolitiFact will issue a "Pants on Fire" rating to a false statement from a Republican compared to a Democrat. A PoF bias number of 1.5, for example, says a GOP falsehood is 50 percent more likely to receive the "Pants on Fire" rating than a Democrat. It's a simple matter of finding the total falsehoods for each party, then dividing the number of "Pants on Fire" ratings by the total number of falsehoods PolitiFact counted.

The PoF bias number occurs completely independently from the sheer number of falsehoods. But if a party receives no "Pants on Fire" ratings at all in a given year then we can't compare the percentages from each party to see which is more likely. In 2025 PolitiFact issued zero "Pants on Fire" ratings to Democrats.

Important note: After starting his post and creating the graphs, I discovered an error where I had undercounted by one the Democratic "Pants on Fire" by one a few years back. One PoF makes a significant difference in the numbers for the Democrats because they receive so few of them compared to the GOP.

We elected to publish the flawed crafts with this disclaimer. We have the error ironed out, so it's not likely to affect future graphs.

Interestingly, both parties have received fewer "False" and "Pants on Fire" ratings in recent years. That's likely explained by the distorting effect of paid social media fact-checking partnerships, though Facebook's termination of the partnership program in 2025 did not appear to upset the trend.

Despite no PoF bias number for 2025, the 2025 data helps reinforce the past data trends. The percentage of "Pants on Fire" ratings for the GOP was highly consistent with the average since 2007 (25.35 percent compared to 25.99 percent) That percentage over time remains consistent at a level that counts as downright creepy.

If we toss the outliers (Dems 2007, GOP 2023), the GOP slotted right in between 20 and 31 consistently. The Democratic Party side showed no such consistency, ranging from 0 to 30 and showing a relatively consistent downward trend.

What Are We Measuring?

Our research results drive us back toward the basic question: What are we measuring? Is there a tendency to fib that's consistent within the Republican Party, or is there tendency from PolitiFact, driven by bias, that creates the consistency in the numbers?

Given the apparent subjectivity of the "Pants on Fire" rating, the answer seems obvious. Subjective rating scales serve very poorly for measuring objective things. But they serve very well for measuring subjective things.

We at PolitiFact Bias assert that the numbers show us what PolitiFact subjectively believes about Republicans.

And what of the comparatively all-over-the-place numbers for Democrats? What do those show?

This one's tougher to pin down, but we have a few suggestions based on our years of observation.

New fact checkers start out trying to impress on readers their neutrality. We saw this repeatedly with PolitiFact's new state franchises. They tended to start out tough on Democrats and ease up very soon after. With their neutrality established for the readers they felt free to tell the truth (as they saw it) that the GOP lies more. Bryan's (my) Zebra Fact Check did very much the same thing. The first year I put more effort into fact-checking Republicans to give skeptical readers a measure of the website's fairness. Later I fact checked as I pleased.

The years 2010 through 2015 or so show heavy input from key state franchises like PolitiFact Texas and PolitiFact Wisconsin. The states were doing much of PolitiFact's fact-checking in those years, and (not counting PolitiFact Florida) tended to show less bias toward Republicans. 

I thought PolitiFact National emptied the states of their editorial control. But Iforgot PolitiFact's Louis Jacobson posted on X about his meeting with PolitiFact New York's editorial board, suggesting each state maintains a state version of the ruling editorial board.

We think the dwindling number of falsehoods attributed to Democrats simply shows PolitiFact has grown less interested in examining falsehoods from Democrats. There's no reasonable evidence Democrats simply stopped making false statements. Democrats in 2024 weren't held to account for finding President Biden sharp as a tack, for example.