The random-effect DINA model

Hung-Yu HUANG, Wen Chung WANG

Research output: Contribution to journalArticle

10 Citations (Scopus)

Abstract

The DINA (deterministic input, noisy, and gate) model has been widely used in cognitive diagnosis tests and in the process of test development. The outcomes known as slip and guess are included in the DINA model function representing the responses to the items. This study aimed to extend the DINA model by using the random-effect approach to allow examinees to have different probabilities of slipping and guessing. Two extensions of the DINA model were developed and tested to represent the random components of slipping and guessing. The first model assumed that a random variable can be incorporated in the slipping parameters to allow examinees to have different levels of caution. The second model assumed that the examinees’ ability may increase the probability of a correct response if they have not mastered all of the required attributes of an item. The results of a series of simulations based on Markov chain Monte Carlo methods showed that the model parameters and attribute-mastery profiles can be recovered relatively accurately from the generating models and that neglect of the random effects produces biases in parameter estimation. Finally, a fraction subtraction test was used as an empirical example to demonstrate the application of the new models. Copyright © 2014 by the National Council on Measurement in Education.
Original languageEnglish
Pages (from-to)75-97
JournalJournal of Educational Measurement
Volume51
Issue number1
Early online dateMar 2014
DOIs
Publication statusPublished - 2014

Fingerprint

effect
Markov chain
education
simulation
test
attribute
parameter
method
parameter estimation

Citation

Huang, H.-Y., & Wang, W.-C. (2014). The random-effect DINA model. Journal of Educational Measurement, 51(1), 75-97.