Information and communication technologies hold much promise for use in education in developing countries. This study reports on an evaluation conducted on the introduction of computers in the delivery of the mathematics curriculum in one of the provinces of South Africa. Although the request was for an outcome evaluation very early in the implementation of the program, it was tailored in such a way as to fulfill a more formative role. Despite substantial variability in implementation, and in most cases with very weak exposure of the learners to the intervention, sufficient evidence emerged to indicate that this mode of curriculum delivery may be effective. Improvement in mathematics performance was related to a range of variables: some concerned classroom teaching practices, some referred to social differences between the learners, and some to the specific intervention. The strongest of these predictors in the sample was the strength of the intervention: the more time learners spent on using the software to study mathematics, the more improvement they showed from 1 year to the next in their performance in the subject. (Contains 10 tables.)