A randomized controlled trial (RCT) is a study design in which participants are allocated by chance to receive either an intervention (a supplement, drug, or protocol) or a control (placebo, standard of care, or alternative dose). Random assignment mainly reduces selection and confounding bias by balancing known and unknown prognostic factors in expectation; blinding and placebo controls address expectancy, performance, and measurement bias.
Why it matters
The supplement market is saturated with claims supported by mechanism of action or observational studies alone. RCTs are the threshold where a compound moves from "plausible" to "probably works." When Unfair weights a compound highly in recommendation ranking, the supporting evidence is almost always either an RCT or a meta-analysis of RCTs.
Design features that matter
Not all RCTs are equally strong. The features that predict reliability:
- Double-blind — neither participants nor researchers know who received the active compound until analysis. Protects against placebo expectancy, performance bias, and measurement or assessor bias.
- Placebo-controlled — the control group receives an inactive-looking product, helping keep expectancy effects comparable between arms.
- Adequately powered — enough participants to detect the expected effect size without false negatives; small trials (n < 30) generate noisy results.
- Intent-to-treat analysis — participants are analyzed in the group to which they were assigned, even if they dropped out; protects against selective attrition.
- Pre-registered — primary outcomes and analysis plan were declared before data was collected, preventing post-hoc cherry-picking.
An RCT without these features is closer to evidence tier "preliminary" than "robust human outcome."
Known limits
RCTs are still finite. Populations are often narrow (young men, athletes, or clinical patients), durations are often short (8–12 weeks), doses tested may not match real-world supplement labels, and rare adverse events are systematically underrepresented. An RCT showing "creatine improves repeat-sprint performance in collegiate males over 8 weeks" does not automatically generalize to a 55-year-old recreational runner over a year.
How this appears in Unfair
Compounds whose primary evidence is high-quality RCT or RCT meta-analysis are labeled with the highest evidence tier in Unfair's ingredient metadata, which raises their weight in recommendation ranking.
Clinical safety note
"Evidence from an RCT" does not mean "safe for you." Contraindications, medication interactions, and pregnancy considerations still apply regardless of trial quality. Treat trial evidence as directional, not prescriptive.