羅吉式迴歸分析（logistic regression analysis） 是一種其效標變項為二分變 項（事件發生以 Y=1 表示，不發生以 Y=0 表示）而且可以計算每一受試者某事件發生的概 率估計值之迴歸分析法。在這種羅吉式迴歸分析式中，預測變項與效標變項的概率估計值之 間形成 S 形曲線（羅吉式曲線）的關係。 因此，一般直線迴歸分析中所用的最小平方法並 不宜用來估計羅吉式迴歸分析模式中迴歸係數。本文的目的在透過理論分析和實例演示，以 說明如何使用「最大可能性估計法」（ MLE ）來估計多元羅吉式迴歸（ multiple logistic regression ）模式中的迴歸係收，如何考驗其顯著性，以及如何進行多元羅吉式 迴歸模式的適合度考驗。

　　　　　The logistic regression analysis is one of the standard method for regression analysis when the criterion variable is a dichotomous variable (e.g. Y=1 represents the presents the presence of heart disease, and Y=0 represents the absence of it). The primary purposes of this article, in addition to the discussion of the basic concept of logistic regression, are to demonstrate:(1) how to estimate logistic regression coefficients, with method of maximum likelihood estimation;(2) how to test the significance of the coefficients, with Wald test;(3) how to test the goodness of fit of the logistic regression model. A numerical example was used for these purposes and a plot of S-shaped logistic regression curve was formed to demonstrate how to estimate the conditional probability that the outcome characteristic will occur.