During the 1994–1995 academic year, I taught one section of Calculus I and II using a graphing calculator while teaching the remaining sections in the traditional way. Students were randomly assigned to the experimental sections. Evaluations of the experimental sections were conducted to determine whether the use of the graphing calculator should be considered for adoption beyond the experimental sections. The evaluations included observations of in-class student behavior, an analysis of answers to a short questionnaire administered to students during the last class before the final examination, and traditional class grades as a measure of performance. As table 1 indicates, the retention and performance for the experimental sections exceeded those of the other, traditionally taught sections.