| Summary: | This study aims to compare KR-20 reliability levels of “Paper and Pencil Test” developed according to Classical Test Theory and “Individualized Test” developed according to Item Response Theory (Two-Parameter Logistic Model), and the correlation levels of skill measurements obtained via these two methods in a group of students. Individualized test developed in accordance with the Two-Parameter Logistic Model was applied by means of a question pool consisting of 61 multiple-choice items which can be answered in 13 steps. On the other hand, a paper and pencil test of 47 multiple-choice items was applied to the sample student group. After the test developed according to these two methods was applied to the same group, KR-20 reliability coefficient was calculated as 0.67 for the individualized test and as 0.75 for the paper and pencil test prepared according to Classical test theory. Calculated KR-20 reliability coefficients obtained from the study were converted into Fisher Z and tested at the significance level of 0.05. No meaningful difference was detected at the 0.05 significant difference level between the two KR-20 reliability coefficients obtained from the two methods. Pearson Product-Moment Correlation Coefficient was calculated as 0.36 between the points of the individualized test and the measurement results of the paper and pencil test. A positive yet low correlation was observed between the measurement results obtained from the tests developed according to both methods. Consequently, it was seen that at the 0.05 significance level there was no statistically significant difference between KR-20 reliability coefficients of the tests developed according to the two methods and that there was a low correlation between the skill measurements of the students in both tests, but there was no significant correlation at the 0.05 significance level between the skill measurements obtained from both tests.
|
|---|
