Applying computational analysis of novice learners' computer programming patterns to reveal self-regulated learning, computational thinking, and learning performance

Abstract

Educational research on predicting learners’ computer programming performance has yielded practical implications that guide task designs in computer education. There have been attempts to investigate learners’ computer programming patterns using high-frequency and automated data collection. This approach can be considered as process-based analysis as opposed to outcome-based analysis (i.e., the use of test or exam scores). In this process-based approach to investigate learners’ computer programming process, we included two critical constructs in our research, self-regulated learning and computational thinking skills. We aimed to identify learners’ computer programming patterns in the context that novice students learn a computer programming language, Python, in an online coding environment. We examined the relationships between the learners’ coding patterns, self-regulated learning, and computational thinking skills. Initially, we adopted a traditional approach with the aggregate data of learners’ computer programming behaviors. We then utilized a computational analytics approach to learner performance, self-regulated learning, and computational thinking skills, with ever-changing computer programming patterns. In our initial approach, the indicators of aggregate computer programming data were not associated with learners' learning performance and computational thinking skills. In the computational analysis approach, many indicators revealed significant differences between the identified patterns regarding computational thinking skills and self-regulated learning. Recommendations about the use of programming log data analysis methods and future scaffolding for computer programming learners are addressed.