Strategies for Using Online Practice Problems

Document Type


Publication Date

June 2014

Publication Title

2014 ASEE Annual Conference & Exposition


PathFinder is an active website coded in html, asp.net, c#, JavaScript, xml, and MathML. The website assembles ebooks on the fly from an xml database. The ebooks have randomly selected and generated exercises that are automatically graded. Instantaneous feedback is provided to both students and professors regarding performance on on-line exercises. PathFinder is used to deliver an ebook to a first year introductory engineering course. In Fall 2013 12 sections of 20 – 25 students each used the Pathfinder ebook. This provides an opportunity to investigate the effectiveness of on-line practice problems. PathFinder provides on-line practice problems that students can work before completing scored problems. Practice problems can be similar or related to the student’s scored problem. A similar practice problem is identical to the scored problem (same problem statement), but the given input values are different. A related practice problem is different from the scored problem, both in input values and problem statement. Four sets of four problems each were used to investigate 4 scenarios. Scenarios were randomly applied such that each student was exposed to all four scenarios, but on different sets of problems. In all scenarios the fourth problem had no practice problem, so it could be used as a test of the effectiveness of the practice problems provided for the first three problems. In the first scenario, no practice problems were provided. In the second, three similar practice problems were provided. In the third, three related practice problems were provided. In the fourth, the first problem had a similar, the second a related, and the third no practice problem.The effectiveness of each scenario will be evaluated using (1) student scores on the fourth problems, (2) survey questions completed by students after finishing each series of four questions, and (3) examination performance on related problems. The results are expected to provide feedback that can be used to select one scenario for providing practice problems in the future.