Robust Analysis of Variance and Analysis of Designed Experiments

本研究是关于方差分析和统计设计实验的稳健方法及其在交互式图形工作站上的实现。第一阶段的研究将集中于规模的稳健估计，析因设计和部分析因设计中的效果对比和误差项的规模稳健估计，方差分量模型的稳健方法，以及交互图形和用户界面的初始设计。第二阶段将处理方差分析中的其他健壮性问题，并开发原型实现。第三阶段将集中精力开发最后的商业软件产品，使广大用户可以使用强有力的差异分析方法。

This research is on robust methods for analysis of variance and statistically designed experiments and their implementation in interactive graphics workstations. The Phase I research will focus on robust estimates of scale, robust estimates of scale for effects contrasts and error term in factorial and fractional factorial designs, robust methods for variance components models, and initial design of interactive graphics and user interface. Phase II would treat other problems of robustness in analysis of variance, and develop prototype implementations. Phase III would concentrate on developing a final commercial software product which would make robust methods of analysis of variance available to wide range of users.