| 中文摘要 |
近年來，效度證據的思維已更加廣泛而多元。本研究希冀透過「以人為中心」的對照組法，擴大納入教學現場教師的參與，作為2010年採用「以試題為中心」書籤標定法所進行四年級自然科標準設定結果的效度覆核。研究對象為全臺八位國小自然科教師及其任教班級的四年級學生233人。研究工具為自然科表現水準評定表和自然科單一題本標準化測驗，前者供教師逐一評定每位學生的自然科學習表現，屬於基礎以下、基礎、精熟、或者進階之其中一群；後者則是施測於教師所任教班級的學生以連結既有量尺。研究資料採用一般化部分給分模式將學生的二分類作答反應和教師判斷的多分類結果同時估計。結果發現，教師心中所認知最低通過標準，比書籤標定法所得的標準寬鬆，而最高通過標準，則較書籤標定法嚴格。此外，教師判斷整體命中率達52.36%，各表現標準的命中率分別達26.32%、57.14%、58.00%以及54.55%，提供一定程度的外部效度證據。最後提供數項建議供未來研究參考。
The concept of validity evidence has become diverse and multifaceted in recent years. The purpose of the present study is to examine the external validity of science assessment standard setting for 4th grade, which was implemented with the bookmark method in 2010. This study uses “contrasting group method”, an examinee-centered method, to set performance standards. The participants were eight elementary school teachers and their 233 students. The instruments were classification sheet and a particular form of science test. Teachers were instructed to judge the performance of students based on the performance level descriptors and mark in the classification sheet with four levels (basic, basic, proficient, and advanced). In order to link the existing scale and teachers’ grading, the particular form of science test was administered to students. Generalized partial credit model was applied to estimate the dichotomous and polytomous data. The results revealed that the minimum standards of basic level set by contrasting group method was lower than that of the bookmark method, while the standard of advanced level by contrasting group method was higher than that of the bookmark method. Besides, the general hit rate was 52.36%, while hit rates of the performance classifications were 26.32%, 57.14%, 58.00%, and 54.55%. In the conclusion, suggestions for further studies are provided. |
本站僅提供期刊文獻檢索。
