Comparision of item and test psychometric properties considering chance success

In this study, the effect of chance success in the Turkish test of TYT pilot test was examined by the item and test properties according to Classical Test Theory (CTT) and Item Response Theory (IRT) via statistical analysis techniques. Thanks to the data obtained from a large-scale test, the effects of chance success were examined from different theories and perspectives, and a comparison was made. In accordance with this purpose, a ready-made data set for the 2022-2023 academic year obtained from a private institution was used. The research data were obtained by the dichotomous scoring of the answers given to 40 multiple choice questions by 1140 students selected by sampling method from 3085 students of the Turkish test of the TYT pilot test, as 1 and 0. As a result of the research, mathematical and statistical operations were performed and the effect of chance achievement according to CTT and IRT was examined in terms of substance and test properties. The results of the four models belonging to the two theories were compared in terms of similarities and differences. It was revealed that while there is a statistically significant relationship in terms of the item difficulty index, a significant low relationship was found in the item discrimination index, which was observed and cleared of chance success. Descriptive statistics were obtained on the effect of chance success on test characteristics, and it was concluded that the reliability was high in all four models.