The irtQ R package: a user-friendly tool for item response theorybased test data analysis and calibration

Area of Science:

• Psychometrics
• Educational Measurement
• Statistical Software

Background:

• Computerized adaptive testing (CAT) is prevalent in high-stakes health professions exams for efficiency and precision.
• Item response theory (IRT) is fundamental to CAT, enabling dynamic item selection based on examinee ability.
• Accurate parameter estimation is crucial for effective CAT implementation, requiring robust software solutions.

Purpose of the Study:

• Introduce the irtQ R package for simplified IRT-based analysis and item calibration.
• Provide essential tools to support the development and operation of CAT.
• Demonstrate the utility of irtQ for IRT applications, including test data analysis and ability estimation.

Main Methods:

• Utilizes marginal maximum likelihood estimation via the expectation-maximization algorithm for parameter estimation.
• Implements fixed item parameter calibration and fixed ability parameter calibration for pretest item calibration.
• Enables computation of item and test characteristic curves and information functions.

Main Results:

• The irtQ package offers a user-friendly environment for unidimensional IRT models.
• Demonstrates the package's utility through examples using simulated datasets for IRT applications.
• Facilitates efficient adaptive testing research and operations by enhancing IRT analysis capabilities.

Conclusions:

• The irtQ R package provides essential functionalities for IRT analysis, supporting CAT development.
• Its features enhance the capacity for efficient adaptive testing research and operations.
• The package has broader applicability to the development and operation of IRT-based assessments.