臺灣博碩士論文加值系統

本論文探討古典測驗理論中信度的意義,並討論在一般的情況下Cronbach's alpha 係數將如何低估真實信度,在假設試題反應符合 Rasch's logistic 模型下模擬,用 Cronbach's alpha 係數來估計信度,其估計值與理論信度相差卻不大,所以進一步推導,當試題只有兩題時,Cronbach's alpha 係數與理論信度相差不會大於 0.336 ,而當試題題數增加時,兩者差會減少。
論文中亦驗證文獻中常被提及影響 Cronbach's \alpha 係數的三個因素,
看如何影響信度的大小。其中增加題數會使信度增高部分,本文提出了與文獻中不同的看法,若加考了一題與欲測驗的能力無關時,反而會使信度降低,又如大型考試中,出錯試題時,通常的處理方式為全送分,其結果與有沒有考這一道題相同,信度維持不變。
又因當代測驗理論提出以試題訊息量(item information)及試卷訊息量(test information)的概念，來作為評定某個試題或整份試卷的測量準確性，倒有取代古典測驗理論的「信度」，作為評定試卷內部一致性指標之勢。故在本論文最後一部份便探討了兩者之間的關連性,針對兩份試卷,同一群受測對象去估計信度,且分別畫出兩份試卷的訊息函數,藉由信度與訊息函數兩個指標去評定試卷是否適合所受測的考生,可得到一致的結果。

緒論............................................vi
1.古典測驗理論.......................................1
2.Cronbach's alpha 係數..............................5
2.1 古典測驗理論基本假設.............................5
2.1.1 古典測驗理論推導結論...........................6
2.2 信度.............................................8
2.3 Cronbach's alpha 係數與信度估計.................12
3信度的探討.........................................26
3.1 影響信度的因素..................................26
3.1.1 題目數量......................................26
3.1.2 受測者間的異同性..............................32
3.1.3 測驗難度......................................34
3.2 Rasch's logistic 模型下的訊息函數...............36
3.2.1 訊息函數的基本概念............................36
3.2.2 試卷編製......................................38
4.結論..............................................42
附錄A 能力參數的估計................................44
附錄B 探討遺失試題反應時的給分方式..................49
參考文獻............................................51

[1]傅粹馨 (2002) 信度、Alpha 係數與相關議題之研究,教育學刊,18期,163-184 。
[2]林妙香 (2000) 教育測驗統計模式淺談,統計薪傳,第一卷第一期,19-34 。
[3]溫敏杰 \& 葉莉莉 (2000) Cronbach $\alpha$ 再研究,成功大學學報第三十五卷人文社會篇,159-170 。
[4]Anastasi, A. (1982) Psychological testing. New York:
Macmillan.
[5] Crocker, L., & Algina, J. (1986) Introduction to classical
and modern test theory. New York: Holt, Rinehart & Win-ston.
[6] Cronbach, L.J. (1951) Coecient alpha and the internal
structure of tests. Psychometrika, 16, 197-334.51
BIBLIOGRAPHY BIBLIOGRAPHY
[7] Eason, S. (1991)Why generalizabilty theory yields better results than classical test theory: A primer with concrete examples. In B. Thompson (Ed.), Advances in educational research: Substantive ndings, methodological developments (Vol. 1, pp. 83-89). Greenwich, CT: JAI Press.
[8] Gullikson, H. (1987) Theory of mental tests. Hillsdale,
NJ: Lawrence Erlbaum Associates. (Originally published in 1950 by New York: John Wiley & Sons)
[9] Hambleton, R.K. (Ed.) (1983) Applications of item response theory. Vancouver, BC: Educational Research Institute of British Columbia.
[10] Hambleton, R.K., and Swaminathan, H. (1985) Item response theory: Principles and Applications. Boston, MA:Kluwer-Nijho.
[11] Lord, F.M. (1977) Practical applications of item character-
istic curve theory. . Journal of educational measurement,14,117-138.
[12] Lord, F.M. (1980) Applications of item response theory to
practical testing problems. Hillsdale, NJ: LEA.52 BIBLIOGRAPHY BIBLIOGRAPHY
[13] Lord, F.M., and Novick, M.R. (1968) Statistical theories of
mental test scores. Reading, MA: Addison-Wesley.
[14] Lord, F.M., and Novick, M.R. (1968) Handling Missing
Item REsponses Due to Item Nonresponse and Incomplete Designs, MA: Addison-Wesley.
[15] Rowley, G.L. (1976) The reliability of observational mea-
sures. American Educational Research Journal, 13, 51-59.
[16] Sax, G. (1989) Principles of educational and psychologi-
cal measurements and evaluation (3rd ed.). Belmont, CA:
Wadsworth.
[17] Thompson, B. (1994) Guidelines for authors. Educational
and Psychological Measurement, 54, 837-847.53