Three different methods of teaching fraction computation were compared using community college student scores on measures of fraction computation, fraction understanding, and attitude towards mathematics. One method used conventional algorithms; the second, a control for the effect of using a calculator, used conventional algorithms and calculators; the third used calculator-based algorithms. Differences between methods on the three dependent variables were not statistically different. Pretest to posttest differences were significant for all groups with fraction computation scores making the largest contribution to the increase in scores. (Author/MK)