Abstract

To facilitate the use of data collected through the high-frequency phone surveys on COVID-19, the Living Standards Measurement Study (LSMS) team has created the harmonized datafiles using two household surveys: 1) the country’ latest face-to-face survey which has become the sample frame for the phone survey, and 2) the country’s high-frequency phone survey on COVID-19.

The LSMS team has extracted and harmonized variables from these surveys, based on the harmonized definitions and ensuring the same variable names. These variables include demography as well as housing, household consumption expenditure, food security, and agriculture. Inevitably, many of the original variables are collected using questions that are asked differently. The harmonized datafiles include the best available variables with harmonized definitions.

Two harmonized datafiles are prepared for each survey. The two datafiles are: 1. HH: This datafile contains household-level variables. The information include basic household characterizes, housing, water and sanitation, asset ownership, consumption expenditure, consumption quintile, food security, livestock ownership. It also contains information on agricultural activities such as crop cultivation, use of organic and inorganic fertilizer, hired labor, use of tractor and crop sales.
2. IND: This datafile contains individual-level variables. It includes basic characteristics of individuals such as age, sex, marital status, disability status, literacy, education and work.

Geographic coverage

National

Analysis unit

• Households
• Individuals

Universe

The survey covered all de jure households excluding prisons, hospitals, military barracks, and school dormitories.

Kind of data

Sample survey data [ssd]

Sampling procedure

See “Uganda - National Panel Survey 2019-2020” and “Uganda - High-Frequency Phone Survey on COVID-19 2020-2021” documentations available in the Microdata Library for details.

Mode of data collection

Computer Assisted Personal Interview [capi]

Cleaning operations

Uganda National Panel Survey 2019-2020 and Uganda High-Frequency Phone Survey on COVID-19 2020-2021 data were harmonized following the harmonization guidelines (see “Harmonized Datafiles and Variables for High-Frequency Phone Surveys on COVID-19” for more details).

The high-frequency phone survey on COVID-19 has multiple rounds of data collection. When variables are extracted from multiple rounds of the survey, the originating round of the survey is noted with “_rX” in the variable name, where X represents the number of the round. For example, a variable with “_r3” presents that the variable was extracted from Round 3 of the high-frequency phone survey. Round 0 refers to the country’s latest face-to-face survey which has become the sample frame for the high-frequency phone surveys on COVID-19. When the variables are without “_rX”, they were extracted from Round 0.

Response rate

See “Uganda - National Panel Survey 2019-2020” and “Uganda - High-Frequency Phone Survey on COVID-19 2020-2021” documentations available in the Microdata Library for details.

Abstract

The UNPS aims at producing annual estimates in key policy areas; and providing a platform for experimenting with and assessing national policies and programs. Explicitly, the objectives of the UNPS include:

1. To provide information required for monitoring the National Development Strategy, of major programs such as National Agricultural Advisory Services (NAADS) and General Budget Support, and also to provide information for the compilation of the National Accounts (e.g. agricultural production);

2. To provide high quality nationally representative information on income dynamics at the household level as well as information on service delivery and consumption expenditure estimates annually; to monitor poverty and service outcomes in interim years of other national survey efforts, such as the Uganda National Household Survey (UNHS), Uganda Demographic and Health Survey (UDHS) and National Service Delivery Surveys (NSDS);

3. To provide a framework for low-cost experimentation with different policy interventions to e.g. reduce teacher absenteeism, improve ante-natal and post-natal care, and assess the effect of subsidies on agricultural inputs among others;

4. To provide a framework for policy oriented analysis and capacity building substantiated with the UGDR and support to other research which feed into the Annual Policy Implementation Review; and

5. To facilitate randomized impact evaluations of interventions whose effects cannot currently be readily assessed through the existing system of national household surveys.

Geographic coverage

National

Analysis unit

The study describes (including but not limited to):

• Household

• Individual

• Parcel

• Plot

• Community

Kind of data

Sample survey data [ssd]

Sampling procedure

The UNPS is carried out over a twelve-month period (a “wave”) on a nationally representative sample of households, for the purpose of accommodating the seasonality associated with the composition of and expenditures on consumption. The survey is conducted in two visits in order to better capture agricultural outcomes associated with the two cropping seasons of the country. The UNPS therefore interviews each household twice in a year, in visits approximately six months apart.

