AI-enhanced flipped classroom design thinking to improve self-efficacy in open distance learning

Abstrak

Self-efficacy is a critical determinant of learner success in Open and Distance Learning (ODL), where limited interaction and tutor-centred instruction may weaken confidence and autonomy. This study examines how progressively scaffolded learning designs improve self-efficacy by comparing a Conventional Tutorial, a Flipped Classroom with Design Thinking (FCDT), and an AI-enhanced FCDT-AI model. Using a within-subjects repeated-measures design, 26 undergraduate ODL students recruited through purposive sampling experienced all three conditions in a counterbalanced sequence. Self-efficacy was measured with a validated 34-item scale, complemented by reflection logs. Quantitative analysis (Greenhouse–Geisser–adjusted repeated-measures ANOVA) showed a significant, monotonic increase in self-efficacy from the Conventional Tutorial to FCDT and FCDT-AI (p < .001). Mean scores rose from 77.79 to 80.99 and 85.05, respectively, with decreasing coefficients of variation indicating more stable confidence under higher scaffolding. Qualitative findings reinforced this progression: students described tutor dependence and uncertainty under the Conventional Tutorial, greater autonomy and structured mastery under FCDT, and the strongest, most consistent confidence under FCDT-AI due to AI-supported guidance. These results demonstrate that structured flipped design-thinking models, strengthened by generative AI scaffolding, can substantially enhance self-efficacy in ODL contexts. Future studies should investigate long-term effects across larger and more diverse learner populations.