search for: overparameterised

...ons are not meaningful. in: anova.lme(mod2a.REML, mod2b.REML) Well yes, I suppose that’s right, they are not the same fixed effects. But why does REML give them such different Log likelihoods? And what should I do if I want to compare a larger set of models. For example the following, admittedly overparameterised model, can be fitted (slowly) by either method >mod4<-lme(wthole~poly(nplants,2),data=d3,random=~poly(nplants,2)|field/ subplot,method="ML") But this doesn’t work by either method… > mod4<-lme(wthole~nplants+I(nplants^2),data=d3,random=~nplants+I(nplants^ 2)|field/subplo...

Dear List-members, Hopefully someone will help through my confusion: In order to get the same coefficients as we get from the following ## require (MASS) summary ( lm(Gas ~ Insul/Temp - 1, data = whiteside) ) ...................... we need to do the following (if we use model.matrix to specify the model) ## summary ( lm(Gas ~ model.matrix(~ Insul/Temp - 1) - 1, data = whiteside) )