الملخص الإنجليزي
The purpose of this study was to compare Classical Test Theory (CTT) and Item Response Theory (IRT), which is represented in the Two-Parameter Logistic Model (2-PLM). The relationship between student ability estimated from each theoretical framework was investigated. The study also, assessed item parameters invariance, which are represented in difficulty and discrimination parameters, across different grade levels.
The sample consisted of (3082) students from grades seven, eight and nine, who study in Muscat's Schools, in Oman. The data were collected using Raven's Standard Progressive Matrices Test.
Pearson's correlation coefficient and mean square difference (MSD) were used to investigate the relationship between students' scores that were estimated from each theoretical framework. The same previous statistical methods were used to assess item parameters invariance across different grade levels, for each theoretical framework.
The results showed that, while the order of students classical scores were comparable to their scores estimated from the two-parameter logistic model, values of these two groups of scores were relatively different. The results also showed that, item parameters estimated from the two-parameter logistic model were more invariant than those estimated from classical test theory. In addition, it was found that, the difficulty parameter was more invariant than the discrimination parameter, in both frameworks.
