相似字與非相似字呈現方式對國小一年級學生生字學習效果之比較

A Comparison of the Ability of First Graders in Taiwan to Learn Graphemi Cally-Similar Versus Graphemically-Dissimilar Chinese Characters

本研究旨在比較呈現相似字群（例如：官、宮、客、容）與非相似字群（例如：京、核、官、間）對國小一年級學生生字學習之效果。本研究為2（相似字組、非相似字組）×3（高、中、低成就組）前、後測準實驗研究設計。本研究樣本有共有148位國小一年級學生，分屬4個班級；隨機分配其中兩班為相似組，另外兩班為非相似字組；每一班級學生再依照其學期總成績分為高、中、低成就組。實驗材料共為16個國字，兩屬兩類（相似字、非相似字）4組，每組4個字。資料統計分析計採用共變數分析、變異數分析及t－檢定。本研究發現（1）相似字的呈現有利於國小一年級低成就學生之生字學習；（2）高、中成就學生在「看注音寫國字」測驗上，非相似字組的表現顯著優於相似字組，但是在「看字讀音」和「選出正確的國字」測驗上，則相似字組與非相似字組之間無顯著性差異；（3）相似字組與非相似字組在「選出正確的國字」之不同題型（相似字組組合之選項與非相似字組組合之選項）的反應上並未達到顯著性差異，此外，兩實驗組皆一致顯示在相似字組組合之選項的錯誤率高於非相似字組組合之選項。根據上述結果，本研究對中文生字教材設計及未來相關研究提出數點建議。

The purpose of this study was to test whether Chinese characters are best learned when presented in groups of similar characters(e.g., 官, 宮, 客, 容 ) or in groups of dissimilar groups(e.g., 京, 核, 官, 問 ). This was a quasi-experimental study using pretests and posttests. Participants were four classes of first grade students in one elementary school in Taiwan, all taught by the same teacher. Two classes were taught the 16 characters in 4 graphemically similar sets(GS) and the other two classes were taught the same 16 characters in 4 graphemically dissimilar sets (GD). Participants' learning was measured by three researcher-designed tests. This study used a 2(tasks arrangement) X 3(achievement levels)factorial design. Data analysis employed analysis of covariance(ANCOVA) , analysis of variance(ANOVA), and a t-Test. The error analysis on the three tests used an one-way ANOVA, a t-Test. This study found that the effects grouping graphemically similar Chinese characters versus graphemicall dissimilar were slightly different for participants in different levels of achievement. Specifically, low achievers learned significantly better under the GS condition. A average and high achievers showed significantly superior performance under the GD condition on a Written Graphemic Recall test, but there were no significant differences between GS and GD conditions on the Oral Phonological Recall and Visual Graphemic Recognition tests. In the GD condition, low achievers performed significantly less well than high and average, but the difference between high and average was not significant. In the GS condition. high achievers performed significantly better than average and low, but the difference between average and low was not significant.