Post from Truth Social

A CNN polling segment, posted against unfavorable midterm numbers

At about 2:06 p.m. Eastern (18:06 UTC, most likely in Washington, the day after the September 30 funding deadline), between a run of judicial-nomination announcements, the account posted an uncaptioned clip of what appears to be CNN's Harry Enten at his data wall. The analyst sets up "really good Democratic polls coming up from Ohio, Michigan, Iowa," then argues those leads are smaller than past Republican overperformance of September polling. His figures: in 2024 Republicans beat their September polls by "nine points" in Iowa, "six points" in Ohio and "five points" in Michigan, and in 2018 by six, six and eight. He closes on a conditional: if the pattern repeats, the races "could become very, very interesting."

The source matters. CNN is one of the three outlets whose White House access was cut and then ordered restored by Judge Timothy Kelly on September 25. Using that network's analyst as the messenger makes the argument look conceded by an adversary rather than asserted by the interested party, the "even CNN admits" move that recurs whenever he reposts Enten. The post adds no words, so the hostile-source framing does the persuasive work without him having to claim anything.

What the clip does on his behalf

It gives him a method for disbelieving bad news. The segment's own framing, "fool me once, shame on you. Fool me four times," treats current Democratic leads as a repeat of a known error, not as information. In the week of reported record-low approval and Democratic leads in three Midwestern states with open or competitive Senate and governor's races, the response is not to engage the numbers but to circulate a reason they will be wrong. That is rationalization carried out through selection: the threat is acknowledged (the polls exist) and recast as illusory.

It also drops the speaker's hedge. Enten poses a question, "could it happen again?", and ends on "could become very, very interesting." Posted bare, to an audience primed for reassurance, the conditional reads as a forecast. Nothing in the post restores the uncertainty the analyst kept.

And it flatters him by name. The segment's pivot, "Donald Trump outperformed in these three states by significant margins" across 2016, 2020 and 2024, casts him as the cause of the polling misses. When the analyst's imagined skeptic objects that "Donald Trump was on the ballot then," the answer is "Donald Trump's first midterm election," so even the evidence meant to show the effect without him is labeled with his name.

How much the borrowed argument holds

The empirical core is broadly sound: polls did understate Republicans in these states in 2016, 2020 and 2024, and in Ohio's 2018 races. The limit sits inside the clip's own data. In 2018 Michigan, the state where Enten reports the largest Republican overperformance (eight points), Democrats still won both the Senate and governor's races, because their September leads were larger than the shift. Beating polls and winning are different outcomes, and the segment's closing hedge concedes as much. Since these are the broadcaster's figures, they are discussed here and not entered as claims of the subject's.

Register: defensive reassurance, no combat

The post names no enemy and calls no poll "fake" or "rigged," which is milder than his usual handling of bad polling; the restraint belongs to the clip, since he wrote nothing. The posture is defended grandiosity: the protagonist is the candidate the instruments always undercount, and the supporters are a hidden majority the pollsters miss. The redemption arc (trailing in September, vindicated in November) is the same one the earlier same-day text post runs on Iran: "it would take 4-6 weeks... and I did it in one night!" Both reframe a metric that could read as failure into proof of being underestimated. There is no mobilizing language, no target and no call to act, so the post carries no violence signal; its effect on followers is to make discouraging polls easier to dismiss.

Authorship evidence is thin. The early-afternoon timing and a trimmed broadcast clip fit staff curation, while his long habit of personally circulating Enten segments that flatter him fits his own choosing. Either way, the question is who picked the clip, since no one on his side wrote any words.