Klasik Test Kuramı ve Genellenebilirlik Kuramından Puanlayıcılar Arası Tutarlılığın Farklı Yöntemlere Göre Karşılaştırılması

Abstract

In this research, analytical and holistic rubrics, which evaluate the problem solving ability of seventh grade students,were used to be scored by five different mathematics teachers. The reliability analysis of obtained scores from raters have been made with respect to Classical test theory and Generalizability theory and interrater agreement level has been examined. Whether there is any difference between reliability coefficents and interrater agreement level obtained from different techniques from two theories and which tecniques used to gather more information was determined.Consequently, it was found that the obtained reliability coefficients with respect to classical test theory and generalizability theory are parallel and relatively high. However, the kappa statistics states middle level agreement.