A comparison of two robust estimation methods for business surveys
A comparison of two robust estimation methods for business surveys
Two alternative robust estimation methods often employed by National Statistical Institutes in business surveys are two-sided M-estimation and one-sided Winsorisation, which can be regarded as an approximate implementation of one-sided M-estimation. We review these methods and evaluate their performance in a simulation of a repeated rotating business survey based on data from the Retail Sales Inquiry conducted by the UK Office for National Statistics. One- and two-sided M-estimation are found to have very similar performance, with a slight edge for the former for positive variables. Both methods considerably improve both level and movement estimators. Approaches for setting tuning parameters are evaluated for both methods and this is a more important issue than the difference between the two approaches. M-estimation works best when tuning parameters are estimated using historical data but is serviceable even when only live data is available. Confidence interval coverage is much improved by the use of a bootstrap percentile confidence interval.
270-289
Clark, Robert Graham
1e2838fc-3f4b-44f4-96cb-e234598b55b2
Kokic, Philip
5d275ad2-95c8-420a-afab-a44cf215326d
Smith, Paul A.
a2548525-4f99-4baf-a4d0-2b216cce059c
August 2017
Clark, Robert Graham
1e2838fc-3f4b-44f4-96cb-e234598b55b2
Kokic, Philip
5d275ad2-95c8-420a-afab-a44cf215326d
Smith, Paul A.
a2548525-4f99-4baf-a4d0-2b216cce059c
Clark, Robert Graham, Kokic, Philip and Smith, Paul A.
(2017)
A comparison of two robust estimation methods for business surveys.
International Statistical Review, 85 (2), .
(doi:10.1111/insr.12177).
Abstract
Two alternative robust estimation methods often employed by National Statistical Institutes in business surveys are two-sided M-estimation and one-sided Winsorisation, which can be regarded as an approximate implementation of one-sided M-estimation. We review these methods and evaluate their performance in a simulation of a repeated rotating business survey based on data from the Retail Sales Inquiry conducted by the UK Office for National Statistics. One- and two-sided M-estimation are found to have very similar performance, with a slight edge for the former for positive variables. Both methods considerably improve both level and movement estimators. Approaches for setting tuning parameters are evaluated for both methods and this is a more important issue than the difference between the two approaches. M-estimation works best when tuning parameters are estimated using historical data but is serviceable even when only live data is available. Confidence interval coverage is much improved by the use of a bootstrap percentile confidence interval.
Text
comparison of two robust estimation methods - accepted version.pdf
- Accepted Manuscript
More information
Accepted/In Press date: 10 April 2016
e-pub ahead of print date: 23 June 2016
Published date: August 2017
Organisations:
Statistical Sciences Research Institute
Identifiers
Local EPrints ID: 391619
URI: http://eprints.soton.ac.uk/id/eprint/391619
ISSN: 0306-7734
PURE UUID: 48f5987c-8785-477b-b094-2e2302e69d7d
Catalogue record
Date deposited: 20 Apr 2016 08:57
Last modified: 15 Mar 2024 05:29
Export record
Altmetrics
Contributors
Author:
Robert Graham Clark
Author:
Philip Kokic
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics
