Psychometric properties of teacher classroom assessment literacy instrument using Rasch model analysis

Abstract

A teacher classroom assessment literacy (TeCAL) instrument was developed to measure the level of teacher classroom assessment literacy in schools. TeCAL contains 66 multiple choice items with four options based on four constructs namely purpose, measurement, evaluation and use. Thus, this study aims to identify the psychometric properties of TeCAL using Rasch measurement model (RMM) analysis through Winstep software version 3.72.3. The findings show that the compatibility values of mean square (MNSQ) infit and outfit items ranged from 0.64 to 1.46 and 0.40 to 2.23, respectively. The value of MNSQ outfit was outside the set range, but still met the other fit statistics indicator which has a positive point measure correlation (PTMEA) value. In addition, the findings show that the empirical raw variance explained by measures is 38.2%. It was very close to the modeled value of 38.4% with the empirical unexplained variance in 1st contrast being 7.5% less than the maximum controlled 15%. Largest standardized residual correlations identified 10 pairs of dependent items to be less than 0.7. The person and item reliability index values were 0.94 and 0.89 with separation index values of 2.90 and 3.80, respectively. Overall, this psychometric analysis is crucial to ensure that the TeCAL instrument has good quality and meaningful to use.