Comparison of the Methods of Classical Test Theory and Item Response Theory on Vertical Scaling


Cetin E., GELBAL S.

HACETTEPE UNIVERSITESI EGITIM FAKULTESI DERGISI-HACETTEPE UNIVERSITY JOURNAL OF EDUCATION, vol.29, no.3, pp.23-34, 2014 (SSCI) identifier identifier

Abstract

Vertical Scaling is a kind of linking which is used to determine how much the students of adjacent grades or ages have improved in the subject areas such as Maths and Language. The purpose of this research is, as a result of vertical scaling, to establish the pattern of students' improvement in certain levels as class grade increases. The data in this research were obtained from the Achievement Exam (OBBS) results for Turkish primary school students of the 6th, 7th and 8th grades in 2005. Vertical Scaling was conducted based on Classic Test Theory (CTT) and Item Response Theory (IRT). Thurstone (1938) scaling method was used based on CTT, and Expected A Posteriori (EAP) scaling method was used in IRT estimation. As an evaluation criteria of vertical scaling, mean, standard deviation and effect size figures in academic growth were used. At the end of the research, in Thurstone Scaling, standard deviations in Maths and Turkish Tests rise as class levels increase. In the conducts of CTT and IRT, it can be said that standard deviations increase free from the increase or decrease of means.