Three experiments examined conceptual change from reading refutational texts and how such learning interacted with prior knowledge organization. Prior to reading, 3 groups of learners were identified on the basis of their prior knowledge of the targeted concept: 2 groups held misconceptions; 1 group was generally accurate. Experiment 1 tested learning from a text that contrasted a misconception and the correct conception of the phenomenon of airflow against learning from a text that repeated the correct scientific description twice. The 2 reader groups learned from both types of texts about equally. Experiment 2 contrasted a more traditional refutational text to the "repetition" text. Learning was better with the refutational than the repetition text for both misconception groups on both measures. Experiment 3 demonstrated that learners who held largely accurate conceptions prior to reading texts that presented misconceptions preserved their largely accurate performance. Overall, the results suggest that the inclusion of an explicit refutation of the misconception is critical for instigating knowledge revision when readers possess inaccurate prior conceptions.