A simulation study of repeated covariate equating

0580479 - ÚI 2024 RIV BE eng A - Abstrakt
Vařejková, Michaela - Martinková, Patrícia - Potužníková, E.
A simulation study of repeated covariate equating.
The 10th European Congress of Methodology (EAM2023) Book of Abstracts. Ghent: Ghent University, 2023. s. 267-267.
[EAM2023: European Congress of Methodology /10./. 11.07.2023-13.07.2023, Ghent]
Grant CEP: GA ČR(CZ) GA21-03658S; GA TA ČR(CZ) TL05000008
Institucionální podpora: RVO:67985807
Klíčová slova: test equating * simulation study * psychometrics * educational measurement
Obor OECD: Education, general; including training, pedagogy, didactics [and education systems]
https://eam2023.ugent.be/images/eam2023_abstracts_book.pdf

ZÁKLADNÍ ÚDAJE: The 10th European Congress of Methodology (EAM2023) Book of Abstracts. Ghent: Ghent University, 2023. s. 267-267. [EAM2023: European Congress of Methodology /10./. 11.07.2023-13.07.2023, Ghent]. ABSTRAKT: When performing test equating with non-equivalent groups and without an anchor test, one potential solution how to adjust for group differences is to substitute the anchor test score with covariates, such as grades or scores from another test. A key assumption for this approach is that the conditional distribution of test scores, given the covariates, is the same in all groups. In this work, we conduct a simulation study to investigate how different types of violations of the same conditional distribution assumption can affect the resulting equated scores. We consider two non-equivalent groups differing in ability. As covariates, we use two binary variables (which can, for example, refer to student’s educational status and school type) and one continuous variable referring to a score from another test. The continuous variable was generated to be correlated with the other two binary variables. In the simulations, we explore several scenarios differing in sample size and the relationship between the test score and the continuous covariate. We show that if the assumption of equal conditional distribution is not met due to the fact that the covariate itself is measured using different test forms, the accuracy of the resulting equated scores can be improved by equating the covariate before incorporating it into the primary test scores equating algorithm.
Trvalý link: https://hdl.handle.net/11104/0349249