How can I compare variances for different levels of a factor in a mixed effect model?
I'm fitting a mixed effects model (in R using the nlme package) to signal onset time points, like so:
m1.lme <- lme(Onset ~ Signal, random=list(~1|Participant), data=mydata)
I suspect the variance of the onsets is different for various signals. So I add the varIdent argument to my fit:
m2.lme <- lme(Onset ~ Signal, random=list(~1|Participant), data=mydata, weights=varIdent(form=~1|Signal))
Now I want to compare the variances for different levels of the factor signal, i.e. is the variance for signal A significantly bigger than for signal B? Is this possible, and if so, how?
