Just like in the case of Bayesian Analysis, Stata 14 includes an all-new 199-page Stata Item Response Theory Reference Manual which documents how Stata can fit models for binary, ordinal or categorical items. For binary items: Stata 14 includes one-parameter logistic (1PL), two-parameter logistic (2PL) and three-parameter logistic (3PL). For ordinal items: Stata 14 includes graded-response models, rating-scale models and partial-credit models. For categorical items: Stata 14 includes nominal response models. Hybrid models where different items use different models are also supported.
A selected list of available models and features in Stata 14 for IRT models:
Available Models | Available Features |
Binary items: 1PL / 2PL / 3PL | Binary, categorical and ordinal outcomes |
Categorical items: nominal response | Control over item grouping and sorting in output |
Ordinal items: graded response / rating scale / partial credit | Complete suite of postestimation graphs: item characteristic curves, category characteristic curves and item information functions |
Control panel that guides users through analysis |