ASYMPTOTIC EFFICIENCY OF A CLASS OF C-SAMPLE TESTS,
Abstract
A c-sample problem is considered by examining the 2-sample problem and generalizing the results obtained by other authors. The resulting L-test for testing the equality of c continuous probaility distributions on the basis of c independent random samples is proposed. Under suitable regularity conditions and certain assumptions concerning the distribution function, the limiting distribution of L is derived. A number of well known tests are included; e.g. the KruskalWallis H-test, Terpestra's K-sample test, Fisher-Yates-Terry-Hoeffding normal score test, and Van der Waerden's 2-sample test. The asymptotic relative efficiencies in Pitman's sense are obtained and shown to be independent of the number of samples. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1963
- Accession Number
- AD0428118
Entities
People
- Madan Lal Puri
Organizations
- New York University