Abstract
In science education, fostering scientific thinking is important, though few examples of scientific thinking are known. Sakamoto et al. (2007) advocated "meta-understanding of scientific principles" as one area of scientific thinking. If one is capable of meta-understanding, one is able to approach unknown phenomena with the assumption that scientific principles are applicable. Sakamoto and her colleague claim to have verified the existence of meta-understanding and the possibility of teaching it to elementary school students. However, we think they failed to provide proof, because their definition of meta-understanding is somewhat confused. Through inspection of the idea of "meta-understanding of scientific principles" and a process of inference with this meta-understanding, we refined the definition of meta-understanding introducing an important distinction they missed: whether a key causal mechanism of phenomena is known to students or not. Based on our new definition, we designed a clearer experiment which contained two pairs of tasks. We predicted a different response pattern for each pair of tasks, where the original paper expected no difference. As the result of the experiment supported our predictions, we achieved a more reliable verification of meta-understanding.
