中學生現代公民核心素養之測驗編製

Development of a Modern Citizen Core Literacy Test for Middle School Students

本研究旨在編製一信、效度良好的中學生現代公民核心素養測驗，以評估臺灣中學生現代公民核心素養狀況。測驗編製係參考現代公民核心能力指標以及國際公民教育與素養調查研究（International Civic and Citizenship Education Study [ICCS]）的評量架構，編擬符合中學生生活場域與經驗的試題。試題編製完成後經學科專家確認內容效度與試題預試，於正式施測階段，採分層便利取樣，蒐集來自全臺北、中、南、東（含離島）各區共22所學校的中學生共1479份有效樣本，以多向度部分給分試題反應理論模式（multidimensional partial credit model?MPCM））進行資料分析。結果發現：中學生現代公民核心素養測驗具有良好試題內部一致性，認知題與情感行為題的試題區分信度係數（item separation reliability）皆高於.9，五向度的受試者區分信度係數（person separation reliability），介於.593至.882之間，測驗折半信度為.883。效度驗證程序採用試題反應理論效度驗證架構進行，結果顯示本測驗符合五向度測量建構，同時並提供測驗的內容、結構、本質、類推效度與效標關聯效度，顯示本研究所發展的測驗工具具有良好信度與效度。

The purpose of this study was to design the Modern Citizen Core Literacy Test with well reliability and validity for middle school students. To develop the test, the core competency index as well as the assessment framework of the International Civic and Citizenship Education Survey were used as references. The Modern Citizen Core Literacy Test is developed to reflect the life experiences of middle school students. To establish content validity of the test, a panel of experts from middle schools were invited to review the first version of the test. After a pilot study, the test was administered to a total of 1,479 students from 22 middle schools in Taiwan; they were recruited with stratified purposive sampling. Multi-dimensional Partial Credit Model was used to assess the model-data fit. Five-core literacy were identified: moral literacy, democracy literacy, scientific literacy, media literacy and aesthetic literacy. Moreover, the test showed sufficient internal consistency with person separation reliability ranging from .593 to .882. The Split-half reliability is .883. In addition, the present study adopted item response theory based validation procedure. The results revealed that five dimensions of the Modern Citizen Core Literacy Test exhibited an acceptable model-data fit. Criterion-related validity, interpretability evidence, generalizability evidence, and substantive evidence were all supported, indicating that the test has good reliability and validity.