Article ID Journal Published Year Pages File Type
1148199 Journal of Statistical Planning and Inference 2008 9 Pages PDF
Abstract
With reference to the problem of interval estimation of a population mean under model uncertainty, we compare approaches based on robust and empirical statistics via expected lengths of the associated confidence intervals. An explicit expression for confidence intervals arising from a general class of robust statistics is worked out and this is employed to obtain a higher order asymptotic formula for the expected lengths of such intervals. Comparative theoretical results, as well as a simulation study, are then presented.
Related Topics
Physical Sciences and Engineering Mathematics Applied Mathematics
Authors
, ,