Su Haiyan, Liang Hua
Department of Mathematical Sciences, Montclair State University, Montclair, New Jersey 07043, USA.
Comput Stat Data Anal. 2010 Apr 1;54(4):1079-1088. doi: 10.1016/j.csda.2009.10.018.
To compare two treatment effects, which can be described as the difference of the parameters in two linear models, we propose an empirical likelihood-based method to make inference for the difference. Our method is free of the assumptions of normally distributed and homogeneous errors, and equal sample sizes. The empirical likelihood ratio for the difference of the parameters of interest is shown to be asymptotically chi-squared. Simulation experiments illustrate that our method outperforms the published ones. Our method is used to analyze a data set from a drug study.
为了比较两种治疗效果（可描述为两个线性模型中参数的差异），我们提出了一种基于经验似然的方法来对该差异进行推断。我们的方法无需假设误差服从正态分布、误差具有同质性以及样本量相等。结果表明，感兴趣参数差异的经验似然比渐近服从卡方分布。模拟实验表明，我们的方法优于已发表的方法。我们的方法被用于分析一项药物研究的数据集。