In 2009/10, the UNPS set out to track and interview 3,123 households that were distributed over 322 Enumeration Areas (EAs), selected out of 783 EAs that had been visited during the Uganda National Household Survey (UNHS) in 2005/06. The distribution of the EAs covered by the 2009/10 UNPS was such that it included all 34 EAs in Kampala District, and 72 EAs (58 rural and 14 urban) in each of the other regions i.e. Central excluding Kampala, Eastern, Western and Northern which make up the strata.

Within each stratum, the EAs were selected with equal probability with implicit stratification by urban/rural and district (in this order). However, the probabilities of selection for the rural portions of ten districts that had been oversampled by the UNHS 2005/06 were adjusted accordingly. Since most IDP (Internally Displaced People) camps in the Northern region are currently unoccupied, the EAs that constituted IDP camps were not part of the UNPS sample. This allocation allows for reliable estimates at the national, rural-urban and regional levels i.e. at level of strata representativeness which includes: (i) Kampala City, (ii) Other Urban Areas, (iii) Central Rural, (iv) Eastern Rural, (v) Western Rural, and (vi) Northern Rural.

In the UNPS 2010/11, the concept of Clusters instead of EAs was introduced. A cluster represents a group of households that are within a particular geographical area up to parish level. This was done due to split-off households that fell outside the selected EAs but could still be reached and interviewed if they still resided within the same parish as the selected EA. Consequently, in each subsequent survey wave, a subset of individuals was selected for tracking.

The UNPS is part of the long term Census and Household Survey Program hence questionnaires and the timing of data collection are coordinated with the current surveys and census implemented by UBOS.

SAMPLE REFRESH

Starting with the UNPS 2013/14 (Wave 4) fieldwork, one third of the initial UNPS sample was refreshed with the intention to balance the advantages and shortcomings of panel surveys. Each new household will be visited for three consecutive waves, while baseline households will have a longer history of five or six years, given the start time of the sample refresh. This same sample was used for the UNPS 2015/16 (Wave 5)

Once a steady state is reached, each household will be visited for three consecutive years, and at any given time one third of the households will be new, one third will be visited for the second time, and one third for the third (and last) time. The total sample will never be too different from a representative cross-section of the country, yet two-thirds of it will be a panel with a background of a year or two.

New households were identified using the updated sample frames developed by the UBOS in 2013 as part of the preparations for the 2014 Uganda Population and Housing Census.

Mode of data collection

Computer Assisted Personal Interview [capi]

Research instrument

The UNPS had six questionnaires namely: Household Questionnaire; Woman Questionnaire; Agriculture & Livestock Questionnaire; Fisheries Questionnaire; Community Questionnaire and Market Questionnaire. Each of these questionnaires is divided into a number of sections and the number of questions in each section varies accordingly.

It should be noted that in 2013/14, 2015/16 and 2018/19, all questionnaires were administered using the CAPI software except the Fisheries and Market Questionnaires which were not administered.

Abstract

The UNPS aims at producing annual estimates in key policy areas; and providing a platform for experimenting with and assessing national policies and programs. Explicitly, the objectives of the UNPS include:

1. To provide information required for monitoring the National Development Strategy, of major programs such as National Agricultural Advisory Services (NAADS) and General Budget Support, and also to provide information for the compilation of the National Accounts (e.g. agricultural production);

2. To provide high quality nationally representative information on income dynamics at the household level as well as information on service delivery and consumption expenditure estimates annually; to monitor poverty and service outcomes in interim years of other national survey efforts, such as the Uganda National Household Survey (UNHS), Uganda Demographic and Health Survey (UDHS) and National Service Delivery Surveys (NSDS);

3. To provide a framework for low-cost experimentation with different policy interventions to e.g. reduce teacher absenteeism, improve ante-natal and post-natal care, and assess the effect of subsidies on agricultural inputs among others;

4. To provide a framework for policy oriented analysis and capacity building substantiated with the UGDR and support to other research which feed into the Annual Policy Implementation Review; and

5. To facilitate randomized impact evaluations of interventions whose effects cannot currently be readily assessed through the existing system of national household surveys.

