Nineteen Calculus II students were randomly sampled and divided into calculator (n=9) and noncalculator (n=10) groups. These students were asked to "think aloud" while solving 24 Advanced Placement calculus problems. Each student interview was videotaped, coded and analyzed for reasoning process as well as outcome. The results indicated that the use of calculators affected the testing of basic facts and reasoning processes at the individual problem level but basic concepts and principles were unaffected. A follow-up teaching experiment suggested that the content validity of Advanced Placement calculus problems could be more seriously affected if students had greater calculator expertise. (MP)