Evaluating Online Tutorials for Data Structures and Algorithms Courses

Abstract

Abstract We present results from a study investigating the role of online tutorials for data structures and algorithms (DSA) courses in Computer Science. We used principles drawn from research and theories in disciplines such as cognitive science, motivation, and education to design the tutori-als. They were developed as part of the OpenDSA eTextbook project (http://algoviz. org/OpenDSA), an open source, online system combining textbook-quality content with algo-rithm visualizations and interactive exercises. DSA courses emphasize dynamic processes such as how various algorithms work. OpenDSA supports presenting such content in a highly visual manner through the frequent use of slideshows, simulations, and visualizations. Students were also provided a continuous stream of automated assessment questions and interactive exercises, thus providing immediate feedback to the students on their progress. A pilot study was conducted with students in a Computer Science course at Virginia Tech during Fall 2012. We tested three weeks of content on sorting and hashing in a quasi-experimental set-ting and collected quantitative and qualitative data. The data consisted of students’ performance as measured by their grades, students’ perceptions and opinions obtained on surveys, field notes from observing the classes, interview data at the end of the course, and the interaction logs that our system records. After the pilot test, students’ average grade in the treatment group was slightly (but not significantly) better than the control group on the post test. Students’ survey and interview data indicated positive feedback about OpenDSA, with the average response on how well they liked using the OpenDSA materials increasing after use as compared to a similar pre-test question about whether they would like to use such materials.