Geographic coverage

National Coverage.

Analysis unit

The study describes (including but not limited to):

• Household

• Individual

• Parcel

• Plot

• Community

Kind of data

Sample survey data [ssd]

Sampling procedure

The UNPS is carried out over a twelve-month period (a “wave”) on a nationally representative sample of households, for the purpose of accommodating the seasonality associated with the composition of and expenditures on consumption. The survey is conducted in two visits in order to better capture agricultural outcomes associated with the two cropping seasons of the country. The UNPS therefore interviews each household twice in a year, in visits approximately six months apart.

In 2009/10, the UNPS set out to track and interview 3,123 households that were distributed over 322 Enumeration Areas (EAs), selected out of 783 EAs that had been visited during the Uganda National Household Survey (UNHS) in 2005/06. The distribution of the EAs covered by the 2009/10 UNPS was such that it included all 34 EAs in Kampala District, and 72 EAs (58 rural and 14 urban) in each of the other regions i.e. Central excluding Kampala, Eastern, Western and Northern which make up the strata.

Within each stratum, the EAs were selected with equal probability with implicit stratification by urban/rural and district (in this order). However, the probabilities of selection for the rural portions of ten districts that had been oversampled by the UNHS 2005/06 were adjusted accordingly. Since most IDP (Internally Displaced People) camps in the Northern region are currently unoccupied, the EAs that constituted IDP camps were not part of the UNPS sample. This allocation allows for reliable estimates at the national, rural-urban and regional levels i.e. at level of strata representativeness which includes: (i) Kampala City, (ii) Other Urban Areas, (iii) Central Rural, (iv) Eastern Rural, (v) Western Rural, and (vi) Northern Rural.

In the UNPS 2010/11, the concept of Clusters instead of EAs was introduced. A cluster represents a group of households that are within a particular geographical area up to parish level. This was done due to split-off households that fell outside the selected EAs but could still be reached and interviewed if they still resided within the same parish as the selected EA. Consequently, in each subsequent survey wave, a subset of individuals was selected for tracking.

The UNPS is part of the long term Census and Household Survey Program hence questionnaires and the timing of data collection are coordinated with the current surveys and census implemented by UBOS.

SAMPLE REFRESH

Starting with the UNPS 2013/14 (Wave 4) fieldwork, one third of the initial UNPS sample was refreshed with the intention to balance the advantages and shortcomings of panel surveys. Each new household will be visited for three consecutive waves, while baseline households will have a longer history of five or six years, given the start time of the sample refresh. This same sample was used for the UNPS 2015/16 (Wave 5)

Once a steady state is reached, each household will be visited for three consecutive years, and at any given time one third of the households will be new, one third will be visited for the second time, and one third for the third (and last) time. The total sample will never be too different from a representative cross-section of the country, yet two-thirds of it will be a panel with a background of a year or two.

New households were identified using the updated sample frames developed by the UBOS in 2013 as part of the preparations for the 2014 Uganda Population and Housing Census.

Mode of data collection

Computer Assisted Personal Interview [capi]

Research instrument

The UNPS had six questionnaires namely: Household Questionnaire; Woman Questionnaire; Agriculture & Livestock Questionnaire; Fisheries Questionnaire; Community Questionnaire and Market Questionnaire. Each of these questionnaires is divided into a number of sections and the number of questions in each section varies accordingly.

It should be noted that in 2013/14 and 2015/16, all questionnaires were administered using the CAPI software except the Fisheries and Market Questionnaires which were not administered.

Abstract

Uganda has experienced strong economic growth over the past two decades and has made great strides towards improving the quality of life and access to services. In order to continue to promote pro-poor economic growth, the Government of Uganda (GoU) developed the National Development Plan (NDP) and a Joint Budget Support strategy as part of the implementation of the National Development Strategy (NDS).

Uganda recognizes the need for adequate data collection to effectively monitor outcomes of the National Development Strategy (NDS). For this purpose, the Uganda Bureau of Statistics (UBOS) implemented the Uganda National Panel Survey (UNPS) program, with financial and technical support from the Government of Netherlands, and the World Bank Living Standards Measurement Study - Integrated Surveys on Agriculture (LSMS-ISA) project.

