抄録
This research examined relationships between modeling effects and observer's comprehension levels for task structure before observation. In Experiment I, 5-year-olds were divided into three groups with different comprehension levels (Level I being the lowest) derived from the pretest score, and trained under one of four conditions (modeling with instruction, modeling with correct or incorrect cue, modeling only, control). After training, subjects received imitation and generalization test. The results indicated that the Level III showed the best performance in both tests, regardless of the training conditions, however, the Level I was superior to the Level II in both tests. A detailed examination of strategies used in the pretest revealed that the Level I showed less frequent change of strategy than that of the Level II. In Experiment II, subjects who showed no change of strategy in the pretest (consistent group) did better performance than subjects who showed strategy change (inconsistent group) after observation. These results were discussed in terms of the strategies employed in the pretest and the observer's dependence on the model.
