Evidence for Programming Strategies in University Coding Exercises

Abstract

Success in coding exercises is deeply related to the strategy employed by the students to solve coding tasks. In this contribution, we analyze the programming assignments of 600 students from an introductory university course in object-oriented programming. The students were provided unit tests for the assessment of their code, and their editing and testing actions were recorded using an Eclipse plug-in. The primary motivation for this study is to discover the programming strategies used by students for coding exercises with different difficulty levels, and find out if any relation exists between these strategies and the success in solving the coding tasks. More insights into this process will enable educators to provide future students timely, appropriate and constructive feedback on their coding process. Thus, to predict success in the coding exercises, we used indicators from students’ testing behaviour reflecting the time and effort differences between two successive unit test runs. The results show a clear difference in the strategies employed by students within different success levels. The results also highlight ways of providing actionable feedback to the students in a timely and appropriate manner.