Exploring Common Code Reading Strategies Using Scanpath Trend Analysis With a Tolerance
Date of Award
12-2021
Document Type
Dissertation
Degree Name
Doctor of Philosophy in Computer Science
First Advisor
Dr. Ma. Mercedes T. Rodrigo
Abstract
Code reading is a process of visually perceiving the source code to understand the meaning and functions of source code elements and locate defects. While it is considered as an integral part of programming, little is known about the code reading patterns or visual strategies employed by the students while performing code comprehension to detect the bugs. In recent years, eye tracking data has been used to determine code reading patterns, but studies have been limited. Furthermore, no general patterns of code reading have been found because patterns are highly individualistic. Thus, we need to collect and compare multiple scanpaths to generate a representative path that could reveal optimal or suboptimal patterns.
The purpose of this study was to identify common code reading strategies among students engaged in a debugging task. Using Scanpath Trend Analysis (STA) with a tolerance on eye tracking data, common scanpaths of high and low performing students were generated. Results revealed differences in the code reading patterns and code reading strategies of high and low performing students. High performing students follow a bottom-up code reading strategy when debugging complex programs with logical and semantic errors. Top-down code reading strategy is employed when debugging programs with simple control structures, few lines of code and simple error types. The generated common scanpaths of the low performing students, on the other hand, showed erratic code reading patterns which means no evident code reading strategy was used. Empirical evaluation of the effectiveness of the strategies used by high performing students was also conducted and the result suggests that by teaching these strategies to students, improved debugging performance can be observed.
Recommended Citation
Christine Lourrine, Tablatin S., (2021). Exploring Common Code Reading Strategies Using Scanpath Trend Analysis With a Tolerance. Archīum.ATENEO.
https://archium.ateneo.edu/theses-dissertations/753
