[R-sig-ME] Random effects in multinomial regression in R?

[Stuart Luppescu]

On Sat, 2019-03-23 at 10:03 +0000, Doran, Harold wrote:
> I propose it’s
> often ignored not because the issue is not well-known, but because there
> is not (widely distributed) software that easily implements the types of
> corrections needed in this scenario.

I'm not sure if it's "easily" implemented, but you can do this type of
measurement error model in STAN (mc-stan.org), defining the predictor
as a combination of a fixed component and a random error component,
which are determined by the data source. (My work uses the results from
an IRT model which produces an estimate and standard error for each
person.) Information here:

https://mc-stan.org/docs/2_18/stan-users-guide/bayesian-measurement-error-model.html

-- 
Stuart Luppescu
Chief Psychometrician (ret.)
UChicago Consortium on School Research
http://consortium.uchicago.edu

lupp using uchicago.edu