TY - JOUR
T1 - Coding Trails
T2 - 2021 ASEE Virtual Annual Conference, ASEE 2021
AU - Vahid, Frank
AU - Lysecky, Roman
AU - Miller, Bailey Alan
AU - Vanderbeek, Lyssa
N1 - Publisher Copyright:
© American Society for Engineering Education, 2021
PY - 2021/7/26
Y1 - 2021/7/26
N2 - CS instructors desire visibility into student programming behavior, such as seeing the days a student worked, the time spent, and the number of compiles/runs. Such visibility may help find struggling students, prevent or detect cheating, and provide insight into the effect of new policies like points for earlier starts. Such visibility historically has been severely limited due to student use of external tools. Today, many education-focused program auto-graders provide a cloud-based development environment that records much student behavior. Detailed logs are cumbersome to view, especially for large classes, but conversely, summary statistics like averages and standard deviations lose much useful information. This paper introduces the concept of a "coding trail" as an attempt to visually and concisely summarize a student's coding behavior on a programming assignment. Our visual coding trail displays dates, each develop run, each submit run for auto-grading and score, and dramatic changes in code (often a sign of cheating). The coding trail is textual rather than graphical, allowing easy copy-paste, incorporation into spreadsheet gradebooks, and parsing by tools for further analysis, at the expense of some information loss. A version of our coding trail has been implemented in the zyBooks program auto-grader and appeared for over 2,000 courses and 130,000 students in 2020, with numbers growing. This paper introduces the coding trail, discusses various tradeoffs in its design, and points to a variety of uses.
AB - CS instructors desire visibility into student programming behavior, such as seeing the days a student worked, the time spent, and the number of compiles/runs. Such visibility may help find struggling students, prevent or detect cheating, and provide insight into the effect of new policies like points for earlier starts. Such visibility historically has been severely limited due to student use of external tools. Today, many education-focused program auto-graders provide a cloud-based development environment that records much student behavior. Detailed logs are cumbersome to view, especially for large classes, but conversely, summary statistics like averages and standard deviations lose much useful information. This paper introduces the concept of a "coding trail" as an attempt to visually and concisely summarize a student's coding behavior on a programming assignment. Our visual coding trail displays dates, each develop run, each submit run for auto-grading and score, and dramatic changes in code (often a sign of cheating). The coding trail is textual rather than graphical, allowing easy copy-paste, incorporation into spreadsheet gradebooks, and parsing by tools for further analysis, at the expense of some information loss. A version of our coding trail has been implemented in the zyBooks program auto-grader and appeared for over 2,000 courses and 130,000 students in 2020, with numbers growing. This paper introduces the coding trail, discusses various tradeoffs in its design, and points to a variety of uses.
UR - http://www.scopus.com/inward/record.url?scp=85124512822&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85124512822&partnerID=8YFLogxK
M3 - Conference article
AN - SCOPUS:85124512822
SN - 2153-5965
JO - ASEE Annual Conference and Exposition, Conference Proceedings
JF - ASEE Annual Conference and Exposition, Conference Proceedings
Y2 - 26 July 2021 through 29 July 2021
ER -