Use of the Rasch Model in Analyzing Evaluation Questions in End of Year Mathematics Lesson Assessments

The quality of education in Indonesia still needs to be improved by evaluating on the boutiques and student answer sheets. So this research aims: (1) Clarifying the End of Year Mathematics Lesson Assessments (EYA)'s sheets based on the Bloom's Taxonomy; (2) Counting percentages on the boutiques and the correct answers based on the Bloom's Taxonomy; and (3) Analyze the EYA's grain based on the Winsteps app. This research is a qualitative research with therapeutic methods. Data is the documentation method form 35 boutiques of EYA mathematical subjects and 277 sheets of student answers class VIII in teaching 2021/2022. The research results: (1) There are 11 Lower Order Thinking Skills (LOTS) grains, 16 Middle Order Thinking Skills (MOTS) grains, and 8 HOTS grains; (2) Percentage of grains resulting is 31% of LOTS, 46% of MOTS, and 23% of the HOTS with percentages of the correct answers is 51.5% on LOTS, 42% on MOTS, and 35% on HOTS; and (3) Results of analysis using Rasch model obtained 31 of 35 will have been valid with 4 remaining needs revised, the reliability includes the category is sufficient with the alpha value of 0.72, has a sufficiently diverse level of disassembling, the power of the distinguishing is good, and the effectiveness of its refining has 100% function .