IRTBEMM

IRTBEMM implements estimation procedures for Item Response Theory (IRT) models that incorporate guessing and slipping parameters, supporting Bayesian Expected a Posteriori Maximum Likelihood (Bayesian EMM), Bayesian Expected a Posteriori Three-Parameter Model (Bayesian E3M), and corresponding maximum likelihood methods for 3PL, 4PL, 1PL-G, and 1PL-AG models.

Key Features:

• Supported Models: Implements parameter estimation for the three-parameter logistic model (3PL), four-parameter logistic model (4PL), one-parameter logistic model with guessing (1PL-G), and one-parameter logistic model with both guessing and slipping (1PL-AG).
• Bayesian Estimation: Provides Bayesian Expected a Posteriori Maximum Likelihood (Bayesian EMM) and Bayesian Expected a Posteriori Three-Parameter Model (Bayesian E3M) estimation procedures.
• Maximum Likelihood Estimation: Includes maximum likelihood counterparts to the Bayesian EMM and E3M methods.
• Guessing and Slipping Parameters: Explicitly models guessing and slipping effects in item response data.
• Implementation Platform: Distributed as an R package for computational estimation within the R environment.

Scientific Applications:

• Educational Testing: Estimation of item and examinee parameters in multiple-choice assessments where guessing and slipping occur.
• Psychometrics: Parameter recovery and model fitting for psychometric analyses involving 3PL, 4PL, 1PL-G, and 1PL-AG models.
• Empirical Test Data Analysis: Application to real-world datasets for rigorous estimation of guessing and slipping effects in test responses.

Methodology:

Uses Bayesian Expected a Posteriori Maximum Likelihood (Bayesian EMM), Bayesian Expected a Posteriori Three-Parameter Model (Bayesian E3M), and their maximum likelihood counterparts for parameter estimation in the specified IRT models.