Introduction
Inference regarding two related samples, or paired analysis, was addressed in Chapter 3 in discussions of the sign test and the Wilcoxon signed-rank test. When more than two populations are compared, those tests are no longer valid and different tests must be used. The same is true in parametric statistics where a t statistic is used for testing the means of one or two populations, but an F statistic is required for testing parameters from more than two populations. In parametric inference, the F test is used regardless of whether the populations are related or not, although knowledge of the relationship among the samples affects the design of the experiment and thus the analysis. Test statistics have ...
Get A Primer on Nonparametric Analysis, Volume II now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
