This study aims to understand the differences in student learning outcomes and classroom behaviour when using the interpersonal computer, personal computer and pen-and-paper to solve arithmetic exercises. In this multi-session experiment, third grade students working on arithmetic exercises from various curricular units were divided into three groups. The first group used personal computers (netbooks), the second group used an interpersonal computer (ie, one projector with a screen, one computer and one mouse per child) and the third group used pen-and-paper. The results of the experiment indicate that all three groups achieved an increase in learning, as shown by the pretest and posttest scores. No significant difference was found between the interpersonal computer and personal computer groups. This suggests that the key characteristic shared by the two groups is the provision of feedback. The format that such feedback takes, either private (through a personal screen) or public (through a shared screen), is shown to make no difference. However, the results significantly favour groups that are provided with instant feedback (interpersonal computer and personal computer) as opposed to delayed feedback (pen-and-paper).