Decides whether a reported gap between two groups is a real difference or sampling noise, and reports the p-value, the 95 % confidence interval on the difference and an effect size rather than a bare verdict. Use when a claim rests on a comparison — "Model A scored 87% vs Model B's 85%", "the new variant lifted conversion 12%", "is this A/B test result significant?", "run a significance test on these two proportions". Not for pooling several studies into one estimate (use `meta-analysis`) or for planning a test before data exist (use `experimental-design`).
