THE EMPIRICAL EVIDENCE FOR UNIDIMENSIONAL STUDY OF EFFECTIVE TEACHING INSTRUMENT IN THE TRUST SCHOOL PROGRAM (TSP) THROUGH RASCH'S MODEL
JOURNAL OF MECHANICS OF CONTINUA AND MATHEMATICAL SCIENCES
Unidimensionality is one of the key considerations for testing items in meeting assumptions in the Rasch Model measurement. However, these assumptions can sometimes be difficult to meet because many factors need to be taken into account during construction of the item. Therefore, this study focuses on empirical evidence for the unidimensional study of TSP teacher effective teaching instruments using Rasch Model testing. Quantitative survey design was used. A total of 203 TSP teachers were
... teachers were selected through targeted sampling. The research instrument was adapted from the Trust School Teacher Handbook 2018, the Performance Management System for Teachers with fourteen competencies including Seven Pedagogical Pillars. Data were analyzed using Rasch Model framework with Winstep 3.68.2 software. The findings show that the value of raw variants is explained by a 45.2 percent measurement above the 40% confidence interval in the Residual Principal Component Analysis (PCA). In addition, the noise level of the item recorded 5.2 percent. Although the eigenvalues cannot be met to meet the four criteria testing the assumption of unidimensionality, the eigenvalues appear to surpass these assumptions when measured by construction. Overall, this instrument fulfills the assumption of unidimensionality and can be used to measure the effective teaching of TSP teachers. Further studies can be made by comparing the effects of Rasch's unidimensionality in factor analysis.