本研究主要目的有三：（一）探討國小六年級學生在學習「地球的運動」前所具有的另類概念類型，及與空間能力的關係；（二）探討「地球的運動」單元各主要概念是否有發展的先後順序，及「地球形狀概念與重力概念發展」模式的適當性；（三）設計整合的概念改變教學策略，探討高、低空間能力者在不同教學層次與學習方式上的概念改變立即學習效果如何。進一步探討學習效果的持續性如何。並討論影響學生概念改變的可能因素，期能提供教學者及後續研究者之參考。 為完成前述研究目的，本研究進行研究一與研究二兩個研究，所使用的研究工具包括：地球的運動開放式紙筆測驗、地球的運動封閉式紙筆測驗、空間能力測驗、深層處理學生手冊和教學活動設計、淺層處理教學活動設計等，除空間能力測驗外，皆進行預試及修正。在研究一裡，採用問卷調查法及晤談法，以地球的運動開放式紙筆測驗，選取台北市萬興和萬福國小共235名學生為對象，考驗主要概念的發展順序。並依照空間能力測驗的結果找出前後各33﹪的學生列為高低空間能力者，考驗其在科學概念上是否有所差異。本研究一的主要結論如下： 1.高空間能力者在地球形狀、重力概念、晝夜成因、四季成因的科學概念，皆優於低空間能力者。 2.「地球形狀與重力概念的發展模式」是適當的，且各主要概念間有發展的順序存在，而學生的概念是「自然－文化－發展成熟度」三者的函數。 在研究二裡，採用準實驗研究法，首先設計「地球的運動」概念改變教學策略課程，並選取台北市萬興國小及萬福國小六年級各四個班級，共128位學生為對象，進行兩週的教學實驗，隨後進行紙筆測驗立即後測，間隔四週後進行延宕後測。根據研究所獲得資料進行分析，結論如下： 1.本研究之整合概念改變教學策略，有助於學生地球的運動相關概念的改變的立即效果和延宕效果。 2.接受深層處理的概念改變立即效果和延宕效果，皆優於接受淺層處理學習的效果。 3.配對合作學習效果會因為教學層次不同而有所不同，接受深層處理來學習時，配對合作的概念改變立即效果和延宕效果會優於個別學習。但當學生接受淺層處理時，配對合作學習與個別學習的概念改變立即效果和延宕效果並無差異。 4.配對合作學習效果會因為空間能力不同而有所不同，低空間能力學生，配對合作的概念改變立即效果和延宕效果會優於個別學習。但高空間能力學生，配對合作學習與個別學習的概念改變立即效果和延宕效果並無差異。 根據研究結果，本研究提出建議，以提供教育行政單位、國小自然科教師、課程設計者以及未來相關研究之參考。 / The purposes of this dissertation were: (1) to probe into the alternative conceptions of the students about earth's motion, and the relationship between these conceptions and spatial ability; (2) to identify the developmental sequence and model of the main concepts, and verify the model is adequate; (3) to design the instructional strategy of conceptual change and explore the learning effects that high- and low-spatial ability in different processing levels and learning styles. At last, the researcher identified the retention effect, and analyzed the factors that may influence on the conceptual change. For these purposes, the whole study was divided into study Ⅰ and Ⅱ. The instruments employed in this study included: The open-ended question test and the multiple-choice test of earth's motion, the spatial ability test, the curricula and student's manual for conceptual change teaching strategy. Except the spatial ability test, the item analysis and pilot study of the others were conducted to determine the reliability and validity, and promote the quality of this study. The study Ⅰ employed questionnaire survey and interview based on the open-ended question test of earth's motion. The participants were 235 sixth grade students from two elementary schools in Taipei. Through the data, the researcher investigated the developmental sequence of children's conceptual knowledge about the earth's motion. All students were classified as being of high- or low-ability group according to their performance on the spatial ability test. High- and Low-ability students scored upper and lower 33﹪ score of all students separately. Then the researcher examined the difference between high- and low-ability group on the performance of the open-ended question test. The main conclusions of study Ⅰ were as follows: 1. A t-test on the pretest indicated significant discrepancy for ability. In all concepts that included earth's shape, gravity concept, the causes of day/night and seasons, high- ability students outscored low-ability students. 2.The developmental model of earth's shape and gravity concept was adequate. There were developmental sequences among the main concepts. Student's conceptions are the function of physico-cultural and their developmental maturity. The study Ⅱ employed quasi-experimental design. At first, the researcher designed the teaching curricula of conceptual change. Then the researcher experimented on 8 classes of 128 students from two elementary schools in Taipei. Four intact classes were randomly assigned to different experimental groups in each school. Four groups were administered the pretest two weeks before the experimental treatment. After two weeks' instruction, all subjects accepted the posttest and delayed-test of earth's motion after 4 weeks. The results were as follows: 1.The instructional strategy of conceptual change that the researcher integrated had immediate and delayed effects. 2.Students learning by deep processing performed significantly better than the learning by superficial processing on the posttest and delayed-test. 3.The learning effect of cooperative dyads depended on the levels of processing. Students learning by deep processing environment, cooperative learning group did perform better than individual learning group on the posttest and delayed-test. Conversely, the result indicated no significant difference on the posttest and delayed-test between cooperative and individual learning. 4.The learning effect of cooperative dyads depended on the levels of spatial ability. Low-ability students with cooperative learning did perform better than those with individual learning on posttest and delayed-test. The performance of high-ability students showed no significant difference on posttest and delayed-test. Based on the results, these suggestions are proposed for educational administration institution, elementary teachers for science education, curricula designer, and future researchers.