Hi Josh, Thanks for your reply. You're right about letting R's method dispatch system choose the method for summary and inserting more spaces in the code. I was just messing around with the code in the former case and forgot to change it back. As far as not having enough whitespace goes, I've been learning using The R Book and this may have caused me to learn some bad habits. I actually like The R Book a lot and have been able to learn a great deal from it. I've seen a few people comment about the lack of whitespace in the code though, and so this is something I'll need to change. Your comment about editing the contrast matrix is interesting. I tried it out with my own code with the few moments I had this morning (see below). One thing I noticed immediately is that altering the contrast matrix, and then running the model, and then running Dunnett's test using multcomp doesn't work correctly. That is, the reference level for Dunnett's test is still "HI." If you use relevel though, the reference level for Dunnett's test is correct. I haven't had time to investigate this yet. I get the impression though that editing the contract matrix is somewhat like using the "ref=" option in SAS procedures in that the change in the reference parameter doesn't (necessarily) carry over into subsequent analyses. That was one of the things I thought was so great about relevel. You set the reference level once and it stays set until you decide to change it. Anyway, this was very interesting and informative. Thanks for your help. Paul connection <- textConnection(" 101 LO 21 104 LO 18 106 LO 19 110 LO . 112 LO 28 116 LO 22 120 LO 30 121 LO 27 124 LO 28 125 LO 19 130 LO 23 136 LO 22 137 LO 20 141 LO 19 143 LO 26 148 LO 35 152 LO . 103 HI 16 105 HI 21 109 HI 31 111 HI 25 113 HI 23 119 HI 25 123 HI 18 127 HI 20 128 HI 18 131 HI 16 135 HI 24 138 HI 22 140 HI 21 142 HI 16 146 HI 33 150 HI 21 151 HI 17 102 PB 22 107 PB 26 108 PB 29 114 PB 19 115 PB . 117 PB 33 118 PB 37 122 PB 25 126 PB 28 129 PB 26 132 PB . 133 PB 31 134 PB 27 139 PB 30 144 PB 25 145 PB 22 147 PB 36 149 PB 32 ") gad <- data.frame(scan(connection, list(patno=0, dosegrp="", hama=0), na.strings=".")) #### ANOVA #### #gad$dosegrp <- relevel(gad$dosegrp, 3) #contrasts(gad$dosegrp) <- contr.treatment(n = 3, base = 3) model1 <- aov(hama ~ dosegrp, data = gad) summary(model1) #### Dunnett's Test #### library(multcomp) cht <- glht(model1, linfct = mcp(dosegrp = "Dunnett")) summary(cht, test = univariate())
[[alternative HTML version deleted]]
______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.