My R package for dynamic ideal point/IRT models in Stan has reached a milestone with a beta v1.0 release. As it now relies on cmdstanr, it is only available via Github: GitHub - saudiwin/idealstan: idealstan offers item-response theory (IRT) ideal-point estimation for binary, ordinal, counts and continuous responses with time-varying and missing-data inference. Latent space model also included. Full and approximate Bayesian sampling with 'Stan' (www.mc-stan.org).. The great thing of course is that this means idealstan can always use the most up to date Stan via cmdstanr interface.
This is a beta release, so most features are stable, although some have yet to be implemented, in particular a type of marginal effect estimation for latent predictors that will be included in the final 1.0 release. Bug reports are welcome at the Github page above!
Ideal point models are a variant of the item response theory 2-Pl model and are useful in a variety of measurement problems where the aim is to estimate a latent construct from indicators of human behavior, anything from test scores to legislator votes to democracy indexes. You can read more about the package’s approach in this working paper:
And also see a recent application to estimating COVID-19 policy scores: