The authors examined an activity-based listening strategy with first- and third-grade children in mixed-grade dyads. On the basis of theories of cognitive development and previous research, the authors predicted the following: (a) children in an activity-based strategy would recall more story events compared with those in a repetition strategy and (b) children who performed activity would recall more events compared with those who observed activity. In addition, previous visual imagery research suggested that (c) recall in favor of the activity-based strategy would be observed when the toys were removed and imagery instructions were provided. The results confirmed the first prediction that the activity-based strategy would improve children's memory for story content. The second prediction was not supported: Physical manipulation did not improve memory beyond observing the actions performed by a peer. Last, third-grade students benefited from imagery instructions after training, whereas first-grade students did not. The authors discuss the theoretical and education implications of the results. (Contains 2 footnotes and 2 tables.)