Providing software solutions since 1976
Nonparametric Analysis Procedures
- FREQ — One-way to n-way frequency and contingency (crosstabulation) tables
- GAM — Fits generalized additive models
- KDE — Univariate and bivariate kernel density estimation
- LOESS — Nonparametric method for estimating regression surfaces
- NPAR1WAY — Nonparametric tests for location and scale differences across a one-way classification
- TPSPLINE — Provides penalized least squares estimates
Nonparametric Analysis
In statistical inference, or hypothesis testing, the traditional tests are called parametric tests because they depend on the specification of a probability distribution (such as the normal) except for a set of free parameters. Parametric tests are said to depend on distributional assumptions. Nonparametric tests, on the other hand, do not require any strict distributional assumptions. Even if the data are distributed normally, nonparametric methods are often almost as powerful as parametric methods.
Below are highlights of the capabilities of the SAS/STAT procedures that perform nonparametric analysis:
- nonparametric analysis of variance
- nonparametric measures of association
- exact probabilities computed for many nonparametric statistics
- Kruskal-Wallis, Wilcoxon-Mann-Whitney, Friedman's chi-squared, and other rank tests for balanced or unbalanced one-way or two-way designs
- sign and signed rank tests for single samples
- simple linear rank statistics based on Wilcoxon, median, Savage, and Van der Waerden scores
- tests for scale differences include Siegel-Tukey, Ansari-Bradley, Klotz, and Mood
- Kolmogorov-Smirnov, Cramer-von Mises, and Kuiper (two sample only) statistics for comparing empirical distribution functions in several samples
- kernel density estimation
- thin-plate smoothing splines
Statistics and Operations Research Home Page | SAS/STAT Software
