As a beginner in modeling, I’ve encountered a situation where, after running the model and checking it with MCMC, I observed a strong positive correlation between two parameters that represent sensitivity to positive or negative outcomes. I’m wondering how I can address this issue or what steps I can take to resolve this correlation problem.
Any guidance or suggestions would be greatly appreciated.
Howdy!
It would help if you could provide more information about your model and the Stan code. Based on the pairs plots, the first thing that I would do is go back and check your data and code to make sure that there wasn’t a coding error and you aren’t using the same column of data twice in the model under different names.