DEVELOPMENT OF A MEASUREMENT TOOL FOR SELF-REGULATION IN ARTIFICIAL INTELLIGENCE USAGE AMONG SECONDARY SCHOOL STUDENTS

Abstract

This study aimed to (1) develop and examine the quality of a self-regulation in artificial intelligence (AI) usage measurement tool for secondary school students, and (2) establish normative criteria for assessing students’ self-regulation in AI usage. The population consisted of 4,892 upper secondary school students under the Phitsanulok Secondary Educational Service Area Office in the 2025 academic year. The sample consisted of 357 upper secondary school students selected through simple random sampling. The research instrument was a 5-point Likert scale questionnaire comprising 30 items measuring self-regulation in AI usage. Data were analyzed using statistical software, including SPSS for Windows and Mplus. The instrument demonstrated high reliability, with an overall Cronbach’s alpha coefficient of 0.948, indicating a high level of internal consistency. Construct validity was examined through Confirmatory Factor Analysis (CFA) using Mplus. The findings revealed that the model fit the empirical data well (Chi-square = 5.615, p = 0.2298, RMSEA = 0.034, CFI = 0.999, TLI = 0.997, and SRMR = 0.010). Factor loadings ranged from 0.772 to 0.884 and were statistically significant at the .01 level. The variables accounted for 59.5% to 78.1% of the variance, supporting the construct validity of the instrument. Furthermore, normative criteria were established using T-scores and percentile ranks, classifying students’ self-regulation in AI usage into four levels: high, moderately high, moderately low, and low. These norms could be effectively applied for student assessment and development