Jump to a section
General consumer education for adults 21+—not medical, scientific, laboratory, or legal review. Do not delay diagnosis or treatment because of a product claim or this worksheet. Verify current rules and source records for the exact product, place, and date.
THE USEFUL PART
Keep these in your head.
- Copy the full claim and surrounding image, headline, footnote, disclosure, product identity, URL, and date before evaluating it.
- Name the claim type: identity, composition, safety, effect, disease, timing, legality, testing, certification, or comparison claims require different evidence.
- Match the evidence to the exact product or formulation, route, amount, population, comparator, outcome, duration, and conditions; an ingredient study does not automatically validate a finished product.
- Separate an official rule, a public-health recommendation, a scientific study, a laboratory result, a manufacturer instruction, and a personal story instead of ranking them by logo alone.
- Record missing, conflicting, superseded, or inaccessible evidence as unresolved. A disclaimer, badge, influencer, AI summary, or large reference list cannot repair the gap.
Capture the net claim before it moves
Save the exact words, page or post URL, date, seller or speaker, product and variant, images, charts, footnotes, linked study, disclosures, and the action the message asks you to take. Screenshots help preserve a changing ad, but keep the live URL because a screenshot cannot show later corrections, link destinations, or page scope.
Do not isolate one careful sentence from an aggressive headline, product name, before-and-after image, white coat, badge, testimonial, or purchase button. The Federal Trade Commission evaluates both express and reasonably implied claims and the overall impression of an advertisement. The useful unit is therefore the complete message a reasonable reader could take away—not the marketer's preferred paraphrase after a challenge.
Name the kind of claim
| Claim lane | Question it answers | Evidence that does not settle it alone |
|---|---|---|
| Identity or composition | What is in this exact package, in what form and amount? | A product name, ingredient list without amounts, generic certificate, or another lot |
| Safety | Safe for whom, at what exposure, with which medicines, route, duration, and monitoring? | Natural, hemp, nonintoxicating, passed one panel, or no complaint shown |
| Effect or performance | What measurable outcome changed, compared with what, and for how long? | Mechanism, animal or cell result, testimonial, aroma term, or satisfaction survey |
| Disease or treatment | Does the exact intervention prevent, diagnose, treat, mitigate, or cure a condition? | A structure/function phrase, clinician image, preliminary result, or unrelated approved drug |
| Timing | When might an effect begin, peak, end, or leave residual impairment? | A universal countdown, one person's experience, or a different route and formulation |
| Legal or regulatory | Which rule controls this product, activity, person, place, and date? | A federal hemp phrase, shipping offer, disclaimer, state map, or proposed bill |
| Testing or certification | Who tested or certified exactly what, to which method or standard, with what status? | A seal image, QR code, laboratory logo, potency page, or self-created badge |
One sentence can occupy several lanes. “Lab-tested hemp gummies clinically proven safe for sleep in every state” makes composition, testing, safety, treatment, population, and legal claims. Split them. One source is unlikely to answer all six.
Give every source a job and a boundary
| Source record | Useful for | Boundary to preserve |
|---|---|---|
| Statute, regulation, court record, or regulator page | The rule, agency position, license, recall, warning, or required process in its jurisdiction | Not clinical proof, universal law, or a substitute for checking effective dates and local rules |
| Systematic review or evidence assessment | A structured view of multiple studies for a defined question | Can inherit weak or mismatched studies; read population, products, outcomes, dates, and certainty |
| Controlled human study | A defined intervention, comparator, population, outcomes, duration, and harms | Does not automatically transfer to another ingredient, dose, route, product, population, or outcome |
| Observational human study | Associations, patterns, and uncommon or long-term questions | Association is not automatically causation; confounding and measurement limits matter |
| Animal, cell, or chemistry study | Mechanisms, plausibility, hazards, methods, and future research | Cannot by itself establish a consumer health benefit or clinical safety in people |
| Lot report or certificate | Measured results for the identified sample under the named method and reporting system | Not a human-outcome study, every-package guarantee, legal opinion, or proof of conditions after sampling |
| Manufacturer instructions | Exact-model use, compatibility, warnings, maintenance, warranty, and intended conditions | Commercial source with limited scope; not independent comparative proof or a universal safety finding |
| Testimonial, rating, forum post, or personal log | A question, signal, experience, or usability observation worth investigating | Cannot establish general causation, prevalence, safety, authenticity, or a treatment effect |
“Government source” is not one evidence level. A regulator can authoritatively explain the rule it administers while summarizing health evidence at a different level. A medical evidence review can interpret health research without deciding whether a local dispensary is licensed. Record the job each source is doing.
