預測臺灣幼童早期英文閱讀技能之研究

Abstract

本研究探討如何發展一套復合的工具，使之準確地預測台灣幼童在早期英文閱讀技能 (English prereading skills) 的發展。六十八位台灣幼童經過社會經濟地位 (SES)，華語聲韻覺識 (Chinese phonological awareness) 和注音符號 (Chinese graphophonemic knowledge) 的評估後，接受十個星期的英語教學，為期十週的英語教學結束後，研究者評估幼童的英文聲韻覺識 (English phonological awareness) 和英文符號知識 (English graphophonemic knowledge)，再據以統計分析個人的社會經濟地位和早期華語閱讀技能，探討是否能預測英文早期閱讀技術的發展。本研究利用多元回歸分析(Multiple regression analysis) 藉以測試本研究相關工具的預測能力。研究結果發現：1) 社會經濟地位可以預測台灣的幼童在早期英文閱讀技能的發展，2) 早期華語閱讀技能可以有效的預測早期英文閱讀技能的發展，3) 多元回歸分析能用來衡量社會經濟地位和早期華語閱讀技能，使之在預測早期英文閱讀技能更為準確。

The current study researched the possibility of developing a composite instrument that will serve to more accurately predict the development of English prereading skills of Taiwanese kindergarteners than instruments based on a single factor or closely related set of factors. 68 Taiwanese kindergarteners were assessed using the social-background factor of SES as well as the two child-based factors of Chinese phonological awareness and Chinese graphophonemic knowledge. The subjects were then exposed to ten weeks of English instruction. Upon the conclusion of the ten-week course, the children’s English phonological awareness and English graphophonemic knowledge were assessed and statistical analysis was performed in order to determine if SES and Chinese prereading skills serve to predict the development of English prereading skills. Multiple regression analysis was performed in order to determine if a composite instrument consisting of scores from all the above factors and their subcomponents serve to provide greater prediction power than any single factor alone. Results are as follows: (1) A consistent, moderate predictive relationship exists between SES and English prereading skills in Taiwanese kindergarteners, (2) A consistent, strong predictive relationship exists between Chinese prereading skills and English prereading skills, and (3) multiple regression analysis allows for constructing a model based on measures of SES and Chinese prereading that achieves a much higher level of predictive power for English prereading skills than using any one measurement or set of measurements. Important implications from these results are discussed.