The 2005-2009/10 Uganda National Panel Survey (UNPS) is the first "panel" survey done by the LSMS group. The households included were chosen from the 2005-06 Uganda National Household Survey (UNHS).
The sample One of the primary uses of the UNPS is to inform policymaking in advance of the Budget, through descriptive reports that are made ready in time for the initial work on sector budget framework papers.

Survey Objectives The UNPS aimed at producing annual estimates of outcomes and output in the key policy areas and at providing a platform for the experimentation and assessment of national policies and programs.

Explicitly, the objectives of the UNPS include: The UNPS aims at producing annual estimates in key policy areas and at providing a platform for experimenting with and assessing of national policies and programs. Explicitly, the objectives of the UNPS include: 1. To provide information required for monitoring the National Development Strategy, of major programs such as National Agricultural Advisory Services (NAADS) and General Budget Support, and also to provide information to the compilation of the National Accounts (e.g. agricultural production). 2. To provide high quality nationally representative information on income dynamics at the household level and provide annual information on service delivery and consumption expenditure estimates to monitor poverty and service outcomes in interim years of other national survey efforts, such as the Uganda National Household Survey (UNHS), Uganda Demographic and Health Survey (UDHS) and National Service Delivery Surveys (NSDS). 3. To provide a framework for low-cost experimentation with different policy interventions to e.g. reduce teacher absenteeism, improve ante- and post-natal care, or assessing the effect of agricultural input subsidies. 4. To provide a framework for policy-oriented analysis and capacity building substantiated with the UGDR and support to other research which will feed into the Annual Policy Implementation Review. 5. To facilitate randomized impact evaluations of interventions whose effects cannot currently be readily assessed through the existing system of national household surveys.

Geographic coverage

National

Analysis unit

• Households
• Individuals
• Markets
• Communities / facilities

Kind of data

Sample survey data [ssd]

Sampling procedure

Survey Design The UNPS is carried out annually, over a twelve-month period on a nationally representative sample of households, for the purpose of accommodating the seasonality associated with the composition of and expenditures on consumption. The survey is conducted in two visits in order to better capture agricultural outcomes associated with the two cropping seasons of the country. The UNPS will therefore interview each household twice each year, in visits six months apart.

Starting in 2005-09/10, the UNPS has been set out to track and re-interview 3,123 households that were distributed over 322 enumeration areas (EAs), selected out of the 783 EAs that had been visited by the Uganda National Household Survey (UNHS) in 2005-06. The UNPS EAs covered all 34 EAs visited by the UNHS 2005-06 in Kampala District, and 72 EAs (58 rural and 14 urban) in each of the (i) Central Region with the exception of Kampala District, (ii) Eastern Region, (iii) Western Region, and (iv) Northern Region.

Within each stratum, the UNPS EAs were selected from the UNHS 2005-06 EAs with equal probability, and with implicit stratification by urban/rural and district (in this order), except for the rural portions of the ten districts that were oversampled by the UNHS 2005-06. In these districts, the probabilities were deflated, to bring them back to the levels originally intended. Since IDP camps are now mostly unoccupied, the extra EAs in IDP camps are not a part of the UNPS subsample. This allocation strives for reasonably reliable estimates for the rural portion of each region, and for the set of urban areas out of Kampala as a whole, as well as the best possible estimates for Kampala that can be expected from a subsample of the UNHS 2005-06. Therefore, the UNPS strata of representativeness include (i) Kampala City, (ii) Other Urban Areas, (iii) Central Rural, (iv) Eastern Rural, (v) Western Rural, and (vi) Northern Rural.

Prior to the start of the 2005-09/10 field work, 2 UNPS households were also randomly selected in each EA for the purposes of tracking baseline individuals that moved away from original locations since the UNHS 2005-06. The initial UNPS sample was subject to three consecutive waves of data collection after which, parts of the sample was replaced by new households extracted from the updated sample frames developed by the UBOS as part of the 2012 Uganda Population and Housing Census. In addition, the UNPS will fit within the Long-Term Census and Household Survey Program and therefore both the questionnaires and the timing of data collection will be coordinated with the current surveys and census implemented by UBOS.

Note: Detailed description of sampling procedure and calculation of panel weights is presented in "Basic Information Document". The document is provided along with this metadata.

