Thanks, this is super helpful! At this point, I have someone helping me with STAN and a deep knowledge of psychometric thresholds, and I think I’m pretty close to being able to figure everything out. But I do think BRMS is a powerful and easier alternative. I’ll definitely play around with your suggestions though and see if I can get it to work.
The person who has been giving assistance is a researcher at another University, and he’s pretty busy, so I’m definitely trying to take on the bulk of figuring out loose ends…getting everything to work on the super computer within the two day limit.
I think the way forward at this point is using cmdstanr and the reduce_sum argument. I looked through the links you provided and couldn’t quite figure it out…and asked the question here (which has been unanswered): Using "Reduce Sum" to Parallelize Code in Multi-level Psychometric Model. Could you lend some advice on where I might be going wrong?