TESTING THE ASSUMPTIONS OF DIMENSIONALITY, LOCAL INDEPENDENCE AND MONOTONICITY OF CHRISTIAN RELIGIOUS STUDIES MULTIPLE CHOICE QUESTIONS

This study tested the assumptions of Item Response Theorey (IRT) (dimensionality, local independence and monotonicity) of WAEC Christian religious studies multiple choice questions. Instrumentation Research Design was adopted for the study involving 1500 participants. The participants were drawn through multistage sampling procedure. Christian Religious Studies Multiple Choice Questions (CRSMCQ) was the instrument used for data collection. The internal consistency reliability coefficient of the instrument was determined to be 0.81 using Kudder Richardson formula (K-K20). The data collected was analyzed using Normal Ogive Harmonic Analysis Robust Method (NOHARM) and Root Mean Square Reduction (RMSR). The findings of the study showed that the test items of West African Senior School Christian Religious Studies Examination have three dimensions. The test items have good guessing factor index as such the correct response to one does not provide clue to correct response to other items. The study shows that some of the items did not violate the Item Response Theory assumption of item local independence. Teachers and examination bodies should determine the number of dimensions underlying test to understand the number of traits a test is designed to measure.