Hi, I am not sure how Stan is handling the joint distribution of variables with two different scales.
For example, suppose I’d like to fit parameters to the joint distribution of (choice probability, price).
Price is in dollar terms, but choice probability is in between [0,1].
Then how does Stan “know” that the 1 unit of error term for price is not equivalent to the 1 unit of error term for choice probability?
My estimates show that it is “easier” for Stan to adjust parameters on the price side of things to fit the data, which is not what I want.