Applying Item Response Theory Models to Entrance Examination for Graduate Studies: Practical Issues and Insights

Authors : Okan BULUT

Pages : 0-0

Doi:10.21031/epod.17523

View : 123 | Download : 12

Publication Date : 2016-01-02

Article Type : Other Papers

Abstract :Item response theory is a psychometric framework for the design, analysis, and scaling of standardized assessments, psychological instruments, and other measurement tools. Despite its increasing use in educational and psychological assessments across many countries around the world, it has not been applied to any large-scale assessment in Turkey. The purpose of this study is to investigate the fit of unidimensional item response theory models to the Entrance Examination for Graduate Studies which is a high-stake large-scale assessment in Turkey required for applying to graduate programs in Turkish universities. Model assumptions of item response modeling, such as unidimensionality, local independence, and measurement invariance, are examined. Also, model-specific assumptions, such as equal item discrimination and minimal guessing, are evaluated. Findings of this study suggest that the three-parameter IRT model shows the best model-data fit for the Entrance Examination for Graduate Studies. Also, the results of this study highlight potential issues that need to be addressed, such as high omit rates, speededness of the test, and aberrant guessing behaviors.  
Keywords :

ORIGINAL ARTICLE URL