The software just closes without warning when attempting MODEL BUILDING - BEST SUBSETS with GLZ LOGIT.
However the software works fine with MODEL BUILDING - ALL EFFECTS, which suggests to me that there is nothing wrong with the dataset.
Maybe there is a bug in that BEST SUBSETS module? If the problem is too many categorical factors, then I would expect the software to at least providing a warning, not just shut down.
I was attempting GLZ LOGIT with the following:
Dependent variable: one
Categorical factors: 32 (two levels each)
Continuous predictors: none
Count variable: none
Cases: about 2,500.
If I reduce the number of Categorical factors to 10, then it does not crash.
So apparently there is some limit to the number of variables Statistica can admit. What is that number?
Hello, there are virtually no limitations in terms of the number of variables or cases in a Statistica spreadsheet. I am not sure what causes this crash based on your description. I would advise you to open a service request, so that one of our support engineers can work with you to troubleshoot this issue. Here is a KB article of how to get support through the Dell support portal.