Validation of Dyscalculia Risk Detection Instrument in Grade 3 Elementary School Students Using the Rasch Model: Psychometric Analysis and Implications of Inclusive Education

This study aims to evaluate the psychometric quality of a dyscalculia risk screening instrument designed for grade 3 elementary school students. A total of 147 students from four elementary schools in Tangerang were involved as respondents. The instrument consisted of 10 questions and was analyzed using the Rasch model Item Response Theory (IRT) approach with jMetrics software. The results of the analysis show that the items have a fairly wide range of difficulty levels (logit –2.53 to +2.84), but the distribution is not evenly distributed. Some items show misfits for the Rasch model, as well as high error standards on extreme items. Estimated participant ability (theta) shows a logite distribution from –4.62 to +4.55, with measurement reliability at the item level very high (0.9659), but at the moderate participant level (0.5984). The correlation coefficient between the raw score and the theta estimate of 0.9861 supports the validity of the instrument construct. These findings recommend improvements to the distribution of item difficulty levels and revisions to non-model items so that the instrument can be used effectively in inclusion education for early screening of dyscalculia risk.