PISA科學素養之試題認知成份分析

A Cognitive Component Analysis for PISA Science Literacy

認知取向的試題難度分析對教學與評量均有實質上的參考價值，本文整合素養架構與文獻，提出科學素養測驗之試題認知成份架構，描述科學素養測驗的試題難度與認知成份特徵，同時藉由不同國家的試題答對比率的差異討論，提供教學與評量的具體參考資訊。研究中所使用的資料庫是PISA 2006國際評量計畫中施測的103題科學試題。研究中以知識類別的數量、知識層次、科學能力與字數四個成份預測科學素養試題難度，此認知成份模式可預測約52%的臺灣試題難度參數變異。在臺灣表現不佳的試題中，研究者提出不熟悉題型、不會使用題目給定的資料或證據形成推論、無法掌握變因以及無法從圖表中掌握答題的關鍵資訊等四項因素。此四項因素大多是因為學生沒有機會接觸，不熟悉此種類型的試題。因此建議日後教師們可以增強這方面的教學。

The item difficulty cognitive component analysis is useful for the development of curriculum and assessment. This study integrates the framework and the literature perspectives to propose and interpret the item cognitive component model for an assessment of science literacy. The 103 items of The Programme for International Student Assessment (PISA)-Science Literacy were used for this analysis. Four cognitive components were proposed to predict the item difficulty parameters: the number of knowledge category, knowledge level, science competencies, and number of words. The results suggest that cognitive components can predict about 52% of item difficulty variance. The implications of these results for items that students did not perform well and for teaching strategies were also discussed.