Study design is context, not a score.
A randomized trial and an observational study answer different questions and carry different limitations. Our labels make the design visible without automatically declaring the finding “strong” or “weak.”
Syntheses of multiple studies. Their usefulness depends on the quality, similarity and completeness of the included evidence.
Participants are assigned to interventions using randomization. Useful for causal questions, but sample size, blinding, adherence and endpoints still matter.
Interventional human research that may not be randomized or may use other controlled designs.
Cohort, cross-sectional and other observational designs can reveal associations, but confounding and selection effects can limit causal conclusions.
Useful for hypotheses, mechanisms, unusual observations and context, but not automatically equivalent to controlled human outcome evidence.