Transformation and smoothing in sample survey data

Yanyuan Ma, Alan H. Welsh

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

We consider model-based prediction of a finite population total when a monotone transformation of the survey variable makes it appropriate to assume additive, homoscedastic errors. As the transformation to achieve this does not necessarily simultaneously produce an easily parameterized mean function, we assume only that the mean is a smooth function of the auxiliary variable and estimate it non-parametrically. The back transformation of predictions obtained on the transformed scale introduces bias which we remove using smearing. We obtain an asymptotic expansion for the prediction error which shows that prediction bias is asymptotically negligible and the prediction mean-squared error (MSE) using a non-parametric model remains in the same order as when a parametric model is adopted. The expansion also shows the effect of smearing on the prediction MSE and can be used to compute the asymptotic prediction MSE. We propose a model-based bootstrap estimate of the prediction MSE. The predictor produces competitive results in terms of bias and prediction MSE in a simulation study, and performs well on a population constructed from an Australian farm survey.

Original languageEnglish (US)
Pages (from-to)496-513
Number of pages18
JournalScandinavian Journal of Statistics
Volume37
Issue number3
DOIs
StatePublished - Sep 1 2010

Fingerprint

Sample Survey
Survey Data
Prediction Error
Smoothing
Mean Squared Error
Prediction
Model-based
Auxiliary Variables
Finite Population
Nonparametric Model
Parametric Model
Smooth function
Estimate
Bootstrap
Asymptotic Expansion
Survey data
Sample survey
Predictors
Monotone
Simulation Study

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

@article{9c296f5a87594981949f1fd30883b58f,
title = "Transformation and smoothing in sample survey data",
abstract = "We consider model-based prediction of a finite population total when a monotone transformation of the survey variable makes it appropriate to assume additive, homoscedastic errors. As the transformation to achieve this does not necessarily simultaneously produce an easily parameterized mean function, we assume only that the mean is a smooth function of the auxiliary variable and estimate it non-parametrically. The back transformation of predictions obtained on the transformed scale introduces bias which we remove using smearing. We obtain an asymptotic expansion for the prediction error which shows that prediction bias is asymptotically negligible and the prediction mean-squared error (MSE) using a non-parametric model remains in the same order as when a parametric model is adopted. The expansion also shows the effect of smearing on the prediction MSE and can be used to compute the asymptotic prediction MSE. We propose a model-based bootstrap estimate of the prediction MSE. The predictor produces competitive results in terms of bias and prediction MSE in a simulation study, and performs well on a population constructed from an Australian farm survey.",
author = "Yanyuan Ma and Welsh, {Alan H.}",
year = "2010",
month = "9",
day = "1",
doi = "10.1111/j.1467-9469.2010.00691.x",
language = "English (US)",
volume = "37",
pages = "496--513",
journal = "Scandinavian Journal of Statistics",
issn = "0303-6898",
publisher = "Wiley-Blackwell",
number = "3",

}

Transformation and smoothing in sample survey data. / Ma, Yanyuan; Welsh, Alan H.

In: Scandinavian Journal of Statistics, Vol. 37, No. 3, 01.09.2010, p. 496-513.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Transformation and smoothing in sample survey data

AU - Ma, Yanyuan

AU - Welsh, Alan H.

PY - 2010/9/1

Y1 - 2010/9/1

N2 - We consider model-based prediction of a finite population total when a monotone transformation of the survey variable makes it appropriate to assume additive, homoscedastic errors. As the transformation to achieve this does not necessarily simultaneously produce an easily parameterized mean function, we assume only that the mean is a smooth function of the auxiliary variable and estimate it non-parametrically. The back transformation of predictions obtained on the transformed scale introduces bias which we remove using smearing. We obtain an asymptotic expansion for the prediction error which shows that prediction bias is asymptotically negligible and the prediction mean-squared error (MSE) using a non-parametric model remains in the same order as when a parametric model is adopted. The expansion also shows the effect of smearing on the prediction MSE and can be used to compute the asymptotic prediction MSE. We propose a model-based bootstrap estimate of the prediction MSE. The predictor produces competitive results in terms of bias and prediction MSE in a simulation study, and performs well on a population constructed from an Australian farm survey.

AB - We consider model-based prediction of a finite population total when a monotone transformation of the survey variable makes it appropriate to assume additive, homoscedastic errors. As the transformation to achieve this does not necessarily simultaneously produce an easily parameterized mean function, we assume only that the mean is a smooth function of the auxiliary variable and estimate it non-parametrically. The back transformation of predictions obtained on the transformed scale introduces bias which we remove using smearing. We obtain an asymptotic expansion for the prediction error which shows that prediction bias is asymptotically negligible and the prediction mean-squared error (MSE) using a non-parametric model remains in the same order as when a parametric model is adopted. The expansion also shows the effect of smearing on the prediction MSE and can be used to compute the asymptotic prediction MSE. We propose a model-based bootstrap estimate of the prediction MSE. The predictor produces competitive results in terms of bias and prediction MSE in a simulation study, and performs well on a population constructed from an Australian farm survey.

UR - http://www.scopus.com/inward/record.url?scp=77955747848&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77955747848&partnerID=8YFLogxK

U2 - 10.1111/j.1467-9469.2010.00691.x

DO - 10.1111/j.1467-9469.2010.00691.x

M3 - Article

AN - SCOPUS:77955747848

VL - 37

SP - 496

EP - 513

JO - Scandinavian Journal of Statistics

JF - Scandinavian Journal of Statistics

SN - 0303-6898

IS - 3

ER -