摘要：本文為兩項有關水平面概念學習的效果報告：其一是傳統的群體式實驗學習活動；其二是個別式的自我學習活動。群體式實驗學習活動包括連續的三項測試：水平面概念的前測、實作測試、水平面概念的後測。個別式自我學習活動也包括三部份：水平面概念的前測、約五天的自我學習活動、水平面概念的後測。所有的測試都係採用紙筆式示範群測法，實驗對象為小學二年級至國中二年級的學生，參與兩項實驗的學生人數各約為五百人。 學生的概念學習效果係根據他們在三個概念階層間的改變來判別：階層BC為最低的概念層次；階層D為次高的概念層次；階層E為最高的概念層次。 研究的主要發現有：(1)兩種學習實驗的結果都顯示，學生水平面概念階層的改變，都受學習開始時所表現的概念層次的制約；(2)參與群體式實驗學習活動的學生並沒有比參與個別式的自我學習活動的學生表現較好的學習效果；(3)兩種學習方式的結果都顯示原來屬於BC低層次水平面概念的學生，和原來屬於較高層次水平面概念D的學生有不同的學習特性；(4)水面觀察活動對建立階層E的概念層次的效果甚低，甚至有相當數量的學生會因為觀察活動而造成負面的學習效果；(5)學生在面對水平面問題時可以有兩種不同的思考方式，一種是想像容器原來就是靜止放置在桌面上的，另一種是想像現在容器放置方向是由其它放置方向改變而成的，對低層次的學生而言，前一種靜態的思考方式要比較容易，後一種動態的思考方式要比較難。 本研究結果對皮亞傑理論的理解、概念的診斷方法、實驗教學的意義等提供了進一步的瞭解。

　　　　　In this paper we shall report the results of two learning studies on the water-level concept. One study focused on a group experimental learning activity; the other was on a free self-learning activity. The experimental learning activity consisted of three successive group tests. In the second test each student was asked to make close observations of the actual water-level phenomena while answering related questions. The free self-learning activity consisted of two group tests. The second test was held five days later and the students participatants were not informed in advance. All tests were administered to groups in a paper and pencil format. About 500 elementary and high school students participated in each learning activity. Learning effects of the activities were judged by students' conceptual shifts among three stage levels, i.e., the lowest concept level BC that demon- strates no knowledge about the horizontality of the water surface in container; the middle concept level D that demonstrates only occasional knowledge about the horizontality of the water surface; and the final concept level E that demon- strates complete concept of horizontality in all situations. The main findings are: (1) both experimental results indicated that learning of the concept is dependent on the initial conceptual levels of the students; (2) students participating in the experimental learning activity demonstrated no better conceptual learning effect than those participating in free self-learning activity; (3) results of both experiments indicate that the students with original conceptual level BC and those students with original high conceptual level D have very different learning behaviors; (4) immediate observations provoked little conceptual learning when the jar was tilted at an arbitrary angle, i.e., very few students could see the horizontality of the water surface and thus correct their misconceptions; (5) two ways of thinking could be used when the students were facing the problems: one was a static type in which the bottle with the water inside was considered as situated on the table statically; the other was a dynamic type in which the present orientation of bottle on the table with the water inside was considered as rotated from a certain previous orientation. For the students with initial low conceptual level, using the static type was easier than the dynamic type to get the correct answers when facing the water level problems.