Space limitations prohibit a comprehensive discussion of the many statistical methods related to the techniques covered in previous chapters. However, a brief description of some of these alternate methods and comments about them should be made. One general issue is the role of so-called ranked-based or nonparametric methods for comparing groups and performing regression. Such methods are commonly recommended for dealing with nonnormality, so it is important to compare and contrast them with methods based on robust measures of location. Permutation tests are yet another method sometimes recommended for dealing with nonnormality. Another important issue is extending methods for comparing groups to more complicated experimental designs Finally, some comments about software will be made.
KeywordsPermutation Test Bootstrap Method Multiple Group Robust Measure Probability Curve
Unable to display preview. Download preview PDF.