Perceived Usefulness, Perceived Ease of Use, and Self-‎Reported Critical Thinking Among Indonesian Vocational ‎Students: The Indirect Role of Self-Regulated Learning

Abstract

The spread of generative artificial intelligence (AI) in vocational classrooms has raised a concern that easy, on-demand answers may displace the reasoning vocational graduates are expected to master. This cross-sectional survey examined whether students' perceptions of AI are associated with their self-reported critical thinking, and whether self-regulated learning accounts for any such association. Drawing on social cognitive theory and the Technology Acceptance Model, questionnaire data were collected from 365 students in nine state vocational schools (SMK) in Pesisir Selatan Regency, West Sumatra, and analysed with PLS-SEM. All four constructs, including critical thinking, were measured by self-report rather than by performance tasks. Perceived usefulness and perceived ease of use were not associated with critical thinking directly (β = -.026, 95% CI [-.163, .111]; β = .064, 95% CI [-.072, .200]) but were weakly associated with self-regulated learning (β = .161; β = .158), which was in turn the strongest correlate of critical thinking (β = .383, 95% CI [.284, .482]). Both indirect paths were small and positive (β = .062 and .061), a configuration indicating indirect-only mediation. Explanatory power was weak, with R² = .085 for self-regulated learning and .157 for critical thinking, and f² values for all four technology paths below .02. A clustering sensitivity analysis showed that the technology paths lose significance at an intraclass correlation above about .01, so these associations are provisional. Because all constructs were measured on one occasion, the ordering is a theoretical assumption rather than an observed sequence.