Wolfgang Waser
2009-Oct-07 15:28 UTC
[R] repeat(?) measurement ANOVA or (general? mixed?) linear model?
Hi all, after browsing the archives for hours I'm still not sure about the proper analysis for my dataset. I subjected each of about 50 critters (about 10 each in 5 distinct populations) to 4 consecutive treatments (exposure to increasing concentrations), with one measurement per treatment and individual. I, of course, want to know, if there was a treatment and / or a population effect. Since I'm dealing with 'paired' data, and also with a (expected) 'linear' response to a 'linear' treatment, how do I perform the proper analysis? I tried something like: summary(aov(y~group+treatment) summary(lme(y~group+treatment) but I don't see, how these formula would differentiate between a 'paired' or 'unpaired' experimental design. Somewhere in the archives it was mentioned that you could use the error-term to make that distinction (Baron-rpsych.pdf, pp.36-) summary(aov(y~group+treatment+Error(?), na.action=na.omit) summary(lme(y~group+treatment,random=~1|?,na.action=na.omit) but, unfortunately, that level of sophistication is still beyond my statistical powers. Any advice on the matter would be very much appreciated Wolfgang
Dieter Menne
2009-Oct-08 06:56 UTC
[R] re peat(?) measurement ANOVA or (general? mixed?) linear model?
Wolfgang Waser wrote:> > I subjected each of about 50 critters (about 10 each in 5 distinct > populations) to 4 consecutive treatments (exposure to increasing > concentrations), with one measurement per treatment and individual. > > ... > Since I'm dealing with 'paired' data, and also with a (expected) 'linear' > response to a 'linear' treatment, how do I perform the proper analysis? >You could do this with aov, but it is not very flexible. I suggest that for this type of data you should use lme in package nlmer where you can easily specify "repeats" or serial correlation with the random term. The book by Pinheiro/Bates should be in the biologist's bookshelf, it has many example. Getting started without the book is difficult. Dieter -- View this message in context: http://www.nabble.com/repeat%28-%29-measurement-ANOVA-or-%28general--mixed-%29-linear-model--tp25788824p25799282.html Sent from the R help mailing list archive at Nabble.com.