A Comparison of Calibration Methods and Proficiency Estimators Based on Item Response Theory in Vertical Scaling


Creative Commons License

Altun A., KELECİOĞLU H.

HACETTEPE UNIVERSITESI EGITIM FAKULTESI DERGISI-HACETTEPE UNIVERSITY JOURNAL OF EDUCATION, vol.31, no.3, pp.447-460, 2016 (ESCI) identifier

Abstract

In this study the comparison of the vertical scales, which are obtained through the combination of separate and concurrent calibration based on item response theory and Expected A Posteriori,, Maximum A Posteriori and Maximum Likelihood proficiency estimation methods, take place. For this comparison firstly, math tests for the 6th, 7th and 8th grades were composed from the questions asked in SBS (high school entrance exam) between 20082010 years. Then, the vertical scales were developed by using the scores obtained from the answers of 503 6th grade, 502 7th grade and 500 8th grade students to the math tests in different difficulty levels. By using the scale scores which were obtained from these scores placed in the same scale with the vertical scaling process, means and the difference between the means, effect sizes and horizontal distances were calculated so it was tried to be determined how the choices in scaling process affected the vertical scaling