Validation of the Vocabulary Size Test

Authors : Mustafa GÖKCAN, Derya ÇOBANOĞLU AKTAN

Pages : 305-327

Doi:10.21031/epod.1144808

View : 54 | Download : 21

Publication Date : 2022-12-25

Article Type : Research Paper

Abstract :The Vocabulary Size Test insert ignore into journalissuearticles values(VST); is one of the most commonly used assessment tools for measuring English vocabulary size in the field of language testing. Despite its common usage, only a limited number of validity and reliability studies have been carried out with regard to the VST. Besides, they were mostly predicated on the Rasch model. This validation study has attempted to reveal evidence for construct validity for the VST, and to this end, item response theory insert ignore into journalissuearticles values(IRT); analyses were performed based on the three-parameter logistic model insert ignore into journalissuearticles values(3PLM);. The assumptions of IRT were investigated via factor analysis insert ignore into journalissuearticles values(unidimensionality); and Yen’s Q3 statistic insert ignore into journalissuearticles values(local independence);. Detailed differential item functioning insert ignore into journalissuearticles values(DIF); analyses were conducted with Mantel-Haenszel, Lord\`s chi-square test, and Logistic regression methods to add evidence based on internal structure and to check fairness as a lack of measurement bias. The validation results with IRT showed that the 3PLM fitted the data better than the one- and the two-parameter logistic models. DIF results indicated that 10 items exhibited large DIF insert ignore into journalissuearticles values(seven favoring males and three favoring females);. The results further showed that the guessing effect was not negligible for the VST.
Keywords : Language testing, vocabulary assessment, Vocabulary Size Test, item response theory, differential item functioning

ORIGINAL ARTICLE URL