Hi !
I have an issue about GMM and I wondered if you could help me.
I have a dependant variable which is quanti but not continuous (numbers of tries before succeed, min = 4 by definition of the variable), and the histogram is :
I have to construct a model with random and fixed components so I have to use a GMM.
But with this kind of dependant variable, I don’t know which distribution and link I have to choose.
I thought about Poisson but mean (4.5) is too different of variance (1.03).
There is no overdispersion either so I don’t think negative binomial is ok …
I’m a little lost with this one …
If you can help me, I will be very greatful.
Thanks a lot in advance.
Have a nice summer !
Distribution and link in GMM
Re: Distribution and link in GMM
These trainings sound like they could be treated as ordinal data, in which case a cumulative link mixed model would be appropriate.