Previous research reports common student errors in introductory programming (CS1) classes.
Knowing common errors enables us to improve teaching and content to train students to avoid
those errors, and to provide an automated help system like providing hints based on particular
auto-detected errors. Finding and fixing some errors is a part of learning, so our focus is
specifically on errors that cause struggle, meaning excessive time or attempts. Struggle may lead
to giving up, loss of confidence, or cheating. For 89 online auto-graded coding homework
problems in our CS1 class of 100 students (mostly engineering/science majors), we first
automatically determined the 12 problems with the highest struggle rates. Then, we spent about
100 hours manually examining incorrect student submissions to determine what errors caused
struggle and the time spent on each error. Like previous work, we found many common general
errors, like using = rather than ==. However, we also found problem-specific errors, like
misusing a particular library function, leading to a first conclusion that a help system should
allow teachers/authors to add problem-specific hints. Furthermore, we analyzed errors that
caused the longest struggle, and found some uncommon "one-off" errors, leading to a second
conclusion that a help system won't be able to detect all errors and thus might need automated
recommending or alerting for human assistance (or other techniques).
Nabeel Alzahrani is a Computer Science Ph.D. student in the Department of Computer Science and Engineering at the University of California, Riverside. Nabeel's research interests include causes of student struggle, and debugging methodologies, in introductory computer programming courses.
Frank Vahid is a Professor of Computer Science and Engineering at the Univ. of California, Riverside. His research interests include embedded systems design, and engineering education. He is a co-founder of zyBooks.com.
Alex Edgcomb is Sr. Software Engineer at zyBooks.com, a startup spun-off from UC Riverside that develops interactive, web-native learning materials for STEM courses. Alex is also a research specialist at UC Riverside, studying the efficacy of web-native content and digital education.
Online preview: this file is waiting to be converted into a web-viewable format...
Download paper (804 KB)Are you a researcher? Would you like to cite this paper? Visit the ASEE document repository at peer.asee.org for more tools and easy citations.