There are some methods and tests that are considered nonparametric in nature. Compared to parametric tests (e.g., t-test, z-test, F-test, ANOVA), nonparametric tests have the following advantages and a single disadvantage:
- Fewer assumptions are required for the underlying data’s population. Specifically, a nonparametric test does not require that the population be normally distributed. In fact, it does not require any specific distribution and, hence, is sometimes called distribution-free, or tests without specific population parameters (i.e., nonparametric).
- Smaller sample sizes can be used.
- Data with nominal and ordinal scales can be tested.
- Nonparametric methods have lower power and use the data less efficiently. Therefore, if assumptions have been met, it is better to use parametric tests whenever possible.
Some of the most common nonparametric tests are the Runs test for randomness, Wilcoxon test, Lilliefors test, Kruskal–Wallis test, and Friedman’s test.
