「電腦化中文識字量測驗」之發展與信效度研究

Evaluation of the Psychometric Properties of the Computerized Chinese Character Recognition Test: Reliability and Validity

本研究旨在發展國小學童電腦化中文識字量測驗，用以協助教師快速推估學童的中文識字量能力。本測驗的字庫是採用教育部（2000）第二版的「國小學童常用字詞調查報告書」，並分別建立國小一至六年級分年級的版本，測驗字的選擇採用字級分層隨機抽樣，研究有效樣本包括977位國小一至六年級學童。本研究結果，所建置的國小學童電腦化中文識字量測驗，α信度係數介於 .70～.81，折半信度係數介於 .71～.78，重測信度係數介於 .73～.83之間。一至六年級國小學童的識字量隨年級而成長，且四年級到六年級的識字量與閱讀理解成長測驗的效標關聯效度為 .54～.59；與詞彙成長測驗的效標關聯效度為 .59～.62，顯示本研究所建置之電腦化中文識字量測驗的信度與效度尚佳，期待本測驗可做為教師了解學童識字量之評估工具。

This study was to develop a computer based testing of elementary school student’s of Chinese character recognition to aid teachers appraising the size of Chinese characters rapidly. Primary School Children Common Words Investigation Report by the MOE (2000) were used to establish different editings of test from G1 to G6 by stratified random sampling. A total of 977 effective samples were collected from students of G1 to G6. The result: the Cronbach’s alpha between .70 to .81, the split-half reliability between .71 to .78, the test-retest reliability between .73 to .83. The character recognition of elementary school students grew by the years from G1 to G6, and the empirical validity of reading and vocabulary comprehension was statistically significant between .54 to .59 and .59 to .62 respectively from G4 to G6. The results approved the mechanism of computer based testing with stable reliability and validity. The testing can assess the character recognition of elementary school students.