Validation of the Indonesian Version of the Student-Teacher Relationship Scale Using Rasch Analysis

Abstract

This study aims to evaluate the structural validity of the Indonesian version of the Student-Teacher Relationship Scale (STRS) using Rasch Model analysis. While STRS has been validated in several countries, validation in the Indonesian context with a large sample and using Rasch analysis remains limited. The data for this study involved 1,927 junior and senior high school students in Indonesia. The analysis technique used was the Rasch Model with Winsteps, assessing unidimensionality, local dependency, misfit items, and differential item functioning (DIF). The results indicate that STRS-ID demonstrates excellent reliability, with person reliability (0.96), item reliability (0.99), and a Cronbach's Alpha of 0.97. Unidimensionality was achieved with 57.7% of the raw variance explained. However, several misfit items were identified, with ZSTD values falling outside the range of -2.0 to +2.0, as well as DIF on some items by gender, age, and region. This study suggests that revisions are needed for problematic items to improve the measurement accuracy and effectiveness of STRS-ID in the Indonesian context. The contribution of this research lies in providing a psychometrically valid tool for measuring the student-teacher relationship in Indonesia and offering recommendations for refining items that do not meet the expected thresholds.

Keywords: DIF, rasch model, relationship scale, reliability, unidimensionality, structural validity, student-teacher misfit items