An MCMC Algorithm for Estimating the Q-matrix in a Bayesian Framework

TL;DR

This paper introduces a novel MCMC algorithm for estimating the Q-matrix within a Bayesian framework, incorporating correlated attribute estimation, prior modeling, and relabeling techniques, validated through simulations and real data.

Contribution

The paper presents a new MCMC algorithm that estimates the Q-matrix considering attribute correlations and label switching, advancing Bayesian methods in cognitive diagnosis modeling.

Findings

Algorithm accurately estimates the Q-matrix in simulations.

Effective handling of correlated attributes demonstrated.

Empirical application shows practical utility.

Abstract

The purpose of this research is to develop an MCMC algorithm for estimating the Q-matrix. Based on the DINA model, the algorithm starts with estimating correlated attributes. Using a saturated model and a binary decimal conversion, the algorithm transforms possible attribute patterns to a Multinomial distribution. Along with the likelihood of an attribute pattern, a Dirichlet distribution, constructed using Gamma distributions, is used as the prior to sample from the posterior. Correlated attributes of examinees are generated using inverse transform sampling. Closed form posteriors for sampling guess and slip parameters are found. A distribution for sampling the Q-matrix is derived. A relabeling algorithm that accounts for potential label switching is presented. A method for simulating data with correlated attributes for the DINA model is offered. Three simulation studies are conducted to evaluate the performance of the algorithm. An empirical study using the ECPE data is performed. The algorithm is implemented using customized R codes.