「中文線上識字測驗」發展及其在華語遠距教學上的應用

Wed-based Chinese Character Recognition Assessment and Its Application on Distance Education of Chinese

本研究之主要目的為發展一套以中文為第二語言或外語學習者的線上識字測 驗，以應用於華語遠距教學上。本測驗包含120 個字，且是從不同頻率層次下隨機 選出。本研究有兩個主要發現： 第一，在信效度研究方面，結果顯示Cronbach’s Alpha 與折半信度相當良好， 間隔一週之再測信度也具有高度相關。在效度研究結果則發現，本測驗具效標關聯 效度（效標有三項：「中文認字測驗」、「華語聽說讀寫測驗」，以及不同CEFR 等級的參與者），支持線上識字測驗分數作為評估整體華語能力指標的有效性。 第二，本測驗包含本研究識字測驗包含兩種計分：完全計分（該字聲韻調全對 則得分）與部分計分（該字聲韻全對即得分）。結果發現若採用的是部分計分，識 字測驗分數與聽說讀寫測驗分測驗的相關係數一致高於完全計分，此顯示若是採部 分計分，可提高題目鑑別度與測驗分數效度，且更能夠敏感反映CSL ／ CFL 學習 者的華語文能力。本研究結果顯示，理論上中文字層級是中文知識之核心成分，而 在輔助華語教學實務方面，本研究發展的線上識字測驗可作為評估CSL ／ CFL 學 習者華語能力的有用工具。

The study aims at establishing an efficient， valid， and computerized Chinese character recognition assessment in order to meet the increasing needs for online tests in distance education of Chinese. We constructed a Web-based Chinese Character Recognition Assessment system， compiling a test for Chinese Character Recognition Test which included 120 characters from different levels of character frequency in this system. The duration of the assessment was eight minutes. The participants were asked to report the pronunciations of each character that were shown on the screen one at a time. One-week test-retest reliability showed a strong positive correlation， and Cronbach’s Alpha and the split-half reliability demonstrated that this assessment items have good internal consistency. Moreover， a criterion-related validity estimator revealed that the scores of this assessment had significantly positive correlations with three criterions: Graded Chinese Character Recognition Test (Huang， 2001)， Tests of Chinese Listening， Speaking， Reading， and Writing (Extension school of Continuing Education of National Taiwan Normal University， 2010)， and the three groups of different levels of CEFR (below A1， A1， A2 and above). The results suggest that this online character recognition assessment is valid to evaluate overall Chinese language skills. Moreover， the assessment in this study contains two scoring methods: Tone scoring and non-tone scoring. The results indicate that the tone scoring method already possesses the function of distinguishing the level of students’ Chinese competence (because the score for the character recognition test has a significantly positive correlation with the scores for listening， speaking， reading， and writing tests)， and if the non-tone scoring method is adopted (i.e. taking no account of the correctness of tone)， the discrimination of items and the validity of testing scores would be enhanced. Theoretically， this study supports that character-level knowledge is one of crucial components of Chinese language knowledge. Practically， this online character recognition assessment is an efficient and valid tool to evaluate the general Chinese ability of those who take Chinese as a second/foreign language.