Research using the Dimensional Change Card Sort (DCCS) showed that young children are usually able to sort accurately by an initial rule but are unable to switch to a new rule when the two rules conflict. In 2 experiments, the DCCS was modified to study the effects of feedback on 3- to 5-year-old children in a problem-solving task. In Experiment 1, half of the children in each of two age groups (36 to 44 months and 52 to 60 months) were administered the DCCS task using the standard (no feedback) procedure and the other half received feedback on their post-switch responses. Children who received feedback were able to categorize according to the new (correct) rule, whereas the children in the younger age group who did not receive feedback continued to perseverate. Experiment 2 with 3-year-olds replicated the results from Experiment 1 but found that children's successful performance with feedback on the card-sorting task did not lead to improved performance on the post-switch phase of a subsequent DCCS task. Successful performance under conditions of feedback in both studies implies that 3-year-olds are capable of shifting their response mode from one rule to an alternate rule under conditions that offer clear guidance. Poor performance on the standard version is interpreted to be a reflection of the inability to monitor their own task performance in the absence of clear contextual cues. (Contains 1 figure.)