To test the hypothesis that a strategy deficit entirely accounts for retarded persons' inferior short-term memory performance, we attempted to eliminate strategies by selecting a task in which strategies would be difficult to use (memory for the position of a lifted weight) and by having subjects report the use of any strategy. In Experiment 1, discriminability among lifted weights was roughly equated for retarded and nonretarded subjects. Experiment 2 showed that for subjects not reporting use of a strategy, nonretarded subjects performed better than did retarded subjects. Use of a strategy improved the performance of nonretarded and retarded subjects by the same amount. In Experiment 3, a task was devised that made it nearly impossible to use the strategies that had been reported in Experiment 2. This increased the reported use of strategies in both groups, though strategies did not significantly improve performance. Thus, the report of strategies used in Experiments 2 and 3 was negatively related to their effectiveness. Findings of Experiment 2 and 3 suggest that a strategy deficit is not sufficient to account for the memory deficit displayed by retarded subjects.