Evidence-appraisal glossary
Sensitivity analysis
A sensitivity analysis repeats an analysis under different reasonable assumptions or after excluding certain data to see whether the main conclusion still holds. If the result barely changes, the finding is considered robust.
Also called: robustness analysis, leave-one-out analysis.
In evidence synthesis this might mean dropping high risk-of-bias studies, switching between statistical models, or leaving out one study at a time. The important limit is that it only tests robustness to the choices actually varied, not to assumptions left unexamined; a stable result is reassuring but is not proof of correctness, and pre-specifying which analyses to run guards against picking the version that looks best.
This is a plain-language methodology definition for reading research. It is general education, not medical advice.