Students in three grade 7 classes, a grade 6 class and a split grade 5/6 class were provided with calculators for use in their mathematics classes for a period of eight weeks. For the posttest on problem solving the 116 students were randomly assigned to the group using calculators on the posttest or to the group using only paper and pencil. In a second section of the posttest the groups were reversed so that all students took a test in both modes, that is using calculators in one section and using paper and pencil only in the other. Pretest differences were non-significant. Out of eight comparisons on measures of correct answers to problems, three comparisons were statistically significant in favor of the calculator mode of problem solving. The other five differences were non-significant. On measures of problems attempted and methods correct all differences were non-significant except three, and each of these were in favor of the calculator mode. These results are consistent with previous studies. The calculator has no negative effects over short term periods and sometimes has positive effects. (Author)