How we choose what to rate, where the data comes from, how we test a claim, who signs off a score, and what happens publicly when we get something wrong.
Every claim about an effect links to a source — a human trial where one exists, and an honest note where it doesn't. We prefer primary research and reputable aggregators over other people's roundups.
Editorial ratings are never for sale. Paid placements are always labelled and always sit outside the ranking.
We correct errors openly, with a dated note on the page.
We never invent reviews, numbers, experts, or test results. If we haven't verified it, we don't publish it as fact.