So I am doing some Rasch Modelling (simple binary Models and PCM Models) with the eRm-package in R. In the future we may want to use the results to produce feedback for individual experimentees, but (at first) we do not want to recalculate the complete model every time. Thus, I am looking for a possibility to estimate only the person parameters with given item paramters. Has anybody experience with this and maybe even knows a function, preferably from the eRm-package which does the job? I really do not want to implement the whole likelihood-method all by myself.
Asked
Active
Viewed 34 times
0
-
I am not familiar with the package and the methods but from the description, maybe `?eRm:::predict.ppar` is what you are looking for... if not, try providing a reprex, it will increase your chances of getting a better answer (see: https://stackoverflow.com/questions/5963269/) – Marcelo Avila Mar 25 '21 at 15:12
-
Hey thank you for your idea. As far as I see, this method provides the things the other way around. You can insert given person parameters and it return a kind of monte-carlo simulated response data to the questions. I dont think a reprex would really help here, because my quastion is not directly related to a certain programm of mine or anything... – ginger314 Mar 29 '21 at 16:27