Effects of Data Collection Methods on Estimated Household Consumption and Survey Costs : Evidence from an Experiment in the Marshall Islands

In the Pacific, multitopic household surveys have historically gathered expenditure data using open form diaries completed on paper. This methodology is costly to governments, is burdensome for respondents, and takes substantial time to process the...

Full description

Bibliographic Details
Main Authors: Sharp, Michael K., Buffière, Bertrand, Himelein, Kristen, Troubat, Nathalie, Gibson, John
Format: Working Paper
Language:English
Published: World Bank, Washington, DC 2022
Subjects:
Online Access:http://documents.worldbank.org/curated/en/099834004282216747/IDU014657f420d00e042a30a68903563423f5b37
http://hdl.handle.net/10986/37356
id okr-10986-37356
recordtype oai_dc
spelling okr-10986-373562022-04-30T05:10:38Z Effects of Data Collection Methods on Estimated Household Consumption and Survey Costs : Evidence from an Experiment in the Marshall Islands Sharp, Michael K. Buffière, Bertrand Himelein, Kristen Troubat, Nathalie Gibson, John DATA COLLECTION METHODS HOUSEHOLD CONSUMPTION HOUSEHOLD SURVEY SURVEY METHODOLOGY SURVEY DESIGN CAPI HOUSEHOLD INCOME AND EXPENDITURE COMPUTER ASSISTED PERSONAL INTERVIEWING (CAPI) BIAS ECONOMIC STATISTICS In the Pacific, multitopic household surveys have historically gathered expenditure data using open form diaries completed on paper. This methodology is costly to governments, is burdensome for respondents, and takes substantial time to process the results. Noncompliance and partial compliance in diary keeping can artificially inflate poverty measures, biasing economic statistics. This paper reports findings from an experiment in the Marshall Islands comparing the cost and accuracy of several collection methodologies. Variable costs for the status quo diary survey design are between 2.8 and 4.4 times more expensive than a single-visit seven-day recall survey, with the tablet-based diary being even more costly. The highly monitored diaries give similar results to recall but at much greater cost; the status quo yields data of worse quality as effective completion rates with low monitored diaries are only two-thirds the completion rates of recall-based options. Finally, the paper discusses the implementation challenges associated with the different methods in a capacity-constrained environment. 2022-04-29T16:54:00Z 2022-04-29T16:54:00Z 2022-04-28 Working Paper http://documents.worldbank.org/curated/en/099834004282216747/IDU014657f420d00e042a30a68903563423f5b37 http://hdl.handle.net/10986/37356 English Policy Research Working Paper;10029 CC BY 3.0 IGO http://creativecommons.org/licenses/by/3.0/igo World Bank World Bank, Washington, DC Policy Research Working Paper Publications & Research Marshall Islands
repository_type Digital Repository
institution_category Foreign Institution
institution Digital Repositories
building World Bank Open Knowledge Repository
collection World Bank
language English
topic DATA COLLECTION METHODS
HOUSEHOLD CONSUMPTION
HOUSEHOLD SURVEY
SURVEY METHODOLOGY
SURVEY DESIGN
CAPI
HOUSEHOLD INCOME AND EXPENDITURE
COMPUTER ASSISTED PERSONAL INTERVIEWING (CAPI)
BIAS ECONOMIC STATISTICS
spellingShingle DATA COLLECTION METHODS
HOUSEHOLD CONSUMPTION
HOUSEHOLD SURVEY
SURVEY METHODOLOGY
SURVEY DESIGN
CAPI
HOUSEHOLD INCOME AND EXPENDITURE
COMPUTER ASSISTED PERSONAL INTERVIEWING (CAPI)
BIAS ECONOMIC STATISTICS
Sharp, Michael K.
Buffière, Bertrand
Himelein, Kristen
Troubat, Nathalie
Gibson, John
Effects of Data Collection Methods on Estimated Household Consumption and Survey Costs : Evidence from an Experiment in the Marshall Islands
geographic_facet Marshall Islands
relation Policy Research Working Paper;10029
description In the Pacific, multitopic household surveys have historically gathered expenditure data using open form diaries completed on paper. This methodology is costly to governments, is burdensome for respondents, and takes substantial time to process the results. Noncompliance and partial compliance in diary keeping can artificially inflate poverty measures, biasing economic statistics. This paper reports findings from an experiment in the Marshall Islands comparing the cost and accuracy of several collection methodologies. Variable costs for the status quo diary survey design are between 2.8 and 4.4 times more expensive than a single-visit seven-day recall survey, with the tablet-based diary being even more costly. The highly monitored diaries give similar results to recall but at much greater cost; the status quo yields data of worse quality as effective completion rates with low monitored diaries are only two-thirds the completion rates of recall-based options. Finally, the paper discusses the implementation challenges associated with the different methods in a capacity-constrained environment.
format Working Paper
author Sharp, Michael K.
Buffière, Bertrand
Himelein, Kristen
Troubat, Nathalie
Gibson, John
author_facet Sharp, Michael K.
Buffière, Bertrand
Himelein, Kristen
Troubat, Nathalie
Gibson, John
author_sort Sharp, Michael K.
title Effects of Data Collection Methods on Estimated Household Consumption and Survey Costs : Evidence from an Experiment in the Marshall Islands
title_short Effects of Data Collection Methods on Estimated Household Consumption and Survey Costs : Evidence from an Experiment in the Marshall Islands
title_full Effects of Data Collection Methods on Estimated Household Consumption and Survey Costs : Evidence from an Experiment in the Marshall Islands
title_fullStr Effects of Data Collection Methods on Estimated Household Consumption and Survey Costs : Evidence from an Experiment in the Marshall Islands
title_full_unstemmed Effects of Data Collection Methods on Estimated Household Consumption and Survey Costs : Evidence from an Experiment in the Marshall Islands
title_sort effects of data collection methods on estimated household consumption and survey costs : evidence from an experiment in the marshall islands
publisher World Bank, Washington, DC
publishDate 2022
url http://documents.worldbank.org/curated/en/099834004282216747/IDU014657f420d00e042a30a68903563423f5b37
http://hdl.handle.net/10986/37356
_version_ 1764487077843435520