<![CDATA[This study aims to assess the psychometric quality of mathematics evaluation tools using the Rasch Model. The analysis was conducted on 15 test items administered to university students, item difficulty, model fit, reliability, and test information function. The findings of the analysis show that the item difficulty level ranges from −1.50 to 1.75 logits, indicating sufficient variation in difficulty to measure abilities from low to high. Most items show good model fit with Mean Square INFIT and OUTFIT values within an acceptable 0.5–1.5. Item reliability was very high (0.99), indicating stability in the item difficulty hierarchy, while individual reliability was in the moderate category (0.60), reflecting the homogeneity of the respondents' abilities. The test's information function peaks in the ability range of approximately θ = 0 to θ = +1, where the lowest measurement error occurs in that range, making this instrument most accurate in assessing abilities that are average to slightly average. These findings are consistent with Rasch theory and previous research. Overall, the results of this study reinforce that the Rasch Model is effective for assessing and improving mathematics evaluation tools in higher education and provide for the development of more accurate instruments in future research]]>
Copyrights © 2026
