Basic and Applied Statistical Inference

在过去十年中，统计数据推断在确定用于公共政策的基本事实方面发挥了重要作用，例如，估计艾滋病毒感染的流行程度，确定职业或环境暴露导致的癌症风险，以及在法律环境中，例如，涉及就业歧视的案件。这个项目的目的是研究在这些情况下产生的统计问题，并为解决这些问题制订适当的统计方法。一个主要目标是研究估计罕见病患病率的程序，该程序通过在检测前将单个样本组合成批来保持匿名性。这些方法应使卫生专家能够获得公众的合作，因为无法识别被错误分类（假阳性）的感染者和未感染者。将继续强调，当几个科学上似是而非的模型可能是数据的基础，或者当一些观察结果可能是错误的时候，稳健的方法会导致统计上有效的推断。此外，统计方法在分析就业歧视案件中的薪酬、晋升和招聘数据方面继续发挥作用。特别是，将分析沃森诉沃斯堡银行案中各方使用的替代回归方法。

Over the last decade statistical data inference have played an important role in establishing the basic facts used for public policy, e.g. estimating the prevalence of HIV infection, determining the risk of cancer from occupational or environmental exposures, and in the legal setting, e.g. cases concerning employment discrimination. The purpose of this project is to study statistical problems arising in these contexts and to develop appropriate statistical methodologies for their solution. Research on procedures for estimating the prevalence of rare diseases which preserve anonymity by combining individual samples into batches prior to testing is a major goal. These methods should enable health specialists to obtain the cooperation of the public as infected individuals and non-infected persons who are misclassified (false positive) will not be identified. Robust methods which led to statistically valid inferences when several scientifically plausible models may underlie the data or when a few observations may be erroneous will continue to emphasized. In addition, statistical methods continue to play a role in the analysis of pay, promotion and hiring data in employment discrimination cases. In particular, alternative regression approaches will be analyzed that were used by parties in the Watson v. Fort Worth Bank case.