The reliability of high school third grade final exams using Generalizability Theory

The aim of this research was to show the application of Generalizability Theory (GT) in assessing the reliability of final exams’ data. To this end, Persian literature and biology final exams of third grade high school in empirical sciences field were selected which were held at June 2011. Having scientific and practical considerations, the Persian literature course with 61 items and 540 sheets and the biology course with 78 items and 420 sheets were randomly selected from the examination area of one district of Khorramābād city. The correctors of these courses were 9 and 7 individuals, respectively. In Persian literature, items facet with 43.8% and in biology course, the interaction effect of items and students with 64.4% had the highest share in total variance score. The results showed that both students’ scores and the items had high reliability. Also, for each measurement design, D Study was conducted. This research showed that using symmetry principle of the GT, educational researchers could select each of the elements of an educational system as object of measurement and they could estimate its reliability.