=================================y=c(100,200,300,400,500) treatment=c(1,2,3,3,4) block=c(1,1,2,3,3) summary(lm(y~as.factor(treatment)+as.factor(block))) ================================= The aim is to find a model that can estimate the comparison between treatment 1 with 2 and treatment 3 with 4. I have tried all the possible ones ==========================================relevel(as.factor(block),ref="1");relevel(as.factor(treatment),ref="1") relevel(as.factor(block),ref="1");relevel(as.factor(treatment),ref="2") relevel(as.factor(block),ref="1");relevel(as.factor(treatment),ref="3") relevel(as.factor(block),ref="1");relevel(as.factor(treatment),ref="4") relevel(as.factor(block),ref="2");relevel(as.factor(treatment),ref="1") relevel(as.factor(block),ref="2");relevel(as.factor(treatment),ref="2") relevel(as.factor(block),ref="2");relevel(as.factor(treatment),ref="3") relevel(as.factor(block),ref="2");relevel(as.factor(treatment),ref="4") relevel(as.factor(block),ref="3");relevel(as.factor(treatment),ref="1") relevel(as.factor(block),ref="3");relevel(as.factor(treatment),ref="2") relevel(as.factor(block),ref="3");relevel(as.factor(treatment),ref="3") relevel(as.factor(block),ref="3");relevel(as.factor(treatment),ref="4") =========================================== each followed by a line of summary(lm(y~as.factor(treatment)+as.factor(block))) i seem to always get "NaN"s Am I doing something wrong there? What model should I use then? Thanks! casper -- View this message in context: http://n4.nabble.com/model-reparameterization-tp1749621p1749621.html Sent from the R help mailing list archive at Nabble.com.