本研究旨在探究原住民地區幼兒園教師實施STEM方案模組時使用的教學策略，及教學策略如何影響幼兒發展科學實作。本研究採用多重個案研究法，研究參與者為北部地區以泰雅族為主的兩個班級，共有3位教師與25位幼兒（平均年齡為5歲6個月）。蒐集課室觀察資料，並在課程介入前後進行幼兒的科學實作測驗。使用質性資料技巧分析課堂觀察資料，並用Wilcoxon 符號等級檢定幼兒科學實作在課程後是否有進步。本研究有以下五點發現，第一，教師共使用七項教學策略，分別為示範（教師示範與學生示範）、比較與對比、多元表徵（圖像、圖表與角色扮演）、探詢（探詢以歸納與探詢以闡述）、回顧（連結生活經驗與複習）、直接引導（直述與協調糾紛）、及情感支持；第二，分析課堂觀察資料發現，幼兒逐步發展13項不同程度的科學實作，這些實作分別為初步觀察、初步嘗試、提出預測、聚焦、規劃、設計、系統性觀察、操作、測量、紀錄、實驗、分享、及討論；第三，前後測驗分析也發現，幼兒在課程後的科學實作表現有顯著進步；第四，對幼兒來說越複雜的科學實作（例如紀錄），似乎需要教師更多樣的教學策略引導；第五，分析焦點幼兒和教師課堂互動發現，教師針對不同起始能力的幼兒，給予不同的教學策略，以增進其科學實作表現。期待本研究提供幼兒園教師在實施STEM課程時，運用教學策略及觀察幼兒科學實作表現之參
考。

This study aimed to identify the teaching strategies used in a project-based STEM module implemented in an indigenous area and to examine the effects on kindergarteners’ scientific practices. Using a multiple case study method, three teachers and twenty five kindergarteners (average age: six and half year) from two classrooms in an Atayal community in northern Taiwan were recruited. Classroom observation was conducted and a performance-based assessment regarding scientific practices was applied before and after the module. Qualitative data analysis techniques were applied to analyze classroom observation and a Wilcoxon signed-rank test was applied to examine whether the children improved their scientific practices. This study found that: (1) Seven teaching strategies were identified to promote children’s scientific practices: demonstrating, comparing and contrasting, using multiple representations, probing, reviewing, giving direct guidance and providing emotional support. (2) Through analyzing classroom observation, this study found that children improved in thirteen scientific practices. These practices were making initial observations, doing initial testing, making predictions, focusing on key ideas, making a plan, designing procedures, making systematic observations, seeking information, employing equipment and tools, measuring, recording, conducting experiments, sharing and discussing. (3) By examining the pre and post scores of the assessment, it revealed that the children had significantly better performances on scientific practices after taking the module. (4) It seems that the teachers needed to use more teaching strategies to help the children develop more complex scientific practices (e.g., recording). (5) The analysis of the interaction between targeted children and the teachers revealed that teachers employed different teaching strategies to guide children with different initial levels of scientific practices. This study would hopefully provide insight into ways to improve teachers’ teaching strategies and children’s scientific practices when early STEM curricula are implemented.