hello,
I'm attempting to fit an ordered category response model - I'm basically following the steps in Ch 11 using my own dataset but when I switch the estimation method to 2nd order PQL I get the error: design vector at level 2 is the wrong length. I am able to estimate terms using the first order MQL but I get the same error using MCMC
I have 1600 samples (lev 2) with 1 of 5 possible responses (lev 1) and group to which the samples belong as the third and highest level. At this point I have not added any fitted terms to the model. The cons. variable has n=1600 ones and cons.1234 = 6400.
Does anyone have any suggestions about what I might be doing wrong?