Mode of data collection

Face-to-face [f2f]

Research instrument

The UNPS survey was collected using the following five questionnaires: 1. Household Questionnaire 2. Woman Questionnaire 3. Agriculture Questionnaire, (administered to the subset of UNPS households engaged in agricultural activities) 4. Community Questionnaire 5. Market Questionnaire.

Abstract

The National Panel Survey (UNPS) was carried out to collect high quality data on key outcome indicators such as poverty, service delivery, employment and to monitor government's development programs like the National Development Plan (NDP) on an annual basis. The 2010/11 survey collected information on socio-economic characteristics at household, individual and community levels.The NPS 2010/11 was comprised of six modules: the Socio-Economic, Woman, Agriculture, Fisheries, Community and Market Price modules. The survey covered 3,200 households that were scientifically selected and followed for re-interview in 2009/10 and 2010/11. The objectives of the survey were:

1) To provide information required for monitoring the NDP and other development objectives like the JAF, MDGs as well as specific programs such as the National Agricultural Advisory Services (NAADS) among others. 2) To provide high quality nationally representative information on income dynamics at the household level as well as annual consumption expenditure estimates to monitor poverty in years between Uganda National Household Surveys (UNHS). 3) To supply regular data on agriculture in order to characterize and monitor the performance of the agricultural sector.

Geographic coverage

National coverage

Analysis unit

Households

Kind of data

Sample survey data [ssd]

Sampling procedure

The 2010/11 NPS survey maintained the 2009/10 NPS sample design where all the households that were sampled for Wave I (2009/10) were tracked and re-interviewed in Wave II (2010/11). Out of the 7,400 households interviewed during the UNHS 2005/06, 3,200 households were selected for the NPS and the same sample was maintained in both 2009/10 and 2010/11 Panel surveys. During data collection, the population of persons interviewed in Wave II was slightly higher than that of Wave I due to the following reasons: 1. About 20 households that had initially been missed in Wave I were found and successfully interviewed in Wave II. 2. Changes in household composition contributed to the increase in the number of persons that were added to the panel. Most importantly, if a household member split-off from his/her original household (e.g. children leaving home to set up their own household,or a couple separates), all the new households were included/ joined the panel. Inclusion of split-offs was the main way in which panel surveys, maintain sample representativeness over the years. The new households formed are known as Split-off households while the individuals are termed as Movers.

Mode of data collection

Face-to-face paper [f2f]

Response rate

The response rate for the survey is 84%.

Abstract

The overall objective of the UNPS Program is to collect high quality data on key outcome indicators such as poverty, service delivery, governance and employment among others; to monitor Government's development programmes like the NDP and the JAF among others on an annual basis. The specific objectives of the survey are: - To provide information required for monitoring the NDP and other development objectives like the JAF, MDGs as well as specific programs such as the National Agricultural Advisory Services (NAADS) among others. - To provide high quality nationally representative information on income dynamics at the household level as well as annual consumption expenditure estimates to monitor poverty in years between the Uganda National Household Surveys (UNHS) - To supply regular data on agriculture in order to characterize and monitor the performance of the agricultural sector.

Geographic coverage

National coverage

Analysis unit

• Household
• Individual / Person
• Woman aged 15-49
• Community

Sampling procedure

The 2011/12 UNPS survey maintained the 2010/11 UNPS sample design whereby all households that were sampled for Wave II (2010/11) were tracked and re-interviewed in Wave III (2011/12). Out of the 7,400 households interviewed during the UNHS 2005/06, 3,123 households were selected for the panel surveys. As a result, the same sample was maintained in 2009/10, 2010/11 and 2011/12 round of surveys. During data collection, households or individuals that had permanently left the original households to known locations were tracked and interviewed. The new households formed are known as split-off households whereas the individuals are termed as movers.

