This study aimed to develop a Predict–Observe–Explain (POE)-based diagnostic test on corrosion concepts to identify students’ mental model profiles. The research employed a Design and Development Research (DDR) approach, involving design, development, and limited evaluation through expert validation, readability testing, and limited trials. The instrument consisted of POE-based items supported by images, videos, and observation tables, integrating macroscopic, submicroscopic, and symbolic representations. The results indicated good content validity based on evaluations from five experts, with revisions made to the third indicator and several items. The instrument was subsequently refined based on expert suggestions before undergoing readability and reliability testing. Readability results showed that all items were highly comprehensible, with an average score above 3.7. Reliability testing using Cronbach’s Alpha yielded a coefficient of 0.917, indicating very high reliability. Therefore, the developed instrument is valid, reliable, and suitable for identifying students’ mental model profiles.
Copyrights © 2026