A LOCALLY MOST POWERFUL RANK TEST FOR THE LOCATION PARAMETER OF A DOUBLE EXPONENTIAL DISTRIBUTION

Abstract

The locally most powerful rank test (L.M.P.R.T.) for the location parameter of the two sided exponential cumulative density function is examined. Comparing this test with the likelihood ratio test and making use of Pittman's definition of asymptotic relative efficiency (A.R.E.) we find that Birnbaum's test is asymptotically efficient. The A.R.E. of the latter to the likelihood ratio test is one for symmetric distribution and otherwise is shown to vary between zero and infinity. (Author)

Document Details

Document Type
Technical Report
Publication Date
Oct 01, 1962
Accession Number
AD0290541

Entities

People

  • Eugene Laska

Organizations

  • New York University

Tags

Fields of Study

  • Mathematics

Readers

  • Statistical inference.