An Approximate Bartlett Adjustment for Testing the Goodness of Fit of Multinomial Regression Models
针对最多四个类别且每个设计点有足够重复（八个或以上）的情况，提出了一种调整后的似然比统计量，用于检验多项式回归模型的拟合优度。该调整不依赖于具体模型，适用范围广，尤其适用于比例优势模型。
SUMMARY For up to four categories and given sufficient replications per design point (eight or more seem to be enough), an adjusted likelihood ratio statistic for testing the goodness of fit of multinomial regression models is proposed. This adjustment does not depend on the multinomial model used and therefore will have a wide range of applications. It is certainly useful for the proportional odds model, introduced by McCullagh. This proportional odds model is used in some simulation studies to investigate the performance of the new statistic.