mirt: Multidimensional Item Response Theory Analysis Models

mirt is a statistical package for analyzing discrete response data using unidimensional and multidimensional Item Response Theory models, as described by Phil Chalmers in the Journal of Statistical Software (2012). It supports exploratory and confirmatory item factor analysis, bi-factor and two-tier models for testlets, and multiple group analyses for differential item functioning. The package also includes latent class models like DINA and DINO, mixture IRT models, and probabilistic unfolding models.

Use Cases

• Estimating item parameters and latent traits based on unidimensional and multidimensional IRT models.
• Conducting exploratory and confirmatory item factor analysis using EM or MHRM methods.
• Modeling item testlets and local dependencies using confirmatory bi-factor and two-tier models.
• Detecting differential item or test functioning through multiple group analyses and mixed effects designs.
• Fitting latent class and mixture IRT models, such as DINA and DINO, for cognitive diagnosis.

Strengths

• Implements a wide range of IRT models including bi-factor, two-tier, and latent class models.
• Supports both exploratory and confirmatory analysis frameworks as described in the source publication.
• Includes methods for differential item functioning and modeling of item and person covariates.

Limitations

• Row count and dataset size are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Phil Chalmers, as published in the Journal of Statistical Software (2012).

Collection Method

Likely contains simulated or example data for demonstrating the mirt R package's statistical models.