PERFORMANCE COMPARISON OF COMPUTER-BASED TEST AND PAPER-PENCIL TEST USING SUPPORT VECTOR MACHINE

Abstract

The Joint Admissions and Matriculation Board (JAMB) have over the years been in the news over the use of Computer Based Test (CBT) mode over the Paper Pencil Test (PPT) mode for its Unified Tertiary Matriculation Examination (UTME). This study examines the two test modes, and also tries to ascertain which particular mode makes the unified tertiary matriculation examination more comfortable for the students in privileged environment and those in the rural areas. Predicting student performance can be useful to the managements in many contexts. The purpose of this research work is to do a performance evaluation of computer-based and paper-based version of Joint Admissions and Matriculation Board (JAMB) test data conducted in the previous year using a robust Support Vector Machine model. This work attempts to determine whether there is any difference in the performance of student when comparing CBT to identical PPT test mode and also investigate the levels of malpractices involved in both test mode. Experimental results demonstrated CBT has better predictive accuracy and root mean square error compared to PPT