國家教育研究院

　　美國教育部與蘋果、思科、微軟、戴爾電腦等公司，共同創立了二十一世紀關鍵能力聯盟，認為二十一世紀最需要的關鍵能力為：學習與創新的能力、資訊、媒體與科技素養、工作與生活能力。關鍵能力中資訊素養即包含資料的判讀與可能性的預測，對於問題解決的決定有著關鍵性的影響。在此追求e化的世紀，數字是資訊表現的主要媒介，而統計與機率則是解讀和分析數字資訊的重要工具。在國中小學生資料的處理、詮釋與事件可能發生的推測與判斷能力的養成，是培養學生統計素養的先備知識。因此本研究擬編一套跨年段具信效度，適合評量學生資料與可能性能力與其複本常模測驗的工具，做為評估本國學生在第二（三四年級）、第三（五六年級）與第四（七八年級）學習階段，資料與可能性的能力發展情形。並探究與國際間學生資料與可能性的能力差距，作為課程發展、教師教學與評估學生資料與可能性的能力之重要依據。此為第一年研究成果目前已完成預試及分析，初步發現試題發展過程均經歷嚴謹修審提程序，具良好之專家內容效度。試題難度分配廣且與能與學生能力相對應，測驗能區辨不同年段群體，且具理想之分隔信度。

　　US Department of Education and Apple, Cisco, Microsoft, Dell and other companies co-founded twenty-first century key competencies Union, that the twenty-first century the most critical capabilities needed for: learning and innovation ability, information, media and technology literacy, work and life skills. Key competencies in information literacy included data interpretation and the prediction of possibility, which has a critical impact for the decision of problem solving. Numbers is the main medium of information in the pursuit of electronic-century. Statistics and probability is the most important tool for interpretation and analysis of digital information. To cultivate the ability of data analysis, interpretation, and the prediction and judgment of event probably occurring is the prior knowledge of statistical literacy in elementary and junior high school. Therefore, this study intends to compile a set of crossing grade assessment with reliability and validity. The assessment and its duplicate norm tests suit to evaluate the students’ ability of data and chance in order to assess their development in the second stage(third and fourth grade), third stage(fifth and sixth grade) and fourth stage(seventh and eighth grade). In addition to, the project investigate the ability difference of data and chance with international students. The research results will be an important basis for curriculum development, teachers’ instruction as well as evaluation of students’ ability of data and chance. The report is the first year result which had finished pretest data analysis. The finding is that the test has good content validity. Items spread reasonably well along the latent ability and are aligned with various ability levels. The test has satisfactory person separation reliability.