Generating orthogonal latent variables

8

I have to think about it a bit more but this may be fine. Since these are latent constructs the location and scale information is arbitrary (really only care about relative differences between individuals), and in the IRT setup there are item slopes and intercepts which translate the location and scale information of the latent factors to the location and scale of the observed data. I might need to modify the priors for those slopes and intercepts to make sure they still makes sense with QRed factors though.