Non-parametric tests скачать в хорошем качестве

Non-parametric tests

This lecture introduces non-parametric tests, which are distribution-free methods suitable for categorical and ordinal data. It focuses on the chi-squared (χ²) test family, explaining the uncorrected χ² test for association between categorical variables, including step-by-step calculations and a worked example. The lecture also covers specialised variations: Yates' correction for small expected frequencies, the Mantel-Haenszel test for controlling confounding variables through stratification, McNemar's test for paired categorical data, and Fisher's exact test for very small samples. Each method is illustrated with practical examples from medical research, emphasising when each test is appropriate and how to interpret results correctly.