Development of a Software Package for Computing Psychometric Properties of Multiple Choice Tests Using Norm-Referenced Approach.

Abstract

This work is on the development of a software package for computing the psychometric properties of multiple choice tests using norm-referenced approach. Focus was on item indices which are difficulty, discrimination and distracter indices. There are no easily affordable statistical packages in the Nigerian market for the computation of these indices. Most studies involving them have always been done manually due non-availability of affordable statistical software. Hence, the researcher was motivated to embark on this investigation. Three research questions were formulated to guide this study. Analysis of the existing system was done on BILOG-MG statistics. This study deployed the following methodologies: mathematical, structured systems analysis and design methodology (SSADM), the object-oriented analysis and design methodology (OOADM) and prototyping. The design for the study was instrumentation. The population for the study was 225 students which consisted of all the JS3 students of 2016/2017 session in Port Harcourt, Rivers State. The researcher made use of two instruments for the study. The first was the Rivers State Basic Education Certificate Examination (BECE), 2017 multiple choice questions in Mathematics. The second was a group of system software: HTML, PHP, JavaScript and MySQL. The responses of the students were used to test the effectiveness and efficiency of the software package developed in this study. Data analysis was done first with manual computation and second with the software package that was developed. The manual computation took the researcher 180 hours while the electronic computation with the developed software took only six hours to complete. The findings from the results of the study showed that the developed software was accurate in computing the difficulty, discrimination and distactor indices of the test items and it has a very high rate of accuracy. The study recommended that the developed software should be used to compute the difficulty, distractor and discrimination indices of multiple choice tests.