A noninformative bayesian approach for selecting a good post-stratification

Patrick Zimmerman, Glen D Meeden

Research output: Contribution to journalArticle

Abstract

In the standard design approach to survey sampling prior information is often used to stratify the population of interest. A good choice of the strata can yield significant improvement in the resulting estimator. However, if there are several possible ways to stratify the population, it might not be clear which is best. Here we assume that before the sample is taken a limited number of possible stratifications have been defined. We will propose an objective Bayesian approach that allows one to consider these several different possible stratifications simultaneously. Given the sample the posterior distribution will assign more weight to the good stratifications and less to the others. Empirical results suggest that the resulting estimator will typically be almost as good as the estimator based on the best stratification and better than the estimator which does not use stratification. It will also have a sensible estimate of precision.

Original languageEnglish (US)
Pages (from-to)2515-2536
Number of pages22
JournalElectronic Journal of Statistics
Volume12
Issue number2
DOIs
StatePublished - Jan 1 2018

Fingerprint

Post-stratification
Stratification
Bayesian Approach
Estimator
Survey Sampling
Prior Information
Posterior distribution
Assign
Bayesian approach
Estimate

Keywords

  • Finite population sampling
  • Prior information
  • Stepwise Bayes
  • Stratification

Cite this

A noninformative bayesian approach for selecting a good post-stratification. / Zimmerman, Patrick; Meeden, Glen D.

In: Electronic Journal of Statistics, Vol. 12, No. 2, 01.01.2018, p. 2515-2536.

Research output: Contribution to journalArticle

@article{1c576a4822124844b49dad87549c54cb,
title = "A noninformative bayesian approach for selecting a good post-stratification",
abstract = "In the standard design approach to survey sampling prior information is often used to stratify the population of interest. A good choice of the strata can yield significant improvement in the resulting estimator. However, if there are several possible ways to stratify the population, it might not be clear which is best. Here we assume that before the sample is taken a limited number of possible stratifications have been defined. We will propose an objective Bayesian approach that allows one to consider these several different possible stratifications simultaneously. Given the sample the posterior distribution will assign more weight to the good stratifications and less to the others. Empirical results suggest that the resulting estimator will typically be almost as good as the estimator based on the best stratification and better than the estimator which does not use stratification. It will also have a sensible estimate of precision.",
keywords = "Finite population sampling, Prior information, Stepwise Bayes, Stratification",
author = "Patrick Zimmerman and Meeden, {Glen D}",
year = "2018",
month = "1",
day = "1",
doi = "10.1214/18-EJS1461",
language = "English (US)",
volume = "12",
pages = "2515--2536",
journal = "Electronic Journal of Statistics",
issn = "1935-7524",
publisher = "Institute of Mathematical Statistics",
number = "2",

}

TY - JOUR

T1 - A noninformative bayesian approach for selecting a good post-stratification

AU - Zimmerman, Patrick

AU - Meeden, Glen D

PY - 2018/1/1

Y1 - 2018/1/1

N2 - In the standard design approach to survey sampling prior information is often used to stratify the population of interest. A good choice of the strata can yield significant improvement in the resulting estimator. However, if there are several possible ways to stratify the population, it might not be clear which is best. Here we assume that before the sample is taken a limited number of possible stratifications have been defined. We will propose an objective Bayesian approach that allows one to consider these several different possible stratifications simultaneously. Given the sample the posterior distribution will assign more weight to the good stratifications and less to the others. Empirical results suggest that the resulting estimator will typically be almost as good as the estimator based on the best stratification and better than the estimator which does not use stratification. It will also have a sensible estimate of precision.

AB - In the standard design approach to survey sampling prior information is often used to stratify the population of interest. A good choice of the strata can yield significant improvement in the resulting estimator. However, if there are several possible ways to stratify the population, it might not be clear which is best. Here we assume that before the sample is taken a limited number of possible stratifications have been defined. We will propose an objective Bayesian approach that allows one to consider these several different possible stratifications simultaneously. Given the sample the posterior distribution will assign more weight to the good stratifications and less to the others. Empirical results suggest that the resulting estimator will typically be almost as good as the estimator based on the best stratification and better than the estimator which does not use stratification. It will also have a sensible estimate of precision.

KW - Finite population sampling

KW - Prior information

KW - Stepwise Bayes

KW - Stratification

UR - http://www.scopus.com/inward/record.url?scp=85051075094&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85051075094&partnerID=8YFLogxK

U2 - 10.1214/18-EJS1461

DO - 10.1214/18-EJS1461

M3 - Article

AN - SCOPUS:85051075094

VL - 12

SP - 2515

EP - 2536

JO - Electronic Journal of Statistics

JF - Electronic Journal of Statistics

SN - 1935-7524

IS - 2

ER -