Schema training in analogical reasoning.

In 3 experiments, college students learned how to solve 20 verbal analogy problems and took transfer and memory tests. Subjects learned from worked-out examples that emphasized relational terms such as «part-to-whole» or under 3 other instructional conditions that required responding to examples or that excluded mention of relational terms. The former subjects were more accurate and faster then other subjects on solving new problems involving the same relations but less accurate in recognizing words from previous problems. This pattern is inconsistent with active responding theory, which predicts students learn best by generating answers and receiving feedback to problems, and is consistent with active learning theory, which predicts that students learn best by inducing schemas for particular problem types