Run the nine-part evidence-transfer check
- Exact claim.
Does the source test or establish the actual statement, including its certainty and implied meaning?
- Exact material.
Is the evidence about the finished product, a named formulation, one purified compound, a different extract, or the plant broadly?
- Exact route and amount.
Oral, inhaled, topical, transdermal, and prescription routes are not interchangeable. Keep stated amounts and units attached.
- Exact population.
Age, diagnosis, medicines, pregnancy, prior exposure, setting, and exclusion criteria determine who was actually studied.
- Exact comparator.
Placebo, usual care, another active product, baseline, no exposure, or a different group answer different questions.
- Exact outcome.
A laboratory marker, rating scale, symptom report, function, diagnosis, adverse event, and driving task are not synonyms.
- Exact time.
One exposure, several weeks, long-term use, and follow-up after stopping cannot be collapsed into “works.”
- Whole result.
Read effect size, uncertainty, dropouts, adverse events, protocol, analysis, limitations, funding, corrections, and contrary findings—not only the abstract conclusion.
- Current status.
Check whether the page, rule, label, study record, recall, or report was corrected, superseded, withdrawn, retracted, or replaced.
If any field differs, describe the transfer rather than hiding it: “This study tested a purified oral formulation in adults with condition X for Y weeks; it does not establish that this retail gummy has the same composition, exposure, benefit, or risk.” Missing is a result, not permission to generalize.
Use the one-page claim audit record
Print the page or copy these headings into notes. Preserve facts and boundaries rather than awarding a numerical “credibility score.” A single disqualifying mismatch—wrong product, withdrawn paper, unrelated outcome, or prohibited activity—should not be averaged away by several decorative strengths.
Treat these shortcuts as stop signs
- “CLINICALLY PROVEN INGREDIENT”An ingredient study may not test the marketed product, amount, route, population, outcome, or claim.
- “ONE STUDY SHOWS”Find the complete study, registration, corrections, limitations, effect size, and the rest of the relevant evidence.
- “DOCTORS RECOMMEND”Ask which doctors, how selected, what question, which product, what evidence, and whether an endorsement is paid.
- “FDA REGISTERED” OR AGENCY LOGORegistration, listing, a facility record, or a borrowed seal is not approval of every product or claim.
- “LAB TESTED”Demand the exact lot, complete final report, laboratory identity, methods, units, panels, limits, qualifiers, and current status.
- “NO SIDE EFFECTS”Absence from an ad, small study, review set, or personal experience is not proof of no risk.
- “LEGAL IN ALL 50 STATES”Identify the exact product and activity, then check current federal, state, tribal, local, property, carrier, and program rules.
- “MONEY-BACK GUARANTEE”A return promise is not evidence that a health, safety, quality, or legal claim is true.
Turn fact-checking into a better purchase exit
Before payment, keep the claim audit beside the ordinary product record. Verify the lawful seller, exact variant, package and lot identity, ingredients, serving and package math, complete current report, warnings, recall status, storage, return or complaint route, and final delivered cost. A claim that survives a science check can still belong to the wrong lot or an unlawful seller; a traceable package can still carry an unsupported treatment promise.
Verify the seller and regulated-product trail →
Compare compatible packages without ranking potency →
No brand, seller, product, price, coupon, rating, or affiliate link appears here. The repository contains no approved affiliate tracking ID, and a commercial relationship would not change the evidence-transfer test.
Primary and authoritative source shelf
Sources checked October 16, 2026. FTC guidance explains advertising-law principles and does not itself adjudicate a particular claim. FDA and NCCIH pages summarize federal product and health-evidence boundaries; they do not replace the complete underlying study, the exact product record, or current state and local rules.
Big Bud Man standard: know the source, read the batch, protect kids and pets, and never drive impaired.