2 days ago
- Kuiper's test is a rotationally invariant test for comparing a sample to a reference distribution, offering a balance between power and simplicity.
- Hypothesis testing involves trade-offs between false positives (Type I error) and false negatives (Type II error), with power being the ability to detect true differences.
- Q-Q plots visually compare a sample to a reference distribution by plotting quantiles, making it easier to assess deviations from a diagonal line.
- Examples include testing birthday uniformity, incident spacing, commit sizes, justice retirement ages, airspace violations, and treatment effects on scores.
- The article presents a tool that computes Kuiper's test statistic, provides Q-Q plots, and supports multiple theoretical and empirical reference distributions.