This is only partly an R question, but this seems a good place to ask. Like many, if not most, statistics programs, R has no post hoc tests for Kruskal Wallis or Friedman, and as far as I can make out they are not available in any of the R packages. I would be grateful if someone could point me in the right direction if I am wrong. As I said this is not uncommon. Parametric Anova procedures almost always, if not always, come with a post hoc procedure. Why is it not the same with non-parametric anovas. I assume there is some reason for this, and would be very pleased to have it explained. Many thanks, Graham S -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
myotis at cix.compulink.co.uk (Graham Smith) writes:> This is only partly an R question, but this seems a good place to ask. > > Like many, if not most, statistics programs, R has no post hoc tests > for Kruskal Wallis or Friedman, and as far as I can make out they are > not available in any of the R packages. I would be grateful if someone > could point me in the right direction if I am wrong.I think pairwise.wilcox.test is just about as close as one gets for the KW case.> As I said this is not uncommon. Parametric Anova procedures almost > always, if not always, come with a post hoc procedure. Why is it not > the same with non-parametric anovas.The math doesn't work out as neatly... E.g. the distribution of rank sums with some groups different is largely unknown. Er, well, it can be worked out as a permutation distribution I suppose, but it isn't nice and to my knowledge there's nothing that corresponds to the studentized range distribution. Then again, maybe there could be. -- O__ ---- Peter Dalgaard Blegdamsvej 3 c/ /'_ --- Dept. of Biostatistics 2200 Cph. N (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk) FAX: (+45) 35327907 -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
In-Reply-To: <x24rxgo6we.fsf at blueberry.kubism.ku.dk> Peter,>The math doesn't work out as neatly..But does that make them invalid, Conover and others certainly present post hoc tests for non-parametric anovas. Cheers, Graham S -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
On 2/27/01 8:00 AM, Graham Smith wrote:> This is only partly an R question, but this seems a good place to ask. > > Like many, if not most, statistics programs, R has no post hoc tests > for Kruskal Wallis or Friedman, and as far as I can make out they are > not available in any of the R packages. I would be grateful if someone > could point me in the right direction if I am wrong.The math is not as simple as in the parametric case. I've found some non-parametric post-hoc tests in Zar, Biostatistical Analysis, 4th ed, Prentice Hall, Chapter 11. (Nemenyi/Dunn test and some variations thereof, with explanations of advantages/drawbacks). They should be easy to implement, if you know some R. Cheers Kaspar Pflugshaupt -- Kaspar Pflugshaupt Geobotanisches Institut Zuerichbergstr. 38 CH-8044 Zuerich Tel. ++41 1 632 43 19 Fax ++41 1 632 12 15 mailto:pflugshaupt at geobot.umnw.ethz.ch privat:pflugshaupt at mails.ch http://www.geobot.umnw.ethz.ch -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
In-Reply-To: <B6C90B83.81D0%pflugshaupt at geobot.umnw.ethz.ch> Kaspar,> They should > be easy to implement, if you know some R.At the moment I know very little R so implementation may not be that easy for me. I can use them in Statistica, so that isn't an immediate problem, I was as much as anything interested in why so few packages offered these tests Thanks for the pointer to Zar. Cheers, Graham S -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._