Dear Ali,
There really isn't anything we can say, given this amount of information.
It sounds as if you already suspect what is wrong (i.e. the regressor), so have a close look at that.
Note that the varcope can be badly behaved when your design is rank deficient or nearly so.
Check the design efficiency output of the GLM to get a feeling for if this is the case.
All the best,
Mark
On 2 Oct 2013, at 21:48, Ali Golestani <[log in to unmask]> wrote:
> I am getting wrong group result, because varcope for one of the subjects has large values. Can anyone explain to me what could be the reason of getting large cope and varcope?
> Thanks
>
> p.s. I don't think the problem is the dataset, as the cope and varcope I get from it with another regressor is perfectly normal. there must be something wrong with the regressor.
|