Premium
Bias‐calibrated estimation from sample surveys containing outliers
Author(s) -
Welsh A. H.,
Ronchetti Elvezio
Publication year - 1998
Publication title -
journal of the royal statistical society: series b (statistical methodology)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.523
H-Index - 137
eISSN - 1467-9868
pISSN - 1369-7412
DOI - 10.1111/1467-9868.00133
Subject(s) - estimator , outlier , quantile , population , statistics , calibration , sample (material) , sampling bias , robust statistics , mathematics , econometrics , function (biology) , sample size determination , physics , demography , evolutionary biology , sociology , biology , thermodynamics
We discuss the problem of estimating finite population parameters on the basis of a sample containing representative outliers. We clarify the motivation for Chambers's bias‐calibrated estimator of the population total and show that bias calibration is a key idea in constructing estimators of finite population parameters. We then link the problem of estimating the population total to distribution function or quantile estimation and explore a methodology based on the use of Chambers's estimator. We also propose methodology based on the use of robust estimates and a bias‐calibrated form of the Chambers and Dunstan estimator of the population distribution function. This proposal leads to a bias‐calibrated estimator of the population total which is an alternative to that of Chambers. We present a small simulation study to illustrate the utility of these estimators.