DEVELOPMENT OF LEARNING ACHIEVEMENT IN STOICHIOMETRY USING A REFLECTIVE LEARNING PROCESS INTEGRATED WITH ARTIFICIAL INTELLIGENCE (GEMINI AI) FOR GRADE 10 STUDENTS

Abstract

The purposes of this research were to: 1) compare student learning achievement in stoichiometry before and after implementing a reflective learning process integrated with Gemini AI, 2) study student learning development using relative gain scores, and 3) investigate student satisfaction toward the learning management. The research population, obtained through cluster random sampling, consisted of 68 Grade 10 students from two classrooms at Matthayom Prachaniwet School during the second semester of the 2025 academic year. The research instruments included reflective learning lesson plans integrated with Gemini AI, prompt tuning guides for personalized learning, achievement tests, and satisfaction questionnaires. All instruments were validated for content validity (IOC), difficulty index, and discrimination power. Data were analyzed using mean, standard deviation, t-test for dependent samples, and relative gain scores.

The results revealed that: 1) student learning achievement after the intervention was significantly higher than before at the .05 level. 2) The average learning development (gain score) was at a moderate-to-high level. The reflective process integrated with Gemini AI, acting as digital scaffolding, enabled students to identify misconceptions, recognize errors in their calculation steps, and independently improve their cognitive processes. 3) Overall student satisfaction was at the highest level (= 4.75, S.D. = 0.43) reflecting the creation of a safe learning space that reduced anxiety and supported personalized learning potential.