Abstract
In grading examinees' responses to test items, it is not uncommon to find that some examinees omit responses to specific items. The number of omitted responses must be considered in the psychometric analysis of test data. Omitted responses cannot be ignored, as mishandling them can jeopardise the validity of the test. This study investigates the impact of omitted responses on examinee characteristics in Mathematics Achievement Test (MAT), using item response theory (IRT) in Osun State, Nigeria. A descriptive survey research design was employed, with a sample of 600 senior secondary school 3 (SSS 3) students from eight randomly selected schools. The instrument used was a 40-item multiple-choice MAT, adapted from the West African Examinations Council's items, with a reliability coefficient of 0.88. The instrument was content-validated by experts in Mathematics using the Lawshe content validity ratio, giving a 0.82 content validity index. The results indicate significant differences in estimated ability levels among groups, with varying probabilities of examinees producing omitted responses. The study recommends the consideration of omitted responses in IRT-based ability estimation and emphasises the importance of comparable ability groupings. This research contributes to the understanding of the complexities of educational measurement and highlights the need for careful handling of omitted responses to ensure the validity of test inferences. CONTRIBUTION: This study contributes by highlighting the importance of considering omitted responses in MAT, emphasising their impact on estimated ability levels and the validity of test inferences, thus informing fairer assessment practices and enhancing the reliability of educational measurements.