Methodology
Every claim we evaluate follows the same structured process — from identifying what a product actually asserts, to grading the quality of evidence behind it, to issuing a verdict grounded in the literature. This page documents that process in full.
Review Framework
Each review follows a structured framework. The Claim → Evidence → Verdict architecture is consistent across all article types:
Claim Identification
We document the specific efficacy claims made by the product, manufacturer, or category marketing — in their own language, without paraphrase.
Literature Review
We search PubMed, Cochrane, and relevant regulatory databases for primary literature. We prioritize double-blind RCTs with appropriate controls, published in peer-reviewed journals.
Evidence Grading
We score evidence strength on a 5-point scale. A 5/5 requires multiple high-quality RCTs with consistent findings. A 1/5 reflects mechanistic speculation or in vitro data only.
Verdict
We issue one of four verdicts: Claim Supported, Partially Supported, Mixed Evidence, or Claim Unsupported. Verdicts are tied to the specific claim — not the ingredient or product category in general.
Evidence Grading Scale
Our 5-point scale reflects the quality and consistency of the available human evidence:
Strong
Multiple high-quality RCTs with consistent findings and adequate statistical power.
Good
At least one well-designed RCT; minor limitations in sample size, blinding, or follow-up duration.
Moderate
Controlled trials with significant limitations, or a single small RCT without replication.
Weak
Observational data, case series, or poorly controlled trials with high risk of bias.
Insufficient
Mechanistic speculation, animal data, or in vitro studies only — no human clinical evidence.
Independent Lab Testing
For select product reviews, we use third-party laboratory analysis to verify claims that cannot be assessed from published literature alone.
- We partner with third-party laboratories to evaluate product claims.
- Products are purchased at retail price by our team — we do not accept manufacturer-supplied samples for tested reviews.
- Data and lab reports are cited directly in the relevant article.
- Raw data is available on request.
Conflict of Interest Policy
Marketing or Science does not accept affiliate revenue. We do not accept manufacturer-supplied samples for any review. All products reviewed are purchased independently by our team or partners at retail prices.
Reviewer disclosures are required for all contributors. Any financial or professional relationship with a reviewed brand or its competitors must be declared and will be noted in the article.
Update Policy
All analyses are dated at publication. When material new evidence emerges — a high-quality RCT, a significant regulatory action, or a retraction of a key cited study — we conduct a full review update. Updated articles are flagged with a revision date alongside the original publication date. Minor corrections (typographical, formatting) are made silently.