Estimation in Highly Skewed Data,
Abstract
The problem of inference for the mean of a highly asymmetric distribution is considered. Even with large sample sizes, usual asymptotics (i.e., normal theory) give poor answers, and standard modifications, such as higher moment correction factors, provide little help. We attempt to develop diagnostics to indicate when inferences are likely to be valid, and we examine the performance of several modifications to the standard procedure. The problem is illustrated with data from particle physics.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1992
- Accession Number
- ADP007193
Entities
People
- Shane P. Pederson
Organizations
- Los Alamos National Laboratory