Robust Analysis of Variance and Analysis of Designed Experiments

将进行研究，以产生一个连贯的系统，对设计的实验进行稳健的分析。本课程将研究假设检验和可信区间以及估计器。既要研究混合效应模型和随机效应模型，也要研究固定效应模型。具体地说，将在高细分、高效率的影响和规模估计的基础上，给出一个新的方差表，每条线有三个额外的“均方”。新的均方差将提供有关数据不寻常方面的信息。文中还研究了稳健尺度问题，给出了方差分量的稳健估计方法。在没有复制的情况下，将给出两个到k次方设计的显著效果的测试。为设计性实验提出的交互式动态图形将被评估。

Research will be conducted to produce a coherent system for the robust analysis of designed experiments. Hypothesis tests and confidence intervals as well as estimators will be studied. Mixed effects and random effects models are to be studied as well as fixed effects models. Specifically, a new analysis of variance table with three additional "mean squares" per line will be given based on high-breakdown, high-efficientcy estimates of effects and of scale. The new mean squares will provide information on unusual aspects of the data. Also the robust scale problem will be studied and robust methods for variance components will be given. Tests for significant effects in two to the power k designs without replication will be given. Proposed interactive dynamic graphics for designed experiments will be evaluated.