4.2 Features of the GDINA R package

  • Estimating G-DINA model and a variety of widely-used models subsumed by the G-DINA model, including the DINA model, DINO model, additive-CDM (A-CDM), linear logistic model (LLM), reduced reparametrized unified model (RRUM), multiple-strategy DINA model for dichotomous responses

  • Estimating models within the G-DINA model framework using user-specified design matrix and link functions

  • Estimating Bugs-DINA, DINO and G-DINA models for dichotomous responses

  • Estimating sequential G-DINA model for ordinal and nominal responses

  • Estimating polytomous G-DINA model for ordinal attributes

  • Estimating generalized multiple-strategy CDM and diagnostic tree model for multiple strategies

  • Estimating an extended multiple-choice DINA model

  • Modeling independent, saturated, higher-order, loglinear smoothed, and structured joint attribute distribution

  • Accommodating multiple-group analysis

  • Imposing monotonic constrained success probabilities

  • Validating Q-matrix under the general model framework using various approaches

  • Evaluating absolute and relative item and model fit

  • Comparing models at the test and item levels

  • Detecting differential item functioning using Wald and likelihood ratio test

  • Evaluating classification accuracy

  • Simulating data based on all aforementioned CDMs

  • Providing graphical user interface for users less familiar with R