George Mason University
CDS/CCDS/Statistics Colloquium Series
Seminar Announcement


Better One-Sided Coverage Intervals

Phillip S. Kott

U.S. Department of Agriculture
National Agricultural Statistics Service
Research and Development Division


Research 1, Room 301, Fairfax Campus
George Mason University, 4400 University Drive, Fairfax, VA 22030

Time: 10:30 a.m. Refreshments, 10:45 a.m. Colloquium Talk
Date: March 27, 2009



ABSTRACT

By using an Edgeworth expansion to speed up the asymptotics, we develop one-sided coverage intervals for a proportion based on a stratified simple random sample that work better than well-known competitors. Unlike in the simple random sampling case, it is necessary to estimate the variance of the estimator for the population proportion when the stratum means are allowed to differ. The idealized variance estimator we employ, although not perfect, appears to be useful. Its use with the Edgeworth expansion leads to coverage-interval applications beyond proportions estimated from stratified random samples.