Use of the Rasch model for the psychometric analysis of a diagnostic test of mathematics in primary education: The case of 6th grade students,

The study assessed the assumptions of the Rasch model to establish the validity and reliability of a mathematical diagnostic test with two main objectives: understanding and application. The researcher chose a sample of 296 students from four Moroccan primary schools. The methodology rigorously assessed the Rasch model’s assumptions, such as unidimensionality, local dependence, pseudo-chance, discrimination, reliability, and item fit quality. Descriptive and psychometric analyses were carried out by the researcher to evaluate the quality of each test item. The results indicated that 78% of the items met the unidimensionality criterion, while 15% showed local dependence, suggesting interaction issues between certain items. The difficulty of the items was well distributed across the levels, ranging from hard to easy. However, the researcher removed three 28 items due to suitability and outfit values exceeding 2.5. Additionally, seven students were excluded from the sample due to non-conforming Infit and Outfit values. Although there are three outlier items, most of the test items While three items were identified as outliers, most of the test items fit well with the Rasch model, indicating strong psychometric properties. Adjusting or removing these outliers will enhance the overall validity and reliability of the test. This study underscores the importance of conducting thorough psychometric evaluations of assessments developed by teachers to ensure they accurately measure the skills and abilities of the target student population. The study concludes by suggesting additional research to expand the use of the Rasch model and enhance the quality of educational assessments in Morocco. This type of research would support the country’s broader educational reform initiatives aimed at improving mathematical achievement.