On A Characterization of the Normal Distribution and Its Application to a Goodness-Of-Fit Problem.
Abstract
Goodness-of-fit problems arising in practice often involve nuisance parameters and hence make the hypotheses composite. Exact tests for such problems are hard to come by. A method is proposed in this paper whereby an exact test for normality when both the location and scale parameter are unknown, is possible by eliminating the parameters in a very natural way. A theorem involving a characterization of the normal distribution is stated and proved and its application to a goodness-of-fit study is mentioned. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 02, 1970
- Accession Number
- AD0715084
Entities
People
- V. Seshadri
Organizations
- Southern Methodist University