1. Author suggests us using expert's judgments. As xuelan's mention, actually the cognitive processes is impossible to be known.
2. really a problem
3. Is it a latent regression as the design matrix known?
4. yes, author didn't mention this in this artical even if in her dessertation in 2012.
5. see wolfe 1981, I think it is the point let the probability in 3P normal ogive equal to 0.5
6. I don't know now. it need to spend time to calculate.
8. yes, prior should be correct.
9. Do you think whether the lower information can be claimed a bad result under the information are correct calculated as family information?