Mode of data collection

Face-to-face [f2f]

Research instrument

The 2011/12 UNPS questionnaires comprised of the following survey instruments: 1) Household Questionnaire: Core and rotating modules 2) Agriculture Questionnaire (for the subset of UNPS households engaged in agricultural activities) 3) Price Questionnaire 4) Community / Facility Questionnaires for schools, health facilities and other facilities (potentially conducted on a rotating basis and not in every year of the UNPS)

Cleaning operations

The 2011/12 round of UNPS used a computerized system of data collection whereby field staff directly captured information using Ultra Mobile Personal Computers (UMPCs) during data collection. The UMPCs were loaded with a data entry application with in-built range and consistency checks to ensure good quality data. Field Team Leaders run checks on the data while still in the field thereafter electronically transmitting it to UBOS Headquarters for verification. Every team was facilitated with an internet modem, a generator and extra UMPC batteries to ensure uninterrupted power supply and internet connectivity while in the field.

Abstract

The 2005-2009/10 Uganda National Panel Survey (UNPS) is the first "panel" survey done by the LSMS group. The households included were chosen from the 2005-2006 Uganda National Household Survey (UNHS).The sample One of the primary uses of the UNPS is to inform policymaking in advance of the Budget, through descriptive reports that are made ready in time for the initial work on sector budget framework papers. The UNPS aimed at producing annual estimates of outcomes and output in the key policy areas and at providing a platform for the experimentation and assessment of national policies and programs. Explicitly, the objectives of the UNPS include:

1. To provide information required for monitoring the National Development Strategy, of major programs such as National Agricultural Advisory Services (NAADS) and General Budget Support, and also to provide information to the compilation of the National Accounts (e.g. agricultural production).
2. To provide high quality nationally representative information on income dynamics at the household level and provide annual information on service delivery and consumption expenditure estimates to monitor poverty and service outcomes in interim years of other national survey efforts, such as the Uganda National Household Survey (UNHS), Uganda Demographic and Health Survey (UDHS) and National Service Delivery Surveys (NSDS).
3. To provide a framework for low-cost experimentation with different policy interventions to e.g. reduce teacher absenteeism, improve ante- and post-natal care, or assessing the effect of agricultural input subsidies.
4. To provide a framework for policy-oriented analysis and capacity building substantiated with the UGDR and support to other research which will feed into the Annual Policy Implementation Review.
5. To facilitate randomized impact evaluations of interventions whose effects cannot currently be readily assessed through the existing system of national household surveys.

Geographic coverage

National

Analysis unit

Households

Kind of data

Sample survey data [ssd]

Sampling procedure

Survey Design The UNPS is carried out annually, over a twelve-month period on a nationally representative sample of households, for the purpose of accommodating the seasonality associated with the composition of and expenditures on consumption. The survey is conducted in two visits in order to better capture agricultural outcomes associated with the two cropping seasons of the country. The UNPS will therefore interview each household twice each year, in visits six months apart. Starting in 2005-09/10, the UNPS has been set out to track and re-interview 3,123 households that were distributed over 322 enumeration areas (EAs), selected out of the 783 EAs that had been visited by the Uganda National Household Survey (UNHS) in 2005-06. The UNPS EAs covered all 34 EAs visited by the UNHS 2005-06 in Kampala District, and 72 EAs (58 rural and 14 urban) in each of the (i) Central Region with the exception of Kampala District, (ii) Eastern Region, (iii) Western Region, and (iv) Northern Region.

Within each stratum, the UNPS EAs were selected from the UNHS 2005-06 EAs with equal probability, and with implicit stratification by urban/rural and district (in this order), except for the rural portions of the ten districts that were oversampled by the UNHS 2005-06. In these districts, the probabilities were deflated, to bring them back to the levels originally intended. Since IDP camps are now mostly unoccupied, the extra EAs in IDP camps are not a part of the UNPS subsample. This allocation strives for reasonably reliable estimates for the rural portion of each region, and for the set of urban areas out of Kampala as a whole, as well as the best possible estimates for Kampala that can be expected from a subsample of the UNHS 2005-06. Therefore, the UNPS strata of representativeness include (i) Kampala City, (ii) Other Urban Areas, (iii) Central Rural, (iv) Eastern Rural, (v) Western Rural, and (vi) Northern Rural.

Prior to the start of the 2005-09/10 field work, 2 UNPS households were also randomly selected in each EA for the purposes of tracking baseline individuals that moved away from original locations since the UNHS 2005-06. The initial UNPS sample was subject to three consecutive waves of data collection after which, parts of the sample were replaced by new households extