mirt

Analysis of dichotomous and polytomous response data using latent trait models under the Item Response Theory paradigm. Exploratory models can be estimated via quadrature or stochastic methods, a generalized confirmatory bi-factor analysis is included, and confirmatory models can be fit with a Metropolis-Hastings Robbins-Monro algorithm which may include polynomial or product constructed latent traits. Multiple group analysis and mixed effects designs may be performed for unidimensional or multidimensional item response models for detecting differential item functioning and modelling item and person covariates.