Rebutting a False Claim Without Spreading It

Every time you repeat a false claim to knock it down, you also repeat the false claim. Repetition makes things feel familiar, and familiar things feel true. So the way you rebut matters as much as whether you are right.

The truth sandwich

George Lakoff, a linguist at the University of California, Berkeley, described this method in June 2018 on CNN's Reliable Sources podcast with Brian Stelter. Accounts differ on who named it: Wikipedia credits Stelter, PBS Standards credits Lakoff. The layers:

1. Truth. Open with what is accurate. This sets the frame.
2. The false claim, briefly. Name it once, plainly, without repeating its catchy wording.
3. Truth again. Return to the facts, with the receipt, so the last thing readers see is the accurate version.

The "after" version leads with the actual vote, mentions the false claim once in neutral words, and ends with the facts and a date. Readers come away knowing what really happened.

The research on the truth sandwich is mixed. Studies in 2023 found sandwiches about as effective as other correction formats, and a 2025 study found they did not beat standard corrections. The 2023 studies found any correction beat none at all. Use the sandwich because it costs nothing and keeps your post about the facts.

Verify before you rebut

Be sure the claim is false before you call it false. And read the exact wording of the claim before you answer it. Checklist before posting a rebuttal:

• Find the original. Who first posted it, and when?
• Find the primary record: minutes, agenda, video, the document itself.
• Does the record contradict the claim, or just not mention it? Those are different. "The minutes show no such vote" is stronger than "that's not true."
• Is any part of the claim accurate? Say so. If the board did discuss a book challenge in a separate meeting, your rebuttal must say that, or it becomes the next false claim.
• Cite. Link the minutes, attach the page, give the date.

The Source Checker at jeffapierson.com/journaltools/sourcechecker/ can help when you are weighing where a claim came from. The verification lesson in the journalism module goes deeper.

Screenshot, quote-post or nothing at all

Pick how to handle the original false post.

Size matters here. If a public official with 20,000 followers posts the claim, the claim is already out; showing it and correcting it serves readers. If an account with 40 followers posts it, your quote-post to 3,000 followers is the biggest thing that ever happened to that claim. Post the facts on their own and leave it there.

Labeling AI images

If you use an AI image (from ImageStudio or anything else), say so on the image and in the caption. A small corner tag works: "AI illustration." In the caption: "Illustration made with AI; not a photo of the actual meeting." The module on images has the full rules.

The platforms require this for realistic content. As of October 2026, YouTube requires disclosure when realistic content was made or meaningfully altered with AI, like a real person appearing to say something they didn't or a realistic scene that never happened. It exempts things like color fixes, captions and AI help writing a script. TikTok's community guidelines say synthetic or manipulated media showing realistic scenes must be clearly disclosed, with a sticker or caption such as "synthetic" or "altered." Meta said in February 2024 it would require users to disclose photorealistic AI video and realistic AI audio, and may penalize accounts that don't.

Go further than the rules. Label every AI image, even an obvious cartoon. When you are rebutting false claims, your credibility is the whole game. One unlabeled AI image found by an opponent will be screenshot and used against everything else you publish.

And never use an AI image to illustrate the thing you are rebutting. A fake photo of empty library shelves, even labeled, keeps the false picture in people's heads.

Correcting yourself in public

You will get something wrong. A vote count, a date, a name spelled two ways. Fix it the same day, where people will see it.

• Edit if the platform allows, and add "Corrected:" with what changed. Editing rules differ a lot by platform and account type, and some platforms do not allow edits at all, so check yours before you need it.
• Post the correction as its own post when the mistake mattered or traveled. "Correction: The library vote was 3 to 2, not 4 to 1 as I posted Tuesday. The Elm Street and Ridgeview branches are affected, as reported."
• Reply to the original with the correction so anyone who finds the old post sees it.
• Fix the article, with a correction note at the bottom giving the date and what changed.
• Don't delete without saying so. Screenshots of the mistake will outlive the post. A silent deletion looks like you were hiding it.

Readers who see you correct yourself plainly trust you more next time. Writers who never admit errors lose the readers who noticed, and someone always notices.