| 摘要: | 本研究旨在編製「臺灣兒童中文文字識詞測驗」，使其運用於在臺灣語言環境下，第一語言為中文的5-10歲學童，補充國內書面文字識詞測驗工具的卻乏。根據Gough和Tunmer(1986)提出的閱讀簡單觀點模式(the Simple View of Reading)，透過本測驗，探討在中文詞彙上，兒童書面詞彙解碼的能力，與其語言理解的關聯。本研究分為預試前訪談、預試和正式施測階段。預試前訪談階段，受試者為5-10歲兒童，試擬87題選擇題後，對10名5-10歲兒童進行一對一訪談，以確保詞彙難度落點，及指導語是否符合兒童理解範圍，並實際讓15位幼稚園大班的學童試答，分為紙本測驗和線上測驗，以方便取樣的方式尋找受試者；並對紙本測驗的受試者進行訪談，觀察進行線上測驗系統的受試者，使用上的流暢度；此外，針對試題請現職教師填寫問卷，評估試題的難易度與適切性。統整訪談受試者意見及教師問卷後，調整87題的選項適切性與難度後，進行預試。預試的受試者為幼稚園大班及國民小學1-4年級的學童，共177個樣本，過程中採團體施測；為了採取電腦化施測，本研究使用PsychoPy軟體建立施測流程，並透過Pavlovia平臺上，提供受試者線上測驗的內容；預試題目共87題。根據古典測驗理論(CTT)和試題反應理論(IRT)，本測驗在預試階段，使用進行古典通過率、古典鑑別度、Rasch難度的分析，作為判定刪題的標準，其中古典通過率的平均值為0.57，古典鑑別度平均值為0.61，Rasch難度的平均值為-0.2，而第51題和第53題，因上述三項指標都偏低，判定為題目太難，故選出85題為正式測驗題目，編制臺灣兒童中文文字識詞測驗。正式施測同樣採取方便取樣，收測對象為幼稚園大班及國民小學1-4年級的學童，共計288個樣本；採用線上測驗，使用電腦、平板為施測工具。正式施測的測驗結果，Rasch難度的平均值為-0.59，古典鑑別度平均值為0.66，古典通過率平均值0.57，整體的平均數48.6，標準差為27.9；進行本研究測驗的同時，也會同時收取「修訂畢保德圖畫詞彙測驗(Peabody Picture Vocabulary Test-Revised,PPVT-R)」做為本研究之效標，因PPVT-R是目前兒童詞彙測驗具代表性，所以作為本測驗之校標關聯效度，是具有說服力的，結果顯示，本測驗與PPVT-R的相關為.771，統計上為顯著，透過PPVT-R證實了本測驗的有效性；在內部一致性的方面，採用Cronbach′s Alpha值的計算結果為.986，折半信度是採用單、雙數把題目分成兩半，分析結果為.98；綜上所述，本測驗具有良好的信度與效度，顯示本測驗是有效測量5-10歲學童的文字識詞能力的；且而本測驗與社會經濟地位(Social-Economic Status, SES)是無顯著結果的。因此本測驗將來是能夠提供教育者，除了聽力測詞彙量之外，另外評估5-10歲學童，在文字識詞能力推估，並有助於教學診斷兒童現階段的文字詞彙能力。;The purpose of this study is to develop a "Chinese Word Recognition Test for Taiwanese children", so that it can be applied to students aged 5-10 whose first language is Chinese in the Taiwanese language environment, which is a supplement to the domestic word recognition test tools. Based on the Simple View of Reading model proposed by Gough and Tunmer(1986), the test compiled in this study, which explores the relationship between children′s ability to decode written vocabulary , and show children′s language understanding in terms of Chinese vocabulary. This study is divided into pre-test interview, pre-test and formal test. In the pre-test interview stage, the subjects were children aged 5-10 years old. After trying 87 multiple choice questions, 10 children aged 5-10 years old were interviewed one-to-one to ensure the difficulty of vocabulary and whether the instruction was in line with the children′s understanding range. In addition, 15 kindergarten students were asked to answer the test, which was divided into paper tests and online tests to find the subjects in a convenient sampling way. The paper test subjects were interviewed, and the fluency of the online test system was observed. In addition, teachers were asked to fill in a questionnaire to evaluate the difficulty and relevance of the questions. After integrating the opinions of the interviewees and the teacher questionnaire, the pre-test was conducted after adjusting the appropriateness and difficulty of the 87 questions. The pre-test subjects were kindergarten senior class and primary school students in grades 1-4, a total of 177 samples, during the process of group testing. In order to take the computerized test, PsychoPy software was used to establish the test process, and the content of online test was provided through the Pavlovia platform. There were 87 pre-test questions. According to Classical Test Theory (CTT) and Item Response Theory (IRT), in the pre-test stage, this test uses the analysis of classical pass rate, classical discrimination and Rasch difficulty as the criteria to determine the deletion of questions. The average value of classical passing rate is 0.57, the average value of classical discrimination is 0.61, and the average value of Rasch difficulty is -0.2. Question 51 and Question 53, due to the above three indicators are low, judged to be too difficult, so 85 formal test questions were selected to compile the Chinese word recognition test for Taiwanese children. A total of 288 samples were collected from kindergarten senior class and primary school students in grades 1-4. Using online test, using computer, tablet as testing tools. For the formal test results, the mean of Rasch difficulty was -0.59, the mean of classical discrimination was 0.66, the mean of classical passing rate was 0.57, the mean of the whole was 48.6, and the standard deviation was 27.9. At the same time, the Peabody Picture Vocabulary test-revised (PPVT-R) will also be used as the criterion of this study, because PPVT-R is a representative Vocabulary Test for children. Therefore, as the calibration correlation validity of this test, it is convincing. The results show that the correlation between this test and PPVT-R is.771, which is statistically significant. The validity of this test is confirmed by PPVT-R. In terms of internal consistency, the calculated result using Cronbach′s Alpha value is.986, while the broken half reliability is divided into two halves using single and even numbers, and the analysis result is.98. In summary, this test has good reliability and validity, indicating that this test is effective in measuring the word recognition ability of children aged 5-10. There was no significant correlation between this test and socioeconomic Status (SES). In the future, this test will be able to provide educators, listening to measure vocabulary, also the assessment of 5 to 10 years old children, in the word recognition ability estimate, and help to teach the diagnosis of children′s current word vocabulary ability. |
