Closed state-o-flux closed 3 years ago
You can just apply the 1.7 constant after the model has been fitted to each parameter estimate (where applicable). There's really no need to do this in the estimation engine itself.
Thanks, this works but then it won't be applied when running fscores(mirt_model) as the inherited model won't actually have changed. No?
The factor score computations would be invariant to a rescaling parameter like D. So even if you included this at the time of estimation the factor scores would be exactly the same under either parameterized scaling.
I'm wondering if it's possible to apply the scaling D factor to alpha parameter estimates in order to estimate the logistic item response model.