Hi everyone, I am wandering if someone was able to get an anova table using the function lmer in lme4 package with an other distribution than gaussian (ex: poisson). In fact, I am using a GLMM with a poisson distribution where I have 2 independant fixed variables (coded as factors with respectively two and three levels) and I am interested with simple and interaction effects (I also have a couple of random variables). However, rather than to have an estimate of each combination (plus intercept), I would like first to know the effects of my 3 sources of variation (not the combination of each level). This works well with a gaussian distribution (using anova()) but not with another one (ex: poisson). I have seen exemple where it seems to work pretty well even with factor variables so I guess there is a way to do it... If someone know how to do, I would greatly appreciate your help. Best regards, Julien ------------------------------------------ Julien Beguin Etudiant au doctorat Facult? de Foresterie et de G?omatique Universit? Laval, local 2113 2405 rue de la Terrasse, G1V 0A6 Qu?bec (Qc) Tel: (418) 656-2131 poste 2620 http://www.cen.ulaval.ca/anticosti/jbeguin.html