A Modular, Practical Test for a Programming Course.

In order to evaluate students' programming skills during a university course, a practical programming test can be administered, in which students are required to implement a short yet complete program according to a provided specification. However, such tests often suffer from drawbacks that prevent comprehensive and accurate assessment of students' abilities. In this paper we identify these drawbacks and then present a modular, practical test that avoids common testing pitfalls, as well as show how to design such a test based on course learning outcomes. A key aspect of our approach is adoption of modularity, which ensures independent and comprehensive verification of learning outcomes. We have used our method to evaluate object-oriented programming skills of undergraduate students over several years and have found that our testing approach has proven its validity and superiority over approaches employed previously.