Comparing the Efficiency of True and Observed Score Equating Methods in Equating Tests Utilizing the Design with an Anchor Test and Nonrandom Groups

Abstract

The aim of the study was to compare the efficiency of true score equating method and observed score equating method in equating tests Utilizing the design that is based on the anchor test and Nonrandom groups. For achieving that aim, two equivalent forms of a physics test were constructed with (20) items for each one of them, in addition to a group of (10) items used as an anchor test. The population of the study consisted of the secondary stage students/ level three in 2008/2009, while the sample of the study was selected from directories of education in the governorate of karak where the sample of the study consisted of two groups: The first group was given the first form of the test in addition to an anchor test at the first semester. The second group was given the second form of the test in the second semester, and these two groups were considered as two non- equivalent groups. Two equating methods that follow the IRT theory were used, the true score equating method and the observed score equating method and four numbers of items of the anchor test (1, 4, 7, and 10). The results and according to the performance of the non-equivalent groups revealed that there was a statistical significance in the means of the equated scores using the true score equating method and the observed score equating method when the number of items of the anchor test were (1, 4, 7) in favor of the observed score equating method, but the differences had no statistical significance when the anchor test had (1) item. The results also revealed that there was no impact of the number of items of the anchor test by using any of the two methods.