就月相盈虧之成因而言，「地球影子遮住月球」為中學生普遍的錯誤 解釋，因此，本研究將探討這一類學生學習「月相盈虧及日、月食」單元的學 習過程及概念改變情形，並試圖自學生的學習過程中，找出概念改愛的關鍵 點。本研究之對象為24位以「地球影子遮住月球」解釋盈虧現象的國三學生， 其中，高、低空間能力的男生、女生各有6人。研究者依分層隨機抽樣的原則， 將此24位學生分成三組:「控制組」、「自我解釋組」及「模型組」，以進行不 同模式之閱讀學習。本研究之實施程序主要分為前測、實驗處理及後測三個階 段。前測及後測皆採個別晤談的方式，旨在確定學生學習前、後的心智模式。 實驗處理則為學生之閱讀學習，分為三組。其中，「控制組」學生將教材閱讀 兩次;「自我解釋組」學生在閱讀同時，並對教材內容加以解釋;「模型組」學 生閱讀時，除解釋教材之外，本研究並提供月相盈虧模型以供其參考。後測結 果顯示，空間能力的高、低影響學生之學習成效。本研究三種不同學習模式中， 高空間能力學生的學習成效皆優於低空間能力學生。此外，高空間能力的學生 當中，「模型組」與「自我解釋組」學生之學習成效無明額差異，「控制組」 學生之學習成效略低。低空間能力的學生當中，「模型組」學生之學習成效最 高，「控制組」次之，「自我解釋組」最低。 自學生之學習歷程分析中，亦獲得類似的結果。能正確解釋盈虧現象的學生，皆 能將教材中平面之日、月、地相對位置關係圖正確地轉換為三度空間之心像;不 具此能力的學生(大都為低空間能力者)，則仍保有完整之「地球影子遮住月球」 造成盈虧現象的想法。此外，低空間能力之「模型組」學生在閱讀過程中，嘗試 藉由月相盈虧模型的輔助以了解教材內容。由此顯示，提供具體模型可能有助於 學生學習此單元。此外，鼓勵高空間能力的學生作自我解釋，可能有助於其形成 完整的知識架構;對低空間能力的學生而言，則可能加深其錯誤之想法。

　　　　　Sevreal earlier studies have found the most common misconceptions of the moonphases is the belief that they are caused by the shadow of the earth. In this study, weinvestigate middle school students' procession of the knowledge of moon phases, anduncover the critical point of the conceptual change of learning about the moon phases. In this study, 24 ninth-grade students are stratfied by gender and spatial abilitiesfirst, and then randomly assigned to three groups equally. One is called selfexplanations group in which students are required to provide explanations of the textwhile reading. The second group is called model group in which the students are provided with a moon-phase model. The students have to explain each phenomenon in accordance with the model. The third model is a control group in which the students arenot asked to do anything specifically. The design of the study has three phases: pretest,treatment, and posttest. The results reveal that the students' spatial ability is a main factor. In general, highspatial ability students outperform low spatial ability students in every group. For highspatial ability students, there is no difference between the students in the model groupand self-explanation group. And the high spatial ability students in both groups are better than the control group. However, the low spatial ability students in the model groupoutperformed those in the control group, the self-explanation group was the worst. Theresults also suggest that it is difficult to learn the moon phases from two dimensionalrepresentations, particularly, for the low spatial ability students. The relevance of thisresearch for instruction, curriculum design, and models of learning is then discussed.