在醫學應用領域中，常有相關性的二元資料，例如家庭中成員是否患有癌症或是否患有白內障 ，而強韌概似函數方法(robust likelihood function)和廣義估計方程式(Generalized estimating equation, GEE) 可以很好的分析這些有相關性的二元資料。 本文的主要目的是在Tsou and Hsiao(2017)發表的論文基礎上，對強韌概似函數方法與廣義估計方程式進一步做更細緻的比較。本文通過模擬與實例分析比較了強韌概似函數方法和廣義估計方程式的型一誤差率、信賴區間上下界與覆蓋率，發現廣義估計方程在小樣本時會因為虛無假設的不同得出相反的結論。 ;In the field of medical applications, there are often related binary data, such as whether a family member has cancer or whether there is a cataract, and the robust likelihood function method and the generalized estimating equation can well analyze these correlated binary data. The main purpose of this paper is to make a more detailed comparison between the robust likelihood function method and the generalized estimating equation(GEE) based on the paper published by Tsou and Hsiao (2017). This paper compares the type I error rate, the upper and lower bounds of the confidence interval and the coverage rate between the robust likelihood function method and the generalized estimation equation through simulation and case analysis.It is found that the generalized estimating equation will draw the opposite conclusion due to the difference of null hypothesis when the sample is small.
