Policies
Corrections
What gets corrected, how it gets corrected, and what has been corrected so far.
How a correction works here
If something published on this site is wrong, the page gets fixed and the fix is described at the bottom of that page, with the date. The original wording stays in the note, because a correction that quietly replaces the old sentence is indistinguishable from never having been wrong.
Three kinds of change are treated differently, and the difference matters more on a site that writes about research than on one that only writes about dinner.
- A factual correction is a number, a name, a year, a journal or a finding that was stated incorrectly. It gets a dated note on the page and a line on this page.
- A clarification is a sentence that was accurate but read as more than it said. It gets a dated note on the page, and it does not get a line here unless the meaning changed.
- A typographical fix is a spelling, a broken link or a wrong quantity in an ingredient list. It gets fixed. It gets no note, because a site that announces every typo is announcing nothing.
A study citation that turns out to point at the wrong paper is a factual correction, not a typo, even when the error is a single digit in an identifier. That is the whole reason identifiers are written out in the text.
Corrections issued
Four, all issued on the same day and all out of the same review. In November 2026 we went back through the four longer pieces that discuss research and anchored every statement in them to the published record, identifier by identifier. Doing that turned up one error we had made repeatedly, in our own words rather than in anybody's numbers: we had described the research on these ingredients as mostly laboratory work on isolated compounds. It is not. There are dozens of human trials, at doses somebody weighed out, and some of them found things.
It changed no verdict. Every claim those pages assess is a claim about a home preparation that nobody has tested, and that is still true. What changed is that the pages now say what the trials on the ingredients actually found — including where they found nothing, where the funding came from a seller, and where two reviews of the same literature contradict each other.
- November 10, 2026 — Claims Review: The Honey, Garlic and Cinnamon Drink. Claim six described the research on the individual ingredients as “most of it done to compounds in glassware at strengths a person drinking a mug of this would never reach”. Wrong. The claim now sets out the human trials on honey, cinnamon and garlic with every identifier linked, including two that were built to test a cholesterol outcome and found none, and two reviews of honey that disagree about the direction of the effect. The verdict on the claim is unchanged.
- November 10, 2026 — Honey, Garlic and Cinnamon: A Traditional Morning Preparation. The research section said the work on the ingredients was “most of it on isolated compounds in a laboratory” and that none of it settled anything. The first was wrong and the second too strong. It now says what the trials found, where they came out empty, and that the doses were capsules and weighed grams rather than a spoon and a clove.
- November 10, 2026 — How the Honey, Garlic and Cinnamon Drink Is Made, Step by Step. The same error in the same words: “the bulk of that work is done to isolated compounds in laboratory conditions”. Corrected, and the walkthrough now points at the claims review for the trials instead of summarising them a second time.
- November 10, 2026 — Claims Review: The Beet, Lemon and Ginger Morning Drink. A reader question was answered with “usually at a concentration nobody reaches by drinking a glass of anything”. One of those trials used 250 mL of beetroot juice a day, which is a quantity a person can drink, and it found a fall in blood pressure that held for four weeks. The answer has been rewritten and the page now lists the trials on each of the three ingredients, including the finding that no randomised trial of ingested lemon exists at all.
Further entries appear above this line in reverse order, each with the date, the page, what was wrong and what it says now.
Asking for one
Write to the editorial address with the page and the sentence. A message that names the sentence gets read the same day; a message that says a page is wrong without saying where takes longer, because the first thing anybody does with it is go looking.
Requests to remove accurate information are not corrections and are handled under the terms. Requests about what the site stores are handled under the privacy policy, which is short.