非理工科系大學生在科學新聞中的統計素養之初探研究

Abstract

隨著大數據時代的發展，各領域在分享資訊時也會使用大量的統計訊息，現代社會的公民能否正確判讀統計資料傳達的訊息相當重要。本研究針對非理工科系大學生了解其閱讀科學新聞文本時，統計素養相關先備知識、CARE統計素養教學指引的學習成效、以及大學時期是否修習統計相關課程對統計素養的影響。研究工具的設計係根據Schield（2010）提出影響統計值需要注意（CARE）的四個面向：「Context變項的脈絡」、「Assembly變項的集合」、「Randomness隨機性」與「Error偏差」，改編成科學新聞中常見的8項統計素養測驗題型與教學概念，分析34位非理工科系大學生在CARE統計素養教學指引前後，統計素養能力的改變。研究結果顯示：(1)非理工科系大學生的統計素養前測中，面對科學新聞中的抽樣偏差與機率的解讀較無問題，但是需要計算的基準值差異以及牽涉到統計學專業名詞的隨機性與混淆變項等三題型，受試者感受到較大的困難。而受試者是否修習統計相關課程分組中，基準值差異與比較類型的選擇兩題型，有修組的表現顯著高於沒修組。(2)教學任務的學習表現上與前測結果相同，受試者需要花較多的心力理解基準值差異、隨機性與混淆變項三個概念。是否修習統計的分組分析與前測結果相同，有修組在基準值差異與比較類型的選擇兩題型顯著高於沒修組。(3)受試者的後測表現與前兩階段表現不同，在面對科學新聞裡的題型，比例的比較與基準值差異題型較難對比教學指引，無法順利解讀語句埋藏的錯誤而使表現下降。受試者是否修習統計的分組則是皆無顯著差異；受試者學習成效在基準值差異、混淆變項、隨機性與偏差共四個題型的後測分數顯著高於前測分數，經過教學任務的指引後，受試者更能掌握科學新聞文本中的埋錯並更正。

With the development of big data, a large amount of statistical information is used when sharing information. It is important for citizens in the modern society to correctly interpret the information conveyed by statistical data. This study explored statistical literacy of non-science-major college students when they read science news. Prior knowledge about statistical literacy, learning outcomes of CARE statistical literacy guidelines, and the influence of whether students take statistics-related courses in college on these measures were examined in this study. The instruments were developed based on Schield (2010), in which factors that influence the interpretation of statistical data were classified into four categories: context, assembly, randomness and error (CARE). We adapted these four factors to create 8 items which are common in science news for statistical literacy test items and concepts to be learned.The results of this study are described as follows. In the pre-test for statistical literacy, non-science-major college students understood the concept of sampling bias and the meaning of probability in science news. But students experienced great difficulties in items concerning baseline value, randomness, and confounding variables. Whether students take statistics-related courses in college resulted in significant differences in items about baseline value and comparison choice. Comprehension levels during instruction were similar to the patterns found in the pre-test. In the post-test, students did not preform well in items about comparison of proportion and baseline value. There was no significant differences between students taking statistics-related courses in college or not. The post-test scores were significantly higher than the pre-test scores in the following four types of questions: baseline value, confounding variables, randomness and errors. After being guided by CARE statisitical literacy guidelines, studentswere better able to detect and correct errors in statistical information mentioned in science news.