Hi,
the following is my R code, where drastic discrepancy is shown between the significant adjusted p values obtained from "mt.maxT (W&Y strong control of FWER)" and "mt.rawp2adjp" (4932 vs. 0). Would you please tell me why this happens, as I really do not know where is the bug in my code? Much Thanks
> ##################################################################
> #mt.maxT procedure:compute permutation adjusted p-values         # 
> #for W&Y step-down multiple testing procedures                   #
> ##################################################################
> resTPBvsP2<-mt.maxT(theta.PBvsP2,classlabel=cl.PBvsP2,test="pairt",side="abs",B=10000)
We're doing 10000 random permutations
b=100   b=200   b=300   b=400   b=500   b=600   b=700   b=800   b=900   ...b=10000 
> no.rawpTPBvsP2<-sum(resTPBvsP2$rawp<0.05);print(no.rawpTPBvsP2)
[1] 4932
> no.adjpTPBvsP2<-sum(resTPBvsP2$adjp<0.05);print(no.adjpTPBvsP2)
[1] 0
> 
> ordTPBvsP2<-order(resTPBvsP2$index)
> ##################################################################
> #rows are sorted first according to their adjusted p-values      #
> #next their unadjusted p values and finally their test statistics# 
> #i.e. resTPBvsP2$adjp==sort(resTPBvsP2$adjp)                     ##################################################################                    
> rawpTPBvsP2<-resTPBvsP2$rawp[ordTPBvsP2]
> #original data order, to be used in mt.rawp2adjp procedure#
> 
> ############################################
> #mt.rawp2adjp:computes adjusted p-values   # 
> #for simple multiple testing procedures    #
> #from a vector of raw (unadjusted) p-values#
> ############################################
> resPBvsP2T<- mt.rawp2adjp(rawpTPBvsP2, proc = c("Holm"))#FWER#
> no.adjpPBvsP2T<-sum(resPBvsP2T$adjp<0.05);print(no.adjpPBvsP2T)
[1] 4932

Thank you,
Anqi 
	[[alternative HTML version deleted]]