At 9:30 in the morning Greenwich time on July 31, the Megyn Kelly AM Update described a poll finding this way: "A new CNN poll finding roughly 1/3 of Democrats and Democratic-leaning independents now describe themselves as Democratic Socialists" [1]. Six and a half hours later, Michael Knowles opened his show with the same finding: "one in three Democrats now identifies as a socialist" [2].
Three things changed in between. "Democratic-leaning independents" is gone, which shrinks the population being described down to people inside one party. "Democratic Socialists," a capitalized label attached to a specific political tendency, became "a socialist," a general one. "Describe themselves as," which is a statement about how people present themselves, became "identifies as," which reads as a settled fact about a person. The AM Update's own tease had already supplied the verb the segment wanted: "New polling exposing socialism's growing grip on the Democrat Party" [1].
The segment underneath that framing is a real story and it earns its airtime. The AM Update played a Fox News Sunday true/false round in which the national co-chair of the Democratic Socialists of America affirmed abolishing the Senate, replacing the presidency and the Supreme Court, abolishing ICE, abolishing borders with an amnesty, defunding the Pentagon, abolishing prisons "in large part," and "government or public ownership of most large corporations" [1]. That is a maximalist program stated plainly by someone with standing to state it. If a third of a party's coalition were signing on to that list, it would be among the more consequential facts in American politics, and a show that leads with it is not manufacturing a topic. Whether those answers describe official DSA policy or one officer's own views is not resolved in the segment.
What this newsroom cannot check is the poll itself. CNN.com does not serve our fetcher, so the sample, the field dates, the margin of error, the base that the one third is taken from and the exact wording of the question are all unverifiable here, and no independent copy of the topline was located. The only text available to us is what the two shows said the poll found, which is exactly why the distance between the two descriptions is the story. Our July 31 piece on the poll's framing handled the framing question and needs no correction; this is the narrower matter of what happened to the sentence over one afternoon.
The compression matters because of the size of the thing being compressed. Two survey houses, working in different modes over different windows, put the Democratic coalition in the same place. Gallup's full-year 2025 aggregate of 13,454 US adults, margin of error one percentage point, found 47% identifying as Democrats or Democratic-leaning independents [3]. Pew's NPORS survey of 5,511 US adults, fielded January 30 to June 18, 2026, found 47% identifying with or leaning toward the Democratic Party [4].
Census Vintage 2024 estimates put the United States at 340,110,988 people, of whom 73,132,720 are under 18, leaving 266,978,268 adults [5]. Forty-seven percent of that is 125,479,786. A third of that is 41,826,595, call it 42 million. These are adults rather than voters, and the distinction is not decoration: both surveys measured adults, and the number carries that definition with it wherever it travels.
The organization those adults are being attached to is very much smaller. DSA's own monthly figures put dues-paying membership at 92,912 in December 2025, up from 50,713 in October 2024 [6]. That growth is real and it is fast, close to a doubling in fourteen months, which is the part of the story the shows have right. DSA has cited a higher figure for mid-2026 and we could not source it, because dsausa.org returns a 403 to our fetcher, so 92,912 is the most recent count we can stand behind.
Forty-two million against 92,912 is a ratio of roughly 450 to 1. That is not an argument that the poll is wrong. A gap of that size between people who accept a label and people who pay dues to an organization is normal, expected and present in every political identity anyone has ever tried to measure. Far more Americans call themselves environmentalists than belong to the Sierra Club, and nobody reads that as evidence the environmentalists are imaginary. The poll, as described on air, measures the size of a sympathetic identity. DSA's roster measures the size of an organized institution. Both numbers can be accurate at the same moment, and the useful thing is knowing which one a given sentence is talking about.
The compression has a cost with a name attached. Abdul El-Sayed is four days out from the August 4 Michigan primary, and the same AM Update that carried the poll line also carried him, at Monday's debate, saying this: "When it comes to the question of socialism, I'm not a socialist. I'm just a capitalist who understands how capitalism works. The biggest threat to capitalism was never the idea of government regulation, it was always monopoly" [1]. A viewer who caught the tease and left before minute thirteen heard a party described as one third socialist and never heard the candidate decline the label.
Dropping "and Democratic-leaning independents" is the load-bearing edit. It converts a self-description offered by part of a coalition, including people who are not in the party at all, into an attribute of the party itself. Six and a half hours is all that took.
