Özet:

The present paper is subject to study and compare empirically item measures of Raven Standard Progressive Matrices (RSPM) Test under Classic Test Theory (CTT) and Item Response Theory (IRT) Models. The findings of research showed that estimated Raven items’ difficulty indexes from CTT were very comparable with those from all IRT models and with the three-parameter model. But compared with item difficulty indexes, the item discrimination indexes from CTT were somewhat less comparable with those from IRT. Both CTT and IRT item difficulty indices (especially two- and three-parameter model) exhibited high invariance across samples. Both CTT and IRT item discrimination indexes which did not exhibit invariance across samples were analysed according to Classical-based and Item response based indices respectively. In general, the findings from this empirical investigation failed to discredit the CTT framework with regard to its alleged inability to produce invariant item statistics. On the other hand, if the RSPM will be used in some testing applications, such as computerized adaptive testing, test equating, IRT models will be preferable, especially two- parameter model

Anahtar Kelimeler: