ON ASYMPTOTICALLY ROBUST COMPETITORS OF THE ONE-SAMPLE T-TEST.
Abstract
Although the t-test is one of the most commonly used statistical procedures, its behavior is somewhat sensitive to the assumption that the observations come from a normal distribution. Recently, it has been shown that 'quick estimators', i.e., estimators which are linear combinations of a few sample quantiles are robust estimators of the location parameter for a large class of symmetric unimodal densities. In order to use the median or any other 'quick estimator' as a test we must estimate its variance, or in large samples its asymptotic variance. The present paper is concerned with estimating (1/f squared) (nu subscript p) where nu subscript p is the true value of the p-th population quantile and f(x) is the density function. The estimator we consider has properties similar to those of Rosenblatt's estimate of a density function.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1967
- Accession Number
- AD0654460
Entities
People
- Daniel A. Bloch
- Joseph L. Gastwirth
Organizations
- Johns Hopkins University