A Rasch Model Analysis of the Suitability of Exam Items in DBS10042 Engineering Science

Abstract

In educational settings, the Rasch Measurement Model is widely used to ensure that assessment instruments produce trustworthy and consistent results. This study applied the Rasch model to analyse the final examination items of the DBS10042 Engineering Science course offered at a Malaysian polytechnic. A total of 240 first-semester engineering students participated in the study. The analysis focused on item fit statistics, item difficulty, and point-measure correlations to determine how well each question aligned with students’ abilities and the intended learning outcomes. Results indicated that most items fell within acceptable fit ranges, confirming their appropriateness for assessing student performance. However, a few items exhibited misfit, suggesting issues such as ambiguity, misalignment with cognitive levels, or excessive difficulty. The study also revealed strong person-item correlations for most items, supporting their diagnostic value in differentiating student proficiency. Overall, the findings highlight the effectiveness of using the Rasch Measurement Model to ensure high quality, fair, and reliable assessments in technical and vocational education contexts.