中華心理學刊 民 87，40 卷，2 期，117-136
Chinese Journal of Psychology 1998, Vol.40, No.2, 117-136
吳庭瑜（國立交通大學工業工程管理學系）；吳明樺（國立交通大學工業工程管理學系）；洪瑞雲（國立交通大學工業工程管理學系）
摘要
歸納推理包含假設形成和假設檢定兩個部分，過去的研究發現，在歸納推理的過程中，人往往只會形成一個假設，且在檢定假設時表現出強烈的證真偏好，有礙正確的推論。本研究的目的即在探討合作學習情境與個別學習、解釋與無解釋，及發問架構提示與無引導提示，三個因素是否有助於歸納推理的表現。161個大學生被隨機指派到上述2×2×2的實驗情境之一去解決16個類似Wason's 2-4-6的歸納推理作業的問題。在練習與學習階段中，合作學習情境的受試者是以兩人一組的方式進行解題，個別情境則是一個人單獨進行。測試階段中所有的受試者皆個別進行測試。研究結果顯示，兩人合作的學習情境在客觀的推理表現或是主觀的滿意度上都有較好的成績，且事後當單獨進行歸納推理時，也正確的發現較多的法則。另一方面，要求受試者對思考行為給與解釋的效果在學習階段時不明顯，但到了測試階段時則明顯的會使受試者產生較多的替代假設，及案例測試，使用較多證偽策略，發現了較多的法則，同時可以減少一般人對自己表現常見的過度信心的傾向。
關鍵詞：歸納推理、合作學習、建構活動、解釋、發問提示
THE EFFECTS OF COLLABORATION, GIVING EXPLANATION, AND GUIDING QUESTIONS ON INDUCTIVE REASONING
Ting-Yu Wu（Department of Industrial Engineering and Management, National Chiao Tung University）；Ming-Hwa Wu（Department of Industrial Engineering and Management, National Chiao Tung University）；Ruey-Yun Horng（Department of Industrial Engineering and Management, National Chiao Tung University）
Abstract
Inductive reasoning has two subcomponents: hypothesis formation and hypothesis testing. Studies have shown that people often fail to form more than one initial hypothesis and furthermore show a strong confirmation bias in hypotheses testing. The purpose of this study was to investigate the effects of collaboration, giving explanation, and guiding questions on inductive reasoning. One hundred and sixty-one undergraduate students were randomly assigned to one of the 2 (collaborative vs. individual learning) × 2 (explanation vs. no explanation) × 2 (guiding vs. no guiding questions) experimental conditions to work on 16 inductive reasoning problems similar to Wason's 2-4-6 task. At the practice and the learning phase, subjects in collaborative condition worked in pairs. At the testing phase, all subjects were tested individually. Results showed that collaborative experience indeed led to more correct rules discovered, both at the learning and the testing phase. Subjects who were asked to provide explanations to their reasoning were, on the other hand, found to generate more alternative hypotheses, use more instances to test their hypotheses, use higher proportion of falsifying testing instances, and discover more rules only at the testing phase. Furthermore, giving explanation also effectively lowered the overconfidence tendency exhibited by subjects in no explanation condition.
Keywords：Inductive reasoning, Collaboration, Explanation, Guiding questions
