Rasch Model Validity and Reliability Test of the Practicality of a Chemistry Learning Module Using the POE (Predict, Observe, Explain) Learning Model Integrated ChatGPT on Colligative Properties of Solutions Material

Abstract

Students' difficulties in understanding chemistry concepts have driven research efforts utilizing the development of artificial intelligence (AI) to support students' independent learning and facilitate comprehension of chemistry material. This study is a type of Research and Development (R&D) using the Rasch Model method to determine the validity and reliability of the module instrument. The practicality score of the module was 82.5%, indicating that the developed module falls into the "practical" category. Most of the item validity scores were classified as valid and acceptable, as the majority of the items (S1, S2, S4, S5, S6, S7, S9, S12, S13, S14, S15, and S16) met all three validity criteria: MNSQ, ZSTD, and Pt-Mean Corr. Items S8, S3, S10, and S11 met two out of the three validity criteria, with ZSTD and Pt-Mean Corr values falling within the valid and acceptable range. The obtained Cronbach’s alpha value was 0.94, indicating that the module has very high reliability with an effective level of internal consistency