How would you distinguish a superiority trial from a noninferiority trial?

Why

The research question and decision boundary must be set before data are analyzed. A superiority trial tests for evidence of a favorable difference, commonly versus placebo or an active comparator. A noninferiority trial typically compares a new treatment against an established active treatment and aims to rule out an unacceptable loss of efficacy.

The noninferiority margin must be justified using clinical judgment and credible evidence of comparator effect, with attention to assay sensitivity and consistency of that effect across settings. Imagine a noninferiority margin of five percentage points for treatment success, with a positive difference favoring the new treatment. The lower confidence bound for the new-minus-control difference must lie above minus five points to support noninferiority for that estimand and analysis.

Equal-looking averages are insufficient if uncertainty is wide. Protocol deviations, missing outcomes and analysis populations deserve special care because they may bias results toward apparent similarity. Noninferiority is not equivalence, and it does not follow from a nonsignificant superiority test.