MCMC Z-G: An IRT Computer Program for Forced-Choice Noncognitive Measurement.

Area of Science:

• Psychometrics
• Educational Measurement
• Organizational Psychology

Background:

• Noncognitive constructs are increasingly measured using Likert scales, which are prone to response distortion.
• Forced-choice formats offer an alternative to mitigate response biases in measurement.
• The Zinnes-Griggs (Z-G) model is a key framework for analyzing forced-choice data.

Purpose of the Study:

• To address the limited availability of software for the Z-G model outside organizational contexts.
• To develop a publicly accessible program for parameter estimation in the Z-G model.
• To facilitate broader application of the Z-G model in educational and managerial research.

Main Methods:

• Development of a Markov chain Monte Carlo (MCMC) estimation program named MCMC Z-G.
• Implementation of a Metropolis-Hastings-within-Gibbs algorithm for parameter estimation.
• Simultaneous estimation of Z-G item and person parameters.

Main Results:

• Successful development of the MCMC Z-G software.
• The program enables parameter estimation for the Z-G model.
• MCMC Z-G is compatible with both Mac OS and Windows platforms.

Conclusions:

• The MCMC Z-G program expands the accessibility of the Z-G model for measuring noncognitive constructs.
• This tool supports researchers in educational and organizational settings using forced-choice measurement.
• Public availability of the software promotes wider adoption and application of advanced psychometric methods.