Abstract
The present study focuses on elementary school children's undergeneralized misconceptions about animals, and examines the effectiveness for modifying these misconceptions of a teaching method using boundary instances. The main hypothesis was that elementary school children's undergeneralized concepts about animals could be effectively changed into a more scientific concept by teaching with boundary instances. 3 experiments were conducted to test this hypothesis; each experiment was performed in the context of classroom science teaching in elementary school fifth grade classes, and consisted of 3 sessions: pretesting, teaching and discussing with a videotaped aid, and post-testing. In the first and third experiments, plankton in water and shellfish respectively were adopted as boundary instances, and a videotaped aid that showed eating, moving, and excreting scenes for each respectively was used. The results of these 2 experiments did not clearly show conceptual change. In the second experiment, a revised version of the videotaped aid that treated scenes of both plankton and shellfish were used. The results showed that the ratio of correct answers increased greatly for all problems, and thus that the hypothesis was supported. The results demonstrated that teaching with paired use of 2 kinds of boundary instances is effective for modifying undergeneralized misconceptions.
