GDINA: An R Package for Cognitive Diagnosis Modeling

Abstract

Cognitive diagnosis models (CDMs) have attracted increasing attention in educational measurement because of their potential to provide diagnostic feedback about students' strengths and weaknesses. This article introduces the feature-rich R package GDINA for conducting a variety of CDM analyses. Built upon a general model framework, a number of CDMs can be calibrated using the GDINA package. Functions are also available for evaluating model-data fit, detecting differential item functioning, validating the item and attribute association, and examining classification accuracy. A grapical user interface is also provided for researchers who are less familar with R. This paper contains both technical details about model estimation and illustrations about how to use the package for data analysis. The GDINA package is also used to replicate published results, showing that it could provide comparable model parameter estimation.