Problem-Based Learning in Teaching Theoretical Computer Science
Roman Bednarik · 2004
ABSTRACT: Problem-based learning has proved to be an effective way to deliver university level courses in various disciplines. However, the foundational theoretical courses of computer science, such as the theory of formal languages or the theory of computability, are still mostly taught in the traditional behaviouristic settings. Beside other negative effects, traditional settings often leads to low student involvement and, as a consequence, to a high rate of withdrawal. In addition, in the traditional lecture-exercise-exam format, teachers ’ assessment of the progress and performance, and student comprehension of subject matter are quite difficult tasks. Very few computer science faculties have reported the use of innovative teaching strategies. Especially, the problem-based learning method is still underutilized. In this paper, we present our experiments and results in teaching a course of theoretical computer science taught in a problem-based learning (PBL) format and supplemented with learning diaries. The experiment was done on the multinational body of students, where the previous learning habits often greatly differed among the learners from different institutions and countries. We report on the methodological challenges of PBL, especially on practical implementation, and on the ways of scheduling, delivering, and assessing the learning process of highly diverse group. We argue that the combination of PBL and learning diaries yields a comprehensive and fair evaluation of learning. Our results show no interaction of PBL with gender or nationality. Our results and experience also indicate that even though the preparation and assessment phases were longer for PBL, the use of PBL in theoretical computer science teaching brings positive effects on important instructional goals, such as students ' comprehension and involvement, and on the course dynamics and atmosphere. 1