Hello everybody, i'm trying to run a lme but am having a problem. My model have many response variables and when i run the anova, the number of interactions (up to six) is great and the p-values of all variables not appear. I wanted to know if i could to control the number of interactions of the model, or if this method is the most advisable for so many variables. Thanks in advance! Follows the procedure I'm using: v.is<-lme(is~direction*envir*region*hour*estom*esl, random=~1|ind/dir/reg, tabela) anova(v.is,test="F") -- View this message in context: http://r.789695.n4.nabble.com/Controlling-the-number-of-interactions-of-a-lme-tp4650183.html Sent from the R help mailing list archive at Nabble.com.
cleberchaves <cleberchaves <at> gmail.com> writes:>[snip]> My model have many response variables and when i run the anova, the number > of interactions (up to six) is great and the p-values of all variables not > appear. > > I wanted to know if i could to control the number of interactions of the[snip]> v.is<-lme(is~direction*envir*region*hour*estom*esl, random=~1|ind/dir/reg, > tabela) > anova(v.is,test="F") >You probably want something like is ~ (direction+envir+region+hour+estom+esl)^2 for example, which would include the main effects and all two-way interactions. See the "Details" section of ?formula for a (terse) description.
Is it, bbolker! Thank you very, very much! -- View this message in context: http://r.789695.n4.nabble.com/Controlling-the-number-of-interactions-of-a-lme-tp4650183p4650290.html Sent from the R help mailing list archive at Nabble.com.