Optimizing Procedural Knowledge Transfer in AI Chatbot-Enhanced Flipped Learning Models on Computer Programming Course

Abstrak

This research investigates whether an AI chatbot can further enhance the effectiveness of the flipped classroom model by facilitating personalized learning in computer programming. A quasi-experimental pretest-posttest with a matched samples control group design was used with a population of 60 in a vocational high school in Indonesia. The experimental treatment group of 30 students received flipped instruction supported by AI chatbot; the 30-student strong control group had traditional flipped instruction. Cognitive test data, psychomotor tests, and self-reported learning questionnaires measured data. Results show that the AI-enhanced flipped classroom outperformed the traditional flipped classroom to a great degree in both personalized learning (adjusted mean difference = 12.30, p < 0.001, partial η² = .335) and procedural knowledge acquisition (mean difference = 9.70, p < 0.001). The effect sizes were large for individualized instruction (Cohen's d = 0.89, 95% CI: 0.75 to 1.03) and medium for procedural knowledge (Cohen's d = 0.62, 95% CI: 0.49 to 0.75). Of particular note was that the experimental group not only did better on coding tasks (d = 1.18), but even more so on debugging efficiency (d = 1.48).

Identifier

• OpenAlex: https://openalex.org/W7168410649
• DOI: https://doi.org/10.21831/elinvo.v11i1.96278