There is little doubt that intervention research based on metacognitive principles has been remarkably successful in improving children's performance on a range of academic tasks. However, to build on this success, at least three modifications need to be made in the way metacognition is usually thought about. First, more attention needs to be given to the developmental issues pertinent to intervention research, and especially to understanding the factors affecting the emergence of the conscious self-regulation of thought. Second, fundamental insights into the nature and development of metacognition require an understanding of the transition from other-regulated to self-regulated thought. Third, researchers should focus on the interaction between metacognitive processes and other self-evaluation issues. In addition, metacognitive training needs to emphasize the importance of social interactions, to treat children as coinvestigators in intervention, and to consider that children of different ages and abilities will be at different phases in the development of metacognitive skills. (FL)