Novice Programmers' Saccadic Patterns in Error Message Comprehension and Syntax Error Identification
DOI:
https://doi.org/10.58459/icce.2024.4975Abstract
Learning debugging is crucial for novice programmers, and as compiler error messages play a vital role in this process, exploring how novice programmers read and interpret these messages has become a focal point for researchers in computer science education. Understanding how novice programmers interpret compiler error messages is essential, as debugging skills are fundamental for their development. In line with this, our prior works investigated how student programmers process syntax errors embedded in Java and C++ programs and the relationship between novice programmers' personalities and their visual attention patterns using fixation-based metrics. Our previous studies were limited in that they did not capture the dynamics of the participants' eye movements. As our next step, in this study, we present saccade- based metrics that we used to gain insights into the problem-solving strategies employed by novice programmers. We found that high and low performing participants demonstrate similar strategies for scanning through code and detecting errors. Furthermore, high performing participants demonstrate longer average saccade lengths, indicative of purposeful exploration and the gathering of more information in each eye movement. Additionally, our findings indicate that high performing participants exhibit a lower average regression ratio, suggesting a greater ability to maintain focus and comprehension throughout problem-solving tasks. In future work, integrating contextual factors such as task complexity, familiarity with programming languages and individual learning styles into the analysis of saccade-based metrics can help identify situational factors shaping problem-solving strategies.