It started with Cat Turd.

In the turbulent media ecosystem that swirls around former president Donald Trump, a pseudonymous Twitter user with that unpleasant appellation is something of a star. (He tweets as “catturd2”; it’s unclear what fate befell “catturd1.”) His combination of anti-establishment hostility and crowd-specific witticisms earned him multiple retweets from Trump back when that was possible and allowed him to carve out a small pocket of influence in MAGA world.

So, on an episode of his podcast last month, he welcomed Rep. Marjorie Taylor Greene (R-Ga.), who shared with the world important new research on Republican responses to voter fraud.

But first, a little housekeeping.

“I don’t even know if I’m saying it right? Cat Turd?” she asked. “Is that right?”

It was, he assured her.

After about 30 minutes of precisely the sort of conversation that you’d expect, Mr. Turd asked Greene how, in future elections, Republicans could “beat the machines” — that is, voting machines that he and Greene incorrectly think somehow illegally influenced the 2020 election. Greene then explained a recent survey that showed exactly how fraught the Republican base’s concerns about election fraud were.

That research eventually trickled up to the New York Times, which reported on it Sunday.

Greene “has told colleagues that she was surprised by a recent survey of Republican voters in her district, according to one person who spoke with her about it,” Jeremy Peters reported. “The informal internal survey found that roughly 10 percent of Republican voters expressed a serious lack of confidence in the security of Georgia’s elections, with 4 percent saying they did not think elections were secure in Georgia and would not vote in future elections, and 6 percent saying they were unsure whether their votes would be counted.”

“The possibility that roughly 10 percent of Republicans would sit out any election,” he continued, “or question whether their votes would be accurately counted — even in a solidly red district like the one held by Ms. Taylor Greene — was something Republican strategists said they found alarming.”

Statistically useful polling is expensive, and members of Congress in secure seats don’t usually poll their constituents in off-years on random policy issues. So perhaps you noticed a red flag in Peters’s description: “informal.” To the “very frisky cat” (as his producer described him), Greene explained how this informal survey worked.

“I did a tele-town hall in my district just last night. And we took a poll during the town hall, and we asked people how they felt about election integrity,” Greene explained. “And we asked them: Press number one if you think your vote counts and you feel our elections are secure. Then we said, press two if you feel your vote does not count and you think your election is not secure. We said press three if you will not vote again. And we said, press four if you’re unsure.”

“The results from the poll are something that are really terrifying to me,” she continued. “I think it was like 18 percent said that, yeah, I trust elections are secure. They were fine with it. There was 73 percent — 73 percent said, no, I don’t trust the elections. I think my vote is stolen. Then the one that’s got me the most, and I’ll tell you why, was number three was people saying that they will not vote. And that was that 4 percent. Four percent sounds like a small amount but when you — and these were Republican voters I had on the phone. These are people that actually vote in primaries and general elections. So when they’re saying, 4 percent are saying, I’m not going to vote, like, flat out, not going to vote, that equates to approximately around close to 11,000 Republican voters in my district.”

She felt confident that this was accurate, she added, because “I talk to regular people all the time.”

It’s not often that someone presents a concise distillation of why their own argument is not worth taking seriously, so let’s take advantage of Greene’s doing so. That she assumes the people with whom she talks are representative of America’s “regular people” is why the survey is somewhat less than useless. People who speak to members of Congress are not regular Americans, particularly when that member of Congress represents a rural Southern district. Nor are those who attended in Greene’s town hall a useful gauge of sentiment; they chose to participate in a telephone conversation hosted by a right-wing politician with whom, it’s safe to say, most agreed. It should be comforting to the GOP, not alarming, that only 4 percent of a group so enamored of Greene that they would spend their evenings on phone calls with her and then weigh in on her question have thrown up their hands about voting in elections.

Yet it nonetheless raised enough concern for Republicans that it eventually reached Peters, who highlighted that 4 percent figure and the 6 percent who had pressed four for “unsure.” He tweaked his report since it was published, adding the qualifier “or question whether their votes would be accurately counted” and adding the critical word “informal,” both of which soften the findings. But that it is still included at all is probably a mistake; there is essentially no value to this finding. A decent rule of thumb is that media outlets should have a higher standard for survey data than does a podcast hosted by someone named after animal excrement.

The utility in Greene’s survey is not the numbers. It is to serve as a reminder that Trump brought a pool of voters to national elections who may not stick around. His own threats that Republicans will sit out 2022 or 2024 if “fraud” isn’t addressed leverage the fact that he has not shown any ability to turn out his most hardcore base for elections in which he is not on the ballot, the central question for Trump’s influence in the party moving forward. The Greene-Trump universe of voters is probably disproportionately unlikely to have voted in 2012 or 2014 and probably disproportionately likely not to be inspired to vote in 2022 or 2024 anyway, false fraud claims notwithstanding.

This is simply an educated guess, however. In order to validate this assumption with all of the statistical accuracy of Greene’s town hall survey, please offer your thoughts in the comments.