Teaching science through STEAM projects is increasingly popular in Indonesian elementary schools, yet one persistent challenge is the lack of psychometrically validated assessment instruments that match what actually happens in those classrooms. This study developed and validated a three-part STEAM-based assessment package—a conceptual post-test (17 items), a project performance rubric (15 criteria), and a student response questionnaire (25 Likert items)—for use across three hands-on projects: a Stick Bridge (Grade 4/Force), an Energy-Efficient House (Grade 5/Energy), and a Simple Flashlight (Grade 6/Electricity). Two expert validators assessed content validity using Aiken's V, and 102 students from SDN 2 Tapa and SDN 3 Batudaa in Gorontalo completed the instruments as part of an empirical field trial. Results showed strong content validity (V̄ = 0.812), acceptable reliability for the 16-item valid questionnaire scale (α = 0.732), and a positive student response rate of 76.9%. Rasch Rating Scale Model analysis revealed that all items fell in the negative logit range, indicating a person-item gap that points toward a need for more challenging items in future revisions. A significant difference between schools was found for individual post-test scores (p = 0.018). These findings confirm that the instruments are content-valid, reliable, and practically usable in Indonesian elementary science classrooms, while also providing clear directions for refinement.
Copyrights © 2026