Review by: Dreier, Matthias (2004-08-09)

A thorough understanding of fundamental algorithms is crucial for both computer scientists and programmers. Unfortunately, it is fairly difficult to learn such algorithms. Using animations to teach these algorithms seems promising, but several studies revealed that animations are less beneficial than hoped. In this publication the use of animations is studied in a “homework” learning scenario instead of a “final exam” scenario.

The authors argue that most studies are conducted in a “final exam” scenario, i. e. students are given a set period of time to learn the algorithms, then the learning material is taken away, and the students answer the questions of the post-test in a fixed amount of time. In contrast, the authors of this publication allowed the students to learn as long as they liked and to use the learning material while answering the questions. One group of six students used animations, another group of six students used still images. Both groups had additional material including exposition, analysis, pseudo-code and figures taken from a text-book on algorithms.

The animation group scored significantly better in the post-test. Also, the animation group learned longer, though not significantly. Both groups answered the questions in approximately the same period of time. Post-test interviews showed that the animation group was more confident of their learning process and felt more relaxed after the test.

The study of Kehoe et al. shows that using animations can benefit the learning of algorithms. However, the authors selected a small group of volunteer computer science students. The results might not be replicated on a more general student population. Nevertheless, the study is valuable for its combination of qualitative and quantitative approach. Many insights result from the observation of the learning process and the post-test interviews.