Displaying 1 result from an estimated 1 matches for "stmassheavy".
2011 Jan 17
1
Using anova() with glmmPQL()
...from this that variation with both roost situation and month are significant, but with no interaction term. So far so good. However, say I were interested only in whether or not those starlings were heavier or lighter than 78g: seemingly, I could change my response variable and analyse like this -
stmassheavy=ifelse(stmass>78,1,0)
lmeres1=lme(fixed=stmassheavy~mnth*roostsitu,random=~1|subject/mnth,na.action=na.exclude,family=binomial)
anova(object=lmeres1,test="Chisq")
but I get errors doing that. After a certain amount of web searching, I find that I'm supposed to use glmmPQL for this...