這篇文章主要是想藉由分界點與潛在變數的觀點，來探討如何利用廣義線性混合模型分析有序的反應變數，內容大致可分成兩個部分：第一部分是廣義線性模型的介紹以及隨機效應的探討；第二部分則是其隨機效應的廣義線性混合模型的介紹以及其參數估計的推導過程。研究的過程中，我們發現到分界點與截距項之間是呈現一對一的關係，因此，分界點在某個程度上可以視為是某種型式的截距項。然而，導入潛在變數與分界點的想法，卻能幫助我們更容易去解釋整個模型的架構，以及反應變數與解釋變數的關係。

The focus of this article is on using Generalized Linear Mixed Models to analyze ordinal response data . Two major issues are discussed: the idea of Generalized Linear Models and random effects are review, and estimation problems releted to of Generalized Linear Mixed Models with random terms are investigated . We found that the concept of thresholds and intercepts in the model are indeed equivalent . However using thresholds and latent variables make it easier for us to explain the structure of a model and the relationship between dependent variables and independent variables.