COPPERS is a mathematical learning environment which aims to support thedevelopment of children's multiplication skills. Two evaluation studies were conducted which examined the extent to which experience with COPPERS would help children to learn that problems may have many different solutions. In particular, the aim of the studies was to examine the importance of various components of the system with respect to learning outcomes. To this end, a number of versions of COPPERS were created varying the presence of these features. The first experiment examined the importance of the two different types of feedback provided by the system and of requiring multiple rather than single answers to a question. Forty subjects aged between six and seven were given pen and paper tests of the sort of questions asked by the computer and then given experience with one version of the program. Further paper tests were given immediately after the intervention and four weeks later. It was found that children in all conditions demonstrated significant improvements. Students who had received both types of feedback provided by the system performed significantly better than those who had seen only one of them. Secondly, for students with initially poorer performance, giving multiple answers per question was important. Finally, it was found that the strategy which children used to calculate their answers to the problems on the computer, was related to their performance at post-test. A second experiment with 50 eight to nine year old children compared three further versions of COPPERS. The same procedure as before was followed and contrasted the effects of requiring four or eight answers per question. It also examined the effects of giving learners the choice of how many answers to give per question and examined whether their choice was related to subsequent performance. A no treatment control group was also included. Children in all experimental groups demonstrated significant improvement in performance, although it was hypothesised that this improvement was based upon the acquisition of different skills. The control group showed no improvement in performance. It was found that students in the experimental groups performed better than those in the control group and that children in the eight answer group performed best. The strategy chosen by children in the autonomous group was found to be mainly unrelated to post-test performance.