RASCH MODEL ANALYSIS OF MATHEMATICAL REASONING ABILITY ITEMS ON LINEAR EQUATIONS FOR JUNIOR HIGH SCHOOL STUDENTS USING MINISTEP

Authors

  • Widuri Ambaika Universitas Muhammadiyah Cirebon image/svg+xml
  • Desy Lusiyana
  • Rifqi Hidayat

DOI:

https://doi.org/10.31000/ggzamp11

Abstract

The limited availability of valid and reliable assessment instruments for measuring junior high school students' mathematical reasoning ability on linear equations motivated this study. Despite widespread recognition of mathematical reasoning as a core competency in school mathematics, few instruments in this domain have been rigorously validated using modern measurement approaches such as the Rasch Model. This study aimed to analyze the item quality of a mathematical reasoning test on linear equations using the Rasch Model via Ministep software, focusing on item validity, instrument reliability, difficulty index, and person-item mapping. A descriptive quantitative approach was conducted with 30 junior high school students chosen through purposive sampling. Four polytomous essay items were developed based on NCTM’s four reasoning indicators, namely drawing conclusions, proposing conjectures, constructing proofs or arguments, and explaining solution steps. Results showed all four items valid: Outfit MNSQ (0.71–0.97), Outfit ZSTD (−0.34 to 0.15), and Point Measure Correlation (0.82–0.84). Person Reliability was 0.87 (separation index 2.63), classified as good. All items were of moderate difficulty (logit: −0.00 to +0.40). The Wright Map indicated students’ average ability exceeded item difficulty, suggesting the need for harder items in future development. Based on the Rasch analysis, the instrument demonstrates acceptable measurement properties and is deemed appropriate for formative assessment of mathematical reasoning in linear equation topics, while also pointing to clear directions for future instrument refinement.

Downloads

Published

2026-11-17