How we grade

Each claim lists the studies behind it. Each study is recorded with its design, size, length, funder, and whether the claim was what it set out to test. A program turns that into a grade. Nobody picks the word.

• Established — two or more placebo-controlled trials found it on their main outcome, or a pooled analysis of such trials did.
• Promising — one placebo-controlled trial found it on its main outcome.
• Tested — no effect found — a placebo-controlled trial tested it as its main outcome and found nothing.
• Early evidence — human studies without a placebo, or a result on a measure the trial wasn’t built around.
• Lab studies only — animals or cells.
• No real evidence — testimonials, company material, tradition, or nothing.

The rules we hold ourselves to

• Testimonials and company white papers never raise a grade.
• A trial that tested a claim and found nothing outranks weaker positive results.
• Who funded a study is always shown, and never changes the grade.
• Every page shows the date its evidence was last checked.