I am running a multilevel mixed-effects negative binomial regression and have two questions I'd like to get some feedback on, if possible. I am primarily interested in random slopes for a continuous variable called ideology that varies with a group variable called category.
I first tried including two random intercepts, id (for my individuals) and category, and a random slope for the variable ideology. These are non-nested random intercepts, so I tried crossed random-effects models, but they never achieved convergence (tried multiple ways).
1) Is it OK to include the id variable in the fixed portion of the model as a series of dummies? It converges this way.
2) When I run the model that way, the random slopes for ideology (that vary per category, the intercept) give me what I expected, which is great. What would be the best way to illustrate the differences between these slopes after running a menbreg model? I have the slopes and standard errors. Should I just show that? Should I calculate confidence intervals or something else?
Thanks
0 Response to Multilevel mixed-effects negative binomial regression
Post a Comment