<- file stat 97postnp.html ->

Kruskal Wallis, posthoc

=======================Hans-Peter Piepho, 12 Feb 1997==========ssc Message-ID: <9702120932.AA15280@fserv.wiz.uni-kassel.de> From: Hans-Peter Piepho <piepho@WIZ.UNI-KASSEL.DE> Subject: Re: Kruskal Wallis Post Hoc Comparisons Question> >I have been using the Kruskal Wallis test quite a bit recently. >I know that there are various post hoc comparison tests for >the one way ANOVA, but I am not aware of what comparison tests >I should be using in the non parametric environment. If anyone >can help me with names of tests or useful references I would be >most appreciative. > I suggest you take a look a the book by Jason Hsu (1996 Multiple comparisons. Chapman and Hall), Section 5.3. Hochberg and Tamhane (1987 Multiple comparison procedures) is another good source. There are a number of "post-hoc" methods. Those based on joint ranking (the same ranks used for Kruskal-Wallis) do not control the experiment-wise Type I error in the strong sense (i.e. they control the Type I error experimentwise only under the global H0). Methods based on pairwise rankings are preferable, for example the single-step procedure by Steel and Dwass. Hans-Peter Piepho _______________________________________________________________________ Hans-Peter Piepho

Nonpar posthoc

=======================Jerrold Zar, 15 Feb 1997==========ssc Message-ID: <s305a532.026@WPO.CSO.NIU.EDU> From: Jerrold Zar <T80JHZ1@WPO.CSO.NIU.EDU> jhzar@niu.edu Subject: Kruskal Wallis Post Hoc Comparisons -Reply Nonparametric multiple-comparison testing, analogous to Tukey, Dunnet, and Scheffe testing, is discussed in Zar, J. H. 1996. Biostatistical Analysis. 3rd ed. Prentice Hall.

Kruskal Wallis

=======================Thom Baguley, 12 Feb 1997==========ssc From: Thom Baguley <t.s.baguley@lboro.ac.uk> Subject: Re: Kruskal Wallis Post Hoc Comparisons Message-ID: <3301F1DC.465C@lboro.ac.uk> Siegel & Castellan dscribe multiple comparison between treatments for Kruskall-Wallace and also comparison of treatments against a control. Siegel, S., & Castellan, N. J., Jr. (1988). Nonparametric statistics for the behavioural sciences (2nd ed.). New York: McGraw-Hill. Ray Meddis covers post-hoc and multiple comparison approaches to non-parametric analyis of various in his book Statistics Using ranks. Meddis, R. (1984). Statistics using ranks: a unified approach. Oxford: Blackwell. Lastly, you could always use Bonferonni methods to control for e.g. all the pairwise comparisons (though this is probably not the preferred solution). * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Document by Rich Ulrich. E-mail to wpilib+@pitt.edu

FAQ top.

Ulrich home page.

Ulrich FAQ. http://www.pitt.edu/~wpilib/stats99.html