Evaluating Anchor-Item Designs for Concurrent Calibration With the GGUM

Area of Science:

• Psychometrics
• Educational Measurement
• Item Response Theory

Background:

• Concurrent calibration with anchor items is efficient for developing large item banks for continuous testing.
• Research on concurrent calibration for ideal point item response theory (IRT) models, specifically the generalized graded unfolding model (GGUM), is limited.
• Existing anchor-item designs effective for dominance IRT models may not directly translate to ideal point IRT models.

Purpose of the Study:

• To develop software for concurrent calibration using the generalized graded unfolding model (GGUM).
• To compare the effectiveness of different GGUM anchor-item designs for parameter recovery.
• To provide practical guidelines for practitioners using ideal point IRT models in computer adaptive testing (CAT).

Main Methods:

• A Monte Carlo simulation study was conducted.
• Three different anchor-item designs were evaluated.
• Parameter recovery was assessed under vertical and horizontal linking scenarios for the GGUM.

Main Results:

• The block-interlaced anchor-item design demonstrated superior parameter recovery across most conditions.
• The study identified specific anchor-item designs that perform well within the GGUM framework.
• Empirical evidence supports the efficacy of concurrent calibration for ideal point IRT models.

Conclusions:

• Concurrent calibration with the GGUM is a viable and effective method for creating large item banks.
• The block-interlaced design is recommended for practitioners implementing anchor-item strategies with the GGUM.
• Findings inform best practices for pretest designs in ideal point computer adaptive testing (CAT).