An Iowa judge dismissed the president's lawsuit against the pollster Ann Selzer and the Des Moines Register on Wednesday, ending a case that had been pending since December 2024 over a survey published days before the presidential election. Judge Scott J. Beattie of the Iowa District Court threw out all three counts and did it with prejudice, which means the same claim cannot be refiled.

The case is interesting less for who won than for how completely the plaintiff lost. The court did not have to decide whether Selzer's poll was conducted in good faith, because it ruled that even taking the allegations as true, the statute the suit relied on did not reach the conduct described. When a complaint fails on the plaintiff's own version of the facts, an appellate court has very little room to work with.

A statute written for purchases got pointed at a prediction

The suit was brought under the Iowa Consumer Fraud Act. That statute exists to protect people from being deceived in a consumer transaction: something is sold, something is misrepresented, someone pays money because of the misrepresentation. The court found that structure decisive. There was no transaction here, and no representation made to induce any of the plaintiffs to act on the strength of it.

Selzer's final Iowa poll, published in early November 2024, showed Kamala Harris leading Trump 47 percent to 44 percent in a state Trump went on to win comfortably. The plaintiffs, who also included Representative Mariannette Miller-Meeks and former state senator Brad Zaun, argued that the survey was skewed and that publishing it amounted to consumer deception.

The gap between those two things is the whole case. A wrong poll published to a general readership does not sell anyone anything. A consumer fraud statute assumes a marketplace relationship: a seller, a buyer, and a misstatement that moved the buyer. None of that exists when a newspaper prints a projection about an election that has not happened yet, which is why the judge concluded that the suit asked Iowa law to do something the legislature never wrote it to do.

The ruling did not depend on whether the poll was honest

Beattie wrote that the poll and the reporting around it are fully protected political speech, and that the suit sought to turn "speech that enjoys the highest category of First Amendment protection" into a liability. He concluded that the protection is too great and the reach is too far.

That is the first ground. The second is more interesting, because the court went out of its way to rule on the fraud claims separately, assuming the plaintiffs' allegations that the poll was intentionally skewed and fabricated were true. Even under that assumption, all three counts failed.

Judges rarely spend effort on a hypothetical when a case can be decided on solid ground. Beattie's decision to resolve the fraud question anyway reads as a signal to the appellate court: there is no version of this complaint that survives, so reversing on the First Amendment question would not revive the case.

A forecast is conjecture, and conjecture is hard to call false

The doctrinal point underneath the fraud ruling is about what a poll is. A statement of fact can be proven false. A prediction about a future event cannot be, at least not in the way a fraud claim requires. When Selzer published her survey, she was offering an estimate of an electorate that had not yet voted.

That distinction is not a technicality. It is the reason opinion and forecasting have been treated differently from factual assertions going back to the earliest American defamation law. If a wrong prediction were actionable, every forecaster in the country would be exposed to suit by anyone disappointed in the outcome, and the safe move would be to predict nothing.

The suit also carried a filing that has become familiar in cases involving the president: a legal action brought over coverage or analysis he objected to, in which the theory of harm grows out of the objection itself. The court's response was procedural but firm. Whatever the poll did, it did not deceive anyone into buying anything, and it did not state a fact that could be shown to be untrue.

A prediction is not a product

The transactional requirement that killed the suit is not a technicality the legislature forgot to fix. It is the boundary between two bodies of law that do different jobs. Consumer protection law regulates commerce, and its remedies assume a marketplace: refunds, damages, injunctions against a seller. Defamation law regulates false statements of fact about a person, and it comes with a set of constitutional guardrails built over sixty years of First Amendment cases, including the actual malice standard for public figures.

When a plaintiff brings a consumer fraud claim over speech, the claim is asking for defamation-style relief under a statute that has none of the defamation guardrails. That is the move the court rejected. The judge noted that poll predictions are conjecture about future events rather than provably false statements of fact, which means they would strain even the defamation standard, and the consumer fraud statute has lower thresholds in some respects because it was written for sellers of goods.

