Proving Item Quality For Measuring Students Critical Thinking Skills In Natural Science Subject: A Rasch Model Measurement Analysis

Abstract

Critical thinking skills (CTS) are important skills in the 21st century and are part of the school curriculum in Indonesia. However, validation of CTS instruments often relies only on content validation, while psychometric validation such as the Rasch model is rarely used. This study aims to validate a critical thinking test using the Rasch model to assess data suitability, measure the CTS of junior high school students in Surabaya, and identify item bias based on gender and school ranking. Data were obtained from 143 junior high school students with 10 multiple-choice questions. The results of the analysis showed that female students tend to have higher CTS than male students, while students from high-ranking schools showed a better distribution of abilities compared to moderate and low-ranking schools. Differential item functioning (DIF) analysis identified one item (EC2) that showed gender bias, and seven items (EC1, EC2, BS1, I1, I2, AC1, ST2) showed DIF based on school ranking. These findings suggest that although this test is generally fair for measuring CTS, there are inequities that need to be addressed to ensure the test is more inclusive and fair for all groups of students. The researcher hopes that this study contributes as a case of reference for future research in assessing CTS in junior high school students