Nonrespondent Subsample Multiple Imputation in Two-Phase Sampling for Nonresponse
Nonresponse is very common in epidemiologic surveys and clinical trials. Common methods for dealing with missing data (e.g., complete-case analysis, ignorable-likelihood methods, and nonignorable modeling methods) rely on untestable assumptions. Nonresponse two-phase sampling (NTS), which takes a ra...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Sciendo
2016-09-01
|
| Series: | Journal of Official Statistics |
| Subjects: | |
| Online Access: | http://www.degruyter.com/view/j/jos.2016.32.issue-3/jos-2016-0039/jos-2016-0039.xml?format=INT |
| id |
doaj-art-54299e3641e04186a353561f64b1e39a |
|---|---|
| recordtype |
oai_dc |
| spelling |
doaj-art-54299e3641e04186a353561f64b1e39a2018-09-02T14:12:33ZengSciendoJournal of Official Statistics2001-73672016-09-0132376978510.1515/jos-2016-0039jos-2016-0039Nonrespondent Subsample Multiple Imputation in Two-Phase Sampling for NonresponseZhang Nanhua0Chen Henian1Elliott Michael R.2Division of Biostatistics & Epidemiology, Cincinnati Children’s Hospital Medical Center, OH 45229, United States of AmericaDepartment of Epidemiology & Biostatistics, College of Public Health, University of South Florida, Tampa, FL 33612-3085, United States of AmericaDepartment of Biostatistics, School of Public Health, University of Michigan and Survey Research Center, Institute for Social Research, University of Michigan, Ann Arbor, MI 48019, United States of AmericaNonresponse is very common in epidemiologic surveys and clinical trials. Common methods for dealing with missing data (e.g., complete-case analysis, ignorable-likelihood methods, and nonignorable modeling methods) rely on untestable assumptions. Nonresponse two-phase sampling (NTS), which takes a random sample of initial nonrespondents for follow-up data collection, provides a means to reduce nonresponse bias. However, traditional weighting methods to analyze data from NTS do not make full use of auxiliary variables. This article proposes a method called nonrespondent subsample multiple imputation (NSMI), where multiple imputation (Rubin 1987) is performed within the subsample of nonrespondents in Phase I using additional data collected in Phase II. The properties of the proposed methods by simulation are illustrated and the methods applied to a quality of life study. The simulation study shows that the gains from using the NTS scheme can be substantial, even if NTS sampling only collects data from a small proportion of the initial nonrespondents.http://www.degruyter.com/view/j/jos.2016.32.issue-3/jos-2016-0039/jos-2016-0039.xml?format=INTDouble samplingmaximum likelihoodmissing datanonignorable missing-data mechanismquality of lifeweighting |
| institution |
Open Data Bank |
| collection |
Open Access Journals |
| building |
Directory of Open Access Journals |
| language |
English |
| format |
Article |
| author |
Zhang Nanhua Chen Henian Elliott Michael R. |
| spellingShingle |
Zhang Nanhua Chen Henian Elliott Michael R. Nonrespondent Subsample Multiple Imputation in Two-Phase Sampling for Nonresponse Journal of Official Statistics Double sampling maximum likelihood missing data nonignorable missing-data mechanism quality of life weighting |
| author_facet |
Zhang Nanhua Chen Henian Elliott Michael R. |
| author_sort |
Zhang Nanhua |
| title |
Nonrespondent Subsample Multiple Imputation in Two-Phase Sampling for Nonresponse |
| title_short |
Nonrespondent Subsample Multiple Imputation in Two-Phase Sampling for Nonresponse |
| title_full |
Nonrespondent Subsample Multiple Imputation in Two-Phase Sampling for Nonresponse |
| title_fullStr |
Nonrespondent Subsample Multiple Imputation in Two-Phase Sampling for Nonresponse |
| title_full_unstemmed |
Nonrespondent Subsample Multiple Imputation in Two-Phase Sampling for Nonresponse |
| title_sort |
nonrespondent subsample multiple imputation in two-phase sampling for nonresponse |
| publisher |
Sciendo |
| series |
Journal of Official Statistics |
| issn |
2001-7367 |
| publishDate |
2016-09-01 |
| description |
Nonresponse is very common in epidemiologic surveys and clinical trials. Common methods for dealing with missing data (e.g., complete-case analysis, ignorable-likelihood methods, and nonignorable modeling methods) rely on untestable assumptions. Nonresponse two-phase sampling (NTS), which takes a random sample of initial nonrespondents for follow-up data collection, provides a means to reduce nonresponse bias. However, traditional weighting methods to analyze data from NTS do not make full use of auxiliary variables. This article proposes a method called nonrespondent subsample multiple imputation (NSMI), where multiple imputation (Rubin 1987) is performed within the subsample of nonrespondents in Phase I using additional data collected in Phase II. The properties of the proposed methods by simulation are illustrated and the methods applied to a quality of life study. The simulation study shows that the gains from using the NTS scheme can be substantial, even if NTS sampling only collects data from a small proportion of the initial nonrespondents. |
| topic |
Double sampling maximum likelihood missing data nonignorable missing-data mechanism quality of life weighting |
| url |
http://www.degruyter.com/view/j/jos.2016.32.issue-3/jos-2016-0039/jos-2016-0039.xml?format=INT |
| _version_ |
1612639580639461376 |