Children's naive theories include misconceptions which can interfere with science learning. This research examined the effect of pairing anomalies with alternative theories, and their order of presentation, on children's belief revision. Children believe that heavy objects sink and light ones float. In a pre-, mid-, and post-test design, 5-year-olds (N = 96) were assigned to one of two conditions, where they were either exposed to an alternative theory about buoyancy and then observed anomalies (Explanation-First), or the reverse (Anomalies-First). At mid-test, children were more likely to revise their beliefs after exposure to an alternative theory than anomalies alone. At post-test, children revised their naïve belief when they had access to an alternative theory before the anomalous evidence than in the opposite order.