Empirical bootstrap bias correction and estimation of prediction mean square error in small area estimation
- 6 April 2012
- journal article
- research article
- Published by Oxford University Press (OUP) in Biometrika
- Vol. 99 (2) , 457-472
Abstract
We develop a method for bias correction, which models the error of the target estimator as a function of the corresponding estimator obtained from bootstrap samples, and the original estimators and bootstrap estimators of the parameters governing the model fitted to the sample data. This is achieved by considering a number of plausible parameter values, generating a pseudo original sample for each parameter and bootstrap samples for each such sample, and then searching for an appropriate functional relationship. Under certain conditions, the procedure also permits estimation of the mean square error of the bias corrected estimator. The method is applied for estimating the prediction mean square error in small area estimation of proportions under a generalized mixed model. Empirical comparisons with jackknife and bootstrap methods are presented.Keywords
This publication has 9 references indexed in Scilit:
- Jackknife estimation of mean squared error of small area predictors in nonlinear mixed modelsBiometrika, 2009
- Mixed model prediction and small area estimationTEST, 2006
- On Parametric Bootstrap Methods for Small Area PredictionJournal of the Royal Statistical Society Series B: Statistical Methodology, 2006
- Bootstrap diagnostics and remediesCanadian Journal of Statistics, 2006
- A unified jackknife theory for empirical best prediction with M-estimationThe Annals of Statistics, 2002
- Bias reduction of maximum likelihood estimatesBiometrika, 1993
- On bootstrap resampling and iterationBiometrika, 1988
- Variance stabilization and the bootstrapBiometrika, 1988
- Accurate Approximations for Posterior Moments and Marginal DensitiesJournal of the American Statistical Association, 1986