Metacognitive strategy patterns in AI-assisted programming: Differences across programming resilience levels

Abstract

The aim of this study was to investigate how senior university students use metacognition when programming AI. 50 senior university students participated in the study. The results show that the participants were divided into two groups as low and high according to their programming resilience in AI programming activities. It was found that participants with high programming resilience performed planning behaviour from metacognitive strategies significantly more than those in the low programming resilience group. There was no significant difference between the groups in terms of monitoring and evaluation strategies, but there was a difference in metacognitive strategy display patterns. The low programming resilience group followed a linear and reactive transition pattern, while the high programming resilience group followed a cyclical and reflective pattern. This suggests that different interventions may be required for different student to learn to program with AI.