A Learning Support Framework for Promoting the Construction of Higher-Quality Code Based on Program Description and Behavior

This study extends a learning support framework proposed in prior research to encourage writing “higherqualitycode”, introducing a new quality indicator for evaluating internal quality. Specifically, within a visual programmingenvironment using a crop harvesting game, we adopted the number of blocks used as an indicator ofinternal quality and designed a quality indicator combining it with external quality (number of harvests): cost performance(number of harvests divided by number of blocks used). We conducted an evaluation experiment with16 participants divided into an experimental group (presented with cost performance-based rankings) and a controlgroup (presented with harvest count-based rankings). Although no statistically significant interaction was found inthe mixed-design ANOVA, from the perspective of effect sizes, the control group tended to improve harvest countwhile the experimental group tended to improve cost performance. Particularly in the water problem, the experimentalgroup improved cost performance while maintaining harvest count, suggesting the possibility that learning alignedwith balancing internal and external quality was conducted. These results suggest that introducing internal qualityas a ranking indicator may lead to different directions in learners’ trial-and-error efforts, though verification throughlarger-scale experiments is necessary due to the limited sample size.