本研究應用問卷調查與教室觀察研究法，探討中學物理教師對科學的本質及物理教學之信念，以侑教師信念如何影響其教學策略與其所營造的教室環境，以提供實際、客觀的教學經驗與啟示。本研究包括觀察、晤談、文件收集、調查訪問、量具的研發與測試等工作，並應用歸納式數據分析法，以求取詳實、客觀的詮釋，所欲瞭解的問題有三方面：一、教師信念，二、教學的實務、三、教室環境、並力求尋找此三方面相互間的動態機制。本研究以建構主義理念為參考座標來衡量教師對科學本質及科學教學的信念，也用同一個座標來評量建構式教學環境。經文獻分析，有關科學本質的構成，本研究選定「科學探討的動態過程」與「科學知識的不確定性」兩項本質，作為探討建構式教室環境的架構，並據以發展「教師科學本質及科學教學信念」及「建構式教室環境」兩種量表。本研究在臺北、高雄、臺中三地區的國中為對象，採二階段叢集隨機取樣，計抽取50個班級三學生1879人侑該班理化任課教師50人，皆別施以「建構式教室環境量表」及「教師科學本質及科學教學信念」，調查資料顯示科學教師的科學本質信念是介於建構主義學派與客觀主義學派的中間地帶，並未完全符合建構式理念；但學生對科學教室環境的感受，卻偏向客觀主義式的教學，顯然教師科學本質信念與其營造的建構式教室環境間的一致性，有著一段落差。經教室的個案觀察研究，教學內容大多為背誦與演算過程反覆練習，以教師為中心的教學傳統，已累積相當大的慣性，學生缺乏建構的機會，教師之所以傾向客觀主義取向的教學，除了深受教師的信念影響之外，也受到來自外界強烈的束縛，其中包括升學聯考的壓力、課程標準僵硬的規定等因素。

　　　　　This paper reports a combined quantitative and qualitative study of the science classroom environment. It focuses on teachers' beliefs about science and science teaching, and the effects these have on the learning environment in a science classroom. By understanding more about this relationship, it is hoped that barriers to the introduction of constructivist teaching approaches in school science can be identified and strategies for enabling teachers, curriculum developers, and policy-makers to overcome them can be proposed. Survey instruments for assessing teacher beliefs and classroom environment have been developed and validated. The "Beliefs About Science and School Science Questionnaire" (BASSSQ) was designed to assess teachers' beliefs about the nature of science and school science. The "Constructivist Learning Environment Survey" (CLES), which was designed to assess students' perceptions of the constructivist nature of their classroom learning environment, was adapted for use in this study. The CLES was administered to students in 50 junior high school science classes. The teachers of each of the classes responded to the BASSSQ. In this paper, we present the following statistical analyses of questionnaire data: (1)internal consistency, reliability, and discriminate validity; and (2)mean scores and standard deviations. The findings one the viability show that the BASSSQ is conceptually sound. The Cronbach alpha reliability of the CLES ranged from 0.73 to 0.93. BASSSQ data analysis suggests the teachers had views on the nature of science that occupied the middle ground of the postmodern-objectivist continuum. On the other hand, the CLES data showed that the learning environment was not often constructivist in nature. There was infrequent opportunity for students to engage in reflective discussions with other students during class. The interpretive study combined classroom observations and interviews with teachers and students. Qualitative analysis revealed the examination-driven nature of teaching. The classes were mainly teacher-centered. The teachers seemed mostly concerned with the content coverage. The student were left with little chance to experience science in their classrooms.