本研究之目的，係探討利用短時間群集評量方式（Time series design），評估初任理化教師教學成效之可行性及有效性。四位初任理化教師及其任教之四班國中二年級學生參與本研究。評量方式乃採用短時間群集設計，追蹤這些學生在理化第八章化學反應的學習成就。研究結果顯示，在職前訓練中專業學科成績及教學表現優異，且具現代科學哲學觀的初任理化教師教學之下，學生的學科成就進步型態，如預期般顯著的優於那些教學表現較差，且具傳統科學哲學觀的初任理化教師所教之學生，顯示出該評量方式在評估初任理化教師之教學成效的可行性。另外有關該評量方式的效度之考驗，由其對高成就與低成就學生的鑑別能力，以及研究工具本身的效標關連效度，進一步顯示其有效性。未來若經進一步其它評量方式比較結果，確具實用性，則在師資檢定授證系統，似乎可將該評量方式當成另外一種評估初任理化教師教學成效的方法，以便提供較為客觀的量化數據給相關人員參考。

　　　　　The purpose of this study was to explore the validity and feasibility of using time series design in the assessment of beginning science teachers' teaching effectiveness. Four beginning physical science teachers and four classes of 8th grade students who were taught by the four teachers participated in this study. The students' learning achievement on the chapter of chemical reaction was assessed by time series design. The results indicate as predicted that students taught by teachers with high GPAs, higher rated teaching performance, and contemporary views about the nature of science made more significant progress on learning achievement than their counterpart who were taught by teachers with low GPAs, lower rated teaching performance, and more traditional view about the nature of science. The convincing criterion-related validity of the instrument and the procedure's high capability of discriminating high and low achieving students suggest that the time series design is effective and feasible in assessing beginning science teachers' teaching effectiveness. Our results suggest that time series design can play a role in the teacher certification system by serving as an alternative of providing quantitative evidence of practicing teachers' teaching performance. More effort are encouraged to verify the validity and feasibility of time series design with other assessment methods.