Article

Equating Test Scores with Different Linkage Plans Using IRT Observed-Score Equating

---

Abstract

Linkage plans are used when equating different test versions from different administrations of standardized achievement tests. The overall aim is to examine item response theory (IRT) observed-score equating when different linkage plans have been used when the average ability among test takers in different administrations and/or the average item difficulty in different administrations differ as well as if there are more variation in test scores in some administrations and if a harder anchor test is used. The results indicate that direct equating without an equating chain always yields the smallest bias, root mean squared error and standard errors, and the longest equating chain always yielded the highest errors. More differences could be seen when the test forms and or the item difficulty varied more between groups and test forms. The equated values from using different linkage plans, also differed depending on the examined scenario. The results are discussed together with limitations, future research and practical recommendations.

Keywords: NEAT design, item-response theory equating, anchor test, item response theory

How to Cite:

Wiberg, M., (2025) “Equating Test Scores with Different Linkage Plans Using IRT Observed-Score Equating”, Practical Assessment, Research, and Evaluation 30(1): 11. doi: https://doi.org/10.7275/pare.3056