Research on mathematics self-efficacy in Indonesia has been widely conducted; however, studies that specifically examine the quality of instruments through rigorous parameter testing remain limited. This study aims to identify valid and reliable items of a mathematics self-efficacy instrument for junior high school students using the Rasch model. The participants consisted of 401 junior high school students from various public and private schools in Serang City, selected using a multistage random sampling technique. Data analysis was conducted using the Rasch model with the Rating Scale Model (RSM) approach, supported by JMetrik software. The results showed that out of 37 analyzed items, 32 items were found to be valid and fit the Rasch model, while 5 items were eliminated due to misfit. The 1–4 rating scale functioned optimally with no indication of category disordering. The reliability indices were classified as very high, both for items (0.99) and persons (0.85), indicating good measurement consistency. Furthermore, the instrument demonstrated a good ability to differentiate students’ levels of mathematics self-efficacy, particularly in the low to moderate range. Therefore, the 32-item mathematics self-efficacy instrument is considered appropriate for effectively and objectively measuring junior high school students’ confidence in learning mathematics.
Copyrights © 2026