Title
Interval Estimation for Proportions
Speaker
Weizhen Wang, Wright State University Department of Mathematics and Statistics
Abstract
Interval estimation for proportions is a basic problem in statistical inference, and is widely used in statistical practice. In this talk, we focus on three related parameters: a single binomial proportion, the difference of two proportions from a matched-pair experiment, and the difference of proportions from two independent binomials. For the case of large sample, we investigate five commonly used intervals, including Agresti-Coull interval (1998), and show that they all have an incorrect asymptotic confidence coefficient. i.e., no matter how large the sample size is, the chance of capturing the parameter may still be below the nominal level by a fixed amount. A concept of uniformly approximate interval is proposed. For small samples, we discuss a construction of one-sided interval, and derive smallest confidence interval for each parameter.
Professor Weizhen Wang received his BS and MS in Peking University in 1987 and 1990, respectively, and completed a PhD in statistics at Cornell University in 1995. After one-year visit at Purdue University, he joined Wright State University and has been a professor of Statistics since 2007. His research interest includes bioequivalence, saturated and adaptive designs, categorical data analysis, foundation of statistics, statistical computation in biochemistry, and dose-response study. Currently he is on a sabbatical leave from Wright State visiting the Division of Biostatistics
