教育心理學的研究日益仰賴統計模型的技術以更深入地探究教育與心理測驗的資料，以及測量變項與內隱無法觀察變項間的關係。如果基本假定可以滿足的話，試題反應理論可以依據既有的參數，複製可觀察的樣本資料。此外，IRT模型允許 在個體的層次上，以個人為單位進行適配度的分析。本文探討個人層次的測量模式適配度統計量，運用仿擬研究及其實資料 (行動導向量表)，說明影響其統計考驗力的可能因素，包括：異常反應組型的類型、題數、試題參數等。文末並討論這些結果的意涵、指出本研究的限制與對未來研究的建議。

Studies of educational psychology apply statistical modeling techniques to explore the test data, and to establish the relationships among the latent variables. The IRT model can reproduce data by the given parameters if the assumption is met. Besides, IRT as a model-based method can detect the individuals inadequately measured by a test. An introduction to person-fit statistics Z1 is provided. The purpose of this study was to find the factors, including types of aberrant response patterns, the number of items, and item parameters, which influence the null distribution of Z1.A simulation study and true data were conducted to investigate the Type I error and the detection rates of Z1.Finally, the results, limitations, and suggestion are discussed.