Testing for Authentic Scientific Underst Anding an Informa Tion Processing Approach

自然學科測驗的新視窗（訊息處理論對測驗的貢獻）

蘇育任(Yu-Jen Su)

《測驗統計年刊》 1期 (1993/12) Pp. 153-163

https://doi.org/10.6773/JRMS.199312.0153

摘要

自然科學本質上說來，是一種訊息強度很大的活動，涵蓋了科學知識、組織的概念、專屬術語、解決問題的技能、操作、觀察與解釋等技能。再簡單的一份自然科試卷，都可能包含上述各種成分，以至於測驗時極易使學生的腦筋負荷過重，而無法真正測量學生對科學知識的理解，故自然科測驗的命題與評分必須十分小心。根據目前頗為風行的訊息處理論，考試時學生解答考題的能力乃是受三個認知因素影響，即：（1）個體工作記憶容量的大小，（2）考題中不相干的雜訊掩蓋過主要訊息，和（3）題目中所使用的術語及文字敘述。本研究旨在介紹英國格拉斯哥大學科教中心，所進行的一系列科學測數與評量之成果，首先簡要地探討訊息處理論，接著依據道理論，逐一探討學生在考試時，受那些認知心理因素的影響。概括說來，實徵性研究結果與訊息處理論的推測十分吻合，他們呼籲測驗學者或教師們要注意設計科學測驗時，必須遵照下列建議，以免試題設計不當，變成只測學生的心理能力而不是在測其學科成就。茲將命題時的一些注意事項臚列如下，以供參考：1.此題究竟要測驗什麼？2.要解這道題目，學生須操作的「訊息數量」有多少？3.學生是否會被本題目中太多「無關的雜訊」迷惑，而找不到真正的答案？4.題目中放入的誘答，是否合理？5.題目的文字敘述、專有名詞等是否合宜？

並列摘要

Science, by its very nature, is an information-intense activity having a mass of factual content, organizing concepts, specialist language, problem-solving skills, manual skills, observational skills and interpretive techniques. Even the simplest exam paper can contain components of all of these with the resultant risk of working memory overload and consequently it may no longer measure what we intend it does. It should never be taken lightly to set and score exam papers.Based upon the currently popular perspective of information processing theory, there are three factors which control a pupil's ability to interpret and handle questions: (a) the limitation on the capacity of the working memory, (b) the noise which swamps the signal, and (c) the language or the words used in an item. This study is trying to interlink theoretical arguments with empirical evidence mostly conducted in the Centre for Science Education, University of Glasgow, Scotland. Starting from a brief overview of the information processing theory, this article explores subsequently the afore-mentioned factors which have effect upon the pupils' performances in the exam situations.Research findings seem to be in favour of the arguments based on information processing model and they call for attention that we make sure to measure the authentic knowledge of science subject rather than a set of psychological factors. When setting a test question, the following guidelines are recommended:1. What am I trying to test? 2. How much information has to be manipulated?3. Does the student have to penetrate a noisy situation to get at what the question is asking?4. Is the noise justified by the objectives being tested?5. Am I 'aggravating the problems above by the words in the item?