Comparability of the Paper-and-Pencil and Computerised Adaptive Versions of the General Scholastic Aptitude Test (GSAT) Senior

Abstract

A computerised adaptive test was constructed from two existing parallel paper-and-pencil versions of the General Scholastic Aptitude Test (GSAT) Senior. Achievement in the GSAT computerised adaptive test was compared to achievement in one form of the GSAT paper-and-pencil test. In computerised adaptive testing the program tailors each test to the examinee's ability level. Based on a statistical method known as Item Response Theory (IRT), the program interactively selects test items which are at the appropriate difficulty level for the individual being tested, thereby allowing a considerable reduction in test length without forfeiting measurement accuracy. The study was undertaken to investigate the equivalence of results obtained with three versions of the GSAT: A paper-and-pencil version, a standard computerised version, and a computerised adaptive version. The standard computerised GSAT was included to study the effects of computerization apart from adaptive testing. The results were used to make adjustments to the GSAT computerised adaptive test to ensure equivalent measurement.