مقارنة بين النظرية الكلاسيكية و النموذج ثلاثي المعلم في انتقاء فقرات اختبار تحصيلي للغة الإنجليزية

الملخص EN

This study aimed to examine the differences between classical test theory (CTT) and the three-parameter logistic model in terms of selecting items for an English achievement test.

to achieve the aim of the study, an English achievement test consists of (42) multiple-choice with four alternatives items was developed and administered on a sample consisted of (1076) female students randomly selected from (11) female secondary schools.

IRT-PRO4, spss-19, AMOS- 24 and Cocron were used for statistical analysis.

the results revealed that (38) items fit the classical test theory with percentage of (90.45%) while (36) items fit the three-parameter logistic model with percentage of (85.71 %), and (36) items out of (42) fit both theories reliability coefficient (KR–20) for the classical test theory was (0.86) while the empirical reliability coefficient value for the three-parameter model was (0.78).

the criterion-related validity coefficient was (0.59) according to the classical theory, and (0.60) according to the three-parameter model.

the results indicated that there are statistically significant differences between the two reliability coefficients resulting from the two theories in the favor of the classical test theory.

results also revealed statistically significant differences between the two criterion-related validity coefficients resulting from the two theories in the favor of the three-parameter logistic model.

overall, IRT provides richer information about measurement precision.

the findings support the use of IRT for test construction and development in the educational fields.