在分層病例對照研究下,針對羅吉斯迴歸模型的適合度檢定,我們考慮在一個廣義的隨機效應模型下建構分數形式檢定統計量。此檢定與梁正憲(2008)所提之動差形式檢定統計量相同。本文進一步地整理Arbogast-Lin檢定,以及實務上常遭誤用的Hosmer-Lemeshow檢定,並針對這幾個檢定進行模擬比較。最後呈述一個關於食道癌研究的範例。 We present the score type test for assessing the adequacy of the logistic regression model in stratified case-control study. This score test is construct based on a random effect model. The proposed test statistic is identical to the moment specification test of Liang(2008). We also review the Arbogast-Lin test and Hosmer-Lemeshow test. The Hosmer-Lemeshow test is valid for prospective data, but if is constantly misapplied for retrospective data. Some comparisons between these methods are performed by simulation studies. An illustration of a research about cancer of the esophagus is provided.