The practical value of the ruling is that it closes a route. A public figure who dislikes coverage can file a defamation suit and expects a hard fight over malice and truth. Filing under a consumer protection statute was an attempt to skip that fight. In Iowa, at least at the trial level, it does not work.

What the plaintiffs would have had to show

The joined plaintiffs included Representative Mariannette Miller-Meeks and former state senator Brad Zaun, both of whom won their races in the same election the poll predicted wrong. Their theory was that the survey deceived Iowa voters, and that they were harmed by it.

That theory has an intuitive pull and a doctrinal problem. To win under the statute, a plaintiff has to show a representation made in connection with a consumer transaction that induced reliance and caused loss. None of the plaintiffs bought anything from Selzer, relied on her poll to their detriment in a commercial sense, or could point to a loss that a court could measure. The deception, if there was one, was aimed at a general audience, and Iowa law does not give a general audience a cause of action over a prediction.

What the plaintiffs would have needed is a different theory entirely, one built on a false statement of fact about them, published with the required state of mind, causing reputational or pecuniary harm. That is a defamation case, and it is a harder one to bring, which is the point of having a harder standard.

With prejudice, and an appeal that has to argue around both grounds

A spokesman for the president's legal team said the case will be appealed to the Iowa Supreme Court. That appeal will have to contend with both grounds at once. If the Iowa appellate court accepts the First Amendment holding, the case ends. If it disagrees, the fraud analysis remains, and the same transactional defects apply.

The dismissal carries a further procedural consequence. A dismissal with prejudice means the complaint cannot be amended and refiled; the plaintiff's only path is appellate. That is a strong signal about how the court read the complaint, because judges usually grant leave to amend when a defect looks fixable. Beattie treated these defects as structural, which is why nothing was left to fix on remand.

The appeal also has a narrower runway than a typical case. The facts are largely undisputed, since the court assumed the plaintiffs' version for the fraud counts. That leaves questions of law, which appellate courts review fresh but which also give them less room to send the case back for more development. If the Iowa Supreme Court agrees that the statute does not reach this conduct, there is no factual dispute left to try.

A plaintiff in that position has one more option, which is to bring a defamation claim instead. Nothing in Wednesday's ruling forecloses it. It would have to contend with the actual malice standard that applies to public figures, which requires knowledge of falsity or reckless disregard for the truth, and with the difficulty of proving that a poll was false when the outcome it predicted had not yet happened.

The dismissal adds Iowa to a pattern. Suits involving the BBC, the New York Times and the Wall Street Journal have drawn similar resistance, and the same day this ruling came down, a federal judge in Washington heard arguments in a separate case over the White House's decision to bar three other outlets from the building. What the Iowa decision adds to that pattern is a reminder that state consumer protection law is a poor fit for disputes about political speech, because the statutes are drafted around commerce.

Selzer, who was represented without charge by the Foundation for Individual Rights and Expression, said the ruling reaffirms the freedom to pursue and publish the truth without political interference, and added that "polling is a social science, not clairvoyance." That is a fair description of the work. It is also, in effect, what the judge said: a poll is a claim about what is likely, offered into a public argument, and the law of consumer fraud was never the right instrument for testing it.

The appeal is a longer fight. For now, the case is over, and the precedent it leaves is one that will matter to anyone who publishes a forecast that turns out to be wrong: being wrong is not the same as being liable, and the statute a plaintiff reaches for has to cover the conduct.

Primary sources

  1. CNN, Judge dismisses Trump's lawsuit against Des Moines Register and pollster J. Ann Selzer, for the terms of the dismissal and the with-prejudice ruling.
  2. Courthouse News Service, Trump suit against Des Moines Register dismissed by Iowa judge, for the court's reasoning on the consumer fraud counts.
  3. Editor and Publisher, Iowa judge dismisses Trump's lawsuit against pollster J. Ann Selzer, for the pollster's statement and representation.
  4. U.S. News, Iowa judge dismisses Trump's lawsuit against Des Moines Register and pollster, for the procedural history and the announced appeal.