I am currently working on my Master thesis and ran a multinomial logistic regression on my data set.
I used "Individual choice" as my dependent variable (with three values; 1=cooperation, 2=defection, 3=boobytrap) and group (1=Chinese, 2=Central Europeans), frame (1= frame 1, 2= frame 2), as well as game theory proficiency (1=No knowledge at all, 2=Very poor, 3=Fairly poor, 4=Mediocre, 5=Fairly good, 6=Very good, 7=Perfect) as my independent variables.
The variables "group" and "frame" are measured nominally, whereas game theory proficiency is measured ordinally.
Now I have been getting the following warning message "Warnings: Unexpected singularities in the Hessian matrix are encountered. This indicates that either some predictor variables should be excluded or some categories should be merged.
The NOMREG procedure continues despite the above warning(s). Subsequent results shown are based on the last iteration. Validity of the model fit is uncertain."
I ran the multinomial logit without the game theory proficiency variable and as a consequence didn't encounter this warning message, so I am guessing the problem must either lie in my dependent variable (categories) having too many values or in the game theory proficiency variable having too many categories.
Would any of you be able to help me out with this problem? I would highly appreciate it!