A psychometric comparison of Sato Test Theory with Classical Test Theory and Item Response Theory

In this study, based on the answers given by the students who participated in the Student Achievement Determination Exam, it is aimed to determine the item characteristics of the mathematics and Turkish sub-tests and student achievement scores with Sato Test Theory (STT), Classical Test Theory (CTT) and Item Response Theory (IRT) and to compare the obtained findings. The research was conducted on 15461 people, 8th-grade students. After the analyzes performed, it has achieved a solid posture that the STT indices examined are statistically consistent even in small samples. When the harmony between the different samples of the STT classifications giving information about the functioning of the items is examined, it was determined that nearly all of the Turkish test items were classified with high compliance percentages among 100 and 200 person samples, and all of them were classified with high compliance percentages among 600 person samples. Classification for Turkish test items has reached a result that has a more stable stance than mathematical test items. At the point of determining the individual and items properties, it is seen that the STT has similar results with other theories and very close results with the CTT and IRT were found especially at the point of identifying the problematic items. The results of this research, it supports the claim that the STT can be evaluated as an alternative test theory, which is psychometric consistent in itself and allows valid and reliable measurements to be made with estimates that do not contradict other test theories.