國家教育研究院

　　國內鮮少以新移民家庭子女為母群體進行全面性、長期性的觀察，了解其學習表現成長趨勢、相關重要影響因素、及教育政策對其學習表現之影響。本研究延續2012年國家教育研究院針對新移民家庭子女的大規模調查結果，從既有臺灣北區新移民家庭子女樣本中，再抽取部分代表性樣本，繼續追蹤新移民家庭子女數學學習表現的進展及其相關影響因素。本研究將四年級的數學學習表現視為起始狀態，假設它會影響學習表現的進展，試圖透過分層叢集抽樣方式，抽出最接近母群分配的樣本。扣除轉學生之後，2014年仍於原校就讀的學生共845人（95.37%），同意接受施測的學生共630人（71.11%），實際到考人數共626人（70.65%），平均年齡為12.36歲。研究工具為數學成就測驗、學生問卷、新移民家庭子女參與學習方案調查表及學校連絡人觀點調查表。數學成就測驗試題自臺灣學生學習成就評量資料庫試題之中挑選，題型為選擇題，總題數78題，以平衡不完全區塊設計題本模式安排題項分配至13個題本，每個題本24題，亦即每位學生於40分鐘之內作答24題，正式施測後各題本Cronbach’s alpha介於.82至.90之間。學生問卷以期望價值理論為基礎建構題項，包括24題背景題項和46題心理構念題項，每構念Cronbach’s alpha介於.74至.92之間。以Jackknife方法估計抽樣標準誤，並以many-faceted Rasch model估計學習表現成長量、學習表現成長量和相關重要影響因素之交互作用。使用的軟體為Acer ConQuest 3。研究發現，新移民家庭子女一年的數學學習成長量為0.479 logit和國際研究結果相近，而經過兩年，「大陸地區與港澳」和「東南亞」和兩組學生於「數與量」、「幾何」、「統計與機率」以及「代數」四項數學主題能力均有進步，其中「幾何」為新移民家庭子女通過率和進步量最不理想的主題。學習方案對於新移民家庭子女數學學習成長的增進，成效不如預期，唯班級常規適應良好的新移民家庭子女四年級至六年級兩年間數學學習表現進步量顯著地高於班級常規適應不佳者。本最後針對補救教學規劃、課綱規劃、新移民家庭三方面提出具體建議。

　　This research is an extension of a large-scale survey on new-immigrant children conducted by the National Academy for Educational Research in 2012. This research chooses a representative subsample from the existing sample of new-immigrant children from northern Taiwan for continued tracking of their mathematics progress and influencing factors. After removing transfer students, 630 students (71.11% of the total) were willing to take the test and 626 (70.65%) were actually tested, with an average age of 12.36. Research tools employed include a mathematics achievement test, a student questionnaire, a learning program questionnaire for new-immigrant children, and a perspective questionnaire for school contacts. The questions used for the mathematics achievement test are multiple choice questions selected from the Taiwan Assessment of Student Achievement item bank. A total of 78 questions were arranged into 13 booklets using the method of balanced incomplete block design. Each booklet contained 24 questions, meaning each student needed to answer 24 questions in 40 minutes. Cronbach’s alpha was shown to be between .82 and .90 for each booklet. The student questionnaire included 24 background information items and 26 psychological construct items. The standard error was estimated using the Jackknife method, and the growth rate as well as the interaction between growth rate and relevant key influence factors were estimated using a many-faceted Rasch model. The software used was Acer ConQuest 3. In general, new-immigrant children’s progress in mathematics performance is 0.479 logit per year, which is not far from international results. Comparison of the 2012 and 2014 test results on four mathematical topics, namely, “numbers and measurement,” “geometry,” “statistics and probability,” and “algebra,” indicate that “geometry” has the worst pass rate for new-immigrant children. Besides, remedial instruction is unable to provide new-immigrant children with lower performance with a chance to catch up, although it does provide them with progress matching their peers. The new-immigrant children who are well adapted to classroom rules have a significantly higher growth rate for mathematics performance than those who do not adapt well to classroom rules. Based on the research findings, this research proposes detailed suggestions with regard to the intended curriculum, remedial instruction project, and practical helping for new-immigrant family.