Abstract

Presence of one or more outliers in the observations affect inference procedure in statistical analysis. Designs robust against presence of a single outlier can be found in the literature for both regression and block design set-ups. In this paper, an attempt has been made to find a robust design in a block design set-up for the estimation of a full set of orthonormal treatment contrasts when there are more than one outlier among the observations. It appears that symmetry and balance play an important role in this study; it is seen that as we deviate from them, designs deviate from robustness.

中文翻译：

---

不存在针对受限类别中存在多个异常值的稳健设计

摘要

观察中存在一个或多个异常值会影响统计分析中的推理过程。可以在回归和块设计设置的文献中找到针对单个异常值的存在的稳健设计。在本文中，已尝试在块设计设置中找到稳健的设计，以在观察值中存在多个异常值时估计一整套正交处理对比。看来对称和平衡在这项研究中起着重要作用；可以看出，当我们偏离它们时，设计就会偏离稳健性。