本文主要是探討應用經驗分布函數對貝他-二項式模式做適合度檢定. 我們提出兩種以經驗分布函數為基礎之統計量的檢定方法, 然後透過蒙地卡羅模擬研究去檢定統計量在對立假設下之不同分布的經驗檢定力. 與 Garren et al. (2001) 的方法相比較, 我們的方法在大多數的情況下是具有較大檢定力的. 最後, 我們應用提出的檢定方法在環境毒性研究的實際資料. Empirical-distribution-function (EDF) goodness-of-fit tests are considered for the beta-binomial model. The testing procedures based on EDF statistics are given. A Monte Carlo study is conducted to investigate the accuracy and power of the tests against various alternative distributions. Our method is found to produce considerably greater power than that of Garren et al. (2001) in most cases. The tests are applied to data sets of the environmental toxicity studies.
