This article presents a critical and systematic review of the science education research literature that explores the response of learners to contradicting information (anomalous data). The review is framed in the cognitive conflict process model (CCPM) and provides an analysis of (1) the types and frequency of possible responses, (2) the conditions by which cognitive conflict is successfully triggered, and (3) the preliminary conditions that eventually favour conceptual changes. The results conclude, among other things, that anomaly-induced cognitive conflict is rather inefficient if triggered in isolation, without supportive processing activities, or without the initial availability of conceptual alternatives. A prospective synthesis is then provided, supporting Ohlsson's view of science education activities that concentrate on cognitive utility rather than emphasising on discrediting initial conceptions. A reflection about the integration of such considerations with contemporary issues is also provided.