I'm trying to run a GZLM analysis with the following characteristics
- type of model: Custom, gamma distribution with identity link function
- Dependent variable is a continuous variable following a gamma distribution (obviously)
The problem is that when I run the model with factor A (categorical) or factor B (dichotomous) it works, but when I add both factors A and B to the model, I get the following message: "The Hessian matrix is singular. Some convergence criteria are not satisfied etc. etc." and the output only shows a Iteration history.
Can anyone tell me if there is anything I should take into account when adding more than 1 factor to the model and how this problem can be solved?
Thanks a lot in advance