[R] simtest for Dunnett's test

Laurent Houdusse l.houdusse at cerep.fr
Fri Aug 13 14:33:50 CEST 2004


Ok,
I use simtest because i retrieve the same result
But i don't know how to do the comparaison of P


Laurent Houdusse
Analyste Programmeur



-----Message d'origine-----
De : Liaw, Andy [mailto:andy_liaw at merck.com] 
Envoyé : vendredi 13 août 2004 14:24
À : Laurent Houdusse; 'r-help at stat.math.ethz.ch'
Objet : RE: RE : [R] simtest for Dunnett's test


> From: Laurent Houdusse
> 
> However, on several tests, my results are correct:
> The t-value seems always to correspond to the q'value calculated by
> SigmaStat...  
> 
> So,What do I have to use to find the same results as
> SigmaStat (value and
> comparaison)?

Seems like you don't have much of a choice:

1. Accept the fact that they are different, and use simtest.  I believe the
authors have good reasons to believe that their procedures are better.

2. Roll your own.  You might be able to take shortcuts by looking for
existing C or Fortran code for Dunnett's procedure and try to link that to
R.

Andy
 
> 
> Laurent Houdusse
> Analyste Programmeur
> 
> 
> 
> -----Message d'origine-----
> De : Liaw, Andy [mailto:andy_liaw at merck.com]
> Envoyé : vendredi 13 août 2004 14:06
> À : Laurent Houdusse; 'r-help at stat.math.ethz.ch'
> Objet : RE: [R] simtest for Dunnett's test
> 
> 
> Before you do any of that, you should realize the fact that
> simtest does
> _not_ do Dunnett's test:  It can use `Dunnett' _contrasts_ to 
> do comparisons
> with a control, but the actual procedure is different.  You 
> are unlikely to
> get the same result as other packages that perform standard 
> Dunnett's test
> (or Tukey's, for that matter).  
> 
> Andy
> 
> > From: Laurent Houdusse
> > 
> > Hi!
> > 
> > I use simtest fonction of multcomp package to compile a Dunnett's 
> > test. I have 10 treatments and one control group, so i create a
> matrix with:
> > 
> > m<-matrix(0,10,11)
> > m[1,1]<--1
> > m[1,2]<-1
> > m[2,1]<--1
> > m[2,3]<-1
> > m[3,1]<--1
> > m[3,4]<-1
> > m[4,1]<--1
> > m[4,5]<-1
> > m[5,1]<--1
> > m[5,6]<-1
> > m[6,1]<--1
> > m[6,7]<-1
> > m[7,1]<--1
> > m[7,8]<-1
> > m[8,1]<--1
> > m[8,9]<-1
> > m[9,1]<--1
> > m[9,10]<-1
> > m[10,1]<--1
> > m[10,11]<-1
> > rownames(m)<-c("vehicle+vehicle vs. morphne
> > 60mg/kg+vehicle","vehicle+vehicle vs. Naloxone 
> > 30mg/kg+vehicle","vehicle+vehicle vs. Naloxone 30mg/kg+Morphine 
> > 60mg/kg","vehicle+vehicle vs. RWJ-69528-000-B 
> > 0.1mg/kg+vehicle","vehicle+vehicle vs. RWJ-69528-000-B 
> > 1mg/kg+vehicle","vehicle+vehicle vs. RWJ-69528-000-B 
> 0.1mg/kg+morphine
> > 60mg/kg","vehicle+vehicle vs. RWJ-69528-000-B 1mg/kg+morphine
> > 60mg/kg","vehicle+vehicle vs. vehicle+acetaminophen 
> > 600mg/kg","vehicle+vehicle vs. RWJ-69528-000-B 
> 0.1mg/kg+acetaminophen
> > 600mg/kg","vehicle+vehicle vs. RWJ-69528-000-B 1mg/kg+acetaminophen
> > 600mg/kg")
> > y<-c(21.5,14.7,16.7,18.8,17.3,15.527,17.551,17.199,15.895,14.3
> > 88,30,30,30,21 
> > .2,14.6,30.83,30.622,30.465,24.563,31.416,30,30,30,30,19.6,17.
> > 61,30.346,20.8 
> > ,17.045,18.129,17.9,15.4,30,15.9,22.9,31.909,25.633,24.6,20.35
> > 5,18.807,30,18 
> > .9,16.4,18.2,22.3,21.3,28.571,17.357,16.983,30.40,26.3,24.4,25
> > ,24.1,14.4,10. 
> > 256,13.266,29.6,15.344,18.69,30,30,25.7,30,28.5,30,30.32,30,19
> > .879,30.716,25 
> > .2,30,30,30,30,30,14.063,28.811,30.396,30.609,30,14.1,30,30,30
> > ,26.322,30.386 
> > ,28.583,15.77,29.341,30,17.1,30,30,30,30,20.451,30,23.967,30.5
> > 07,30,30,30,27
> > .7,30,28.5,22.568,32.379,30.374,28.263)
> > h<-factor(c(rep(1:11, each = 10)))
> > result<-summary(simtest(y ~ h,cmatrix=cbind(0,m)))
> > rownames(result$estimate)
> > result
> > 
> > 
> > I want to compare my results with the results obtained with
> SigmaStat
> > Software. In my results, i retrieve a correct q' value with the
> > simtest t value How can i say if P<0.05 like the SigmStat results?
> > Can i use p raw value, p bonf value, p adj value or other to compare
> > directly? Or can i use the Dunnett's tables?
> > 
> > Thanks for your help!
> > 
> > 
> > 
> > 
> > Laurent Houdusse
> > Analyste Programmeur
> > 
> > ______________________________________________
> > R-help at stat.math.ethz.ch mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-help
> > PLEASE do read the posting guide!
> > http://www.R-project.org/posting-guide.html
> > 
> > 
> 
> 
> --------------------------------------------------------------
> --------------
> --
> Notice:  This e-mail message, together with any attachments, contains 
> information of Merck & Co., Inc. (One Merck Drive, Whitehouse Station, 
> New Jersey, USA 08889), and/or its affiliates (which may be known
> outside the
> United States as Merck Frosst, Merck Sharp & Dohme or MSD and 
> in Japan, as
> Banyu) that may be confidential, proprietary copyrighted 
> and/or legally
> privileged. It is intended solely for the use of the 
> individual or entity
> named on this message.  If you are not the intended 
> recipient, and have
> received this message in error, please notify us immediately 
> by reply e-mail
> and then delete it from your system.
> --------------------------------------------------------------
> --------------
> --
> 
> 


----------------------------------------------------------------------------
--
Notice:  This e-mail message, together with any attachments,...{{dropped}}




More information about the R-help mailing list