Hi I am working on a multiple membership model with brms. My data are individuals (level-1) nested in several social media posts (level-2) so I am looking at the effects of posts on individuals. The difficult part is I have more than 2000 level-2 groups (posts), which gave me the error:
Error: protect(): protection stack overflow
I am wondering if there is any upper limit for number of groups that I can model with brms?
Another question is, when I try a few groups (e.g., 2), I have the error:
Error in FUN(X[[i]], …) : Stan does not support NA (in W_1_1) in data
failed to preprocess the data; sampling not done
Is that because the sum of weights of the incomplete groups is not 1? Because based on my preprocessing of the data, the sum of weights of the 2000+ groups would be 1. I am wondering where this error comes from.
Thanks so much for your attention.