Using IRT Psychometric Analysis in Examining The Quality of Junior Certificate Mathematics Multiple Choice Examination Test Items

This study aims at providing the psychometric analysis of 2010 Botswana mathematics JC paper 1 in determining the quality of the junior certificate mathematics multiple choice examination test items. The mathematics paper 1 consisted of forty (40) multiple choice test items which was constructed using the three year JC mathematics curriculum. The population for the study was all the 36,940 students who sat for the JC mathematics examination in 2010, out of which a sample of 10,000 was selected randomly by the use of SPSS computer software. The students’ responses were analysed using IRT (3PL) model to examine the psychometric parameter estimates of the forty test items which were: item difficulty, item discrimination, and the guessing value. The item characteristics curves were also generated for each test item that fitted the IRT (3PL) model. Twenty three (23) items fitted the 3PLM out of the forty (40) items, and were used in examining the psychometric qualities of the JC mathematics test paper 1.The findings from this study indicated that out of the twenty three (23) items that fitted the IRT model, twelve (12) items were classified as poor test items, ten (10) items were classified as fairly good test items which could be revised or improved and one (1) item was considered to be good test item. It was therefore recommended that examination bodies should consider improving the quality of their test items by conducting IRT psychometric analysis for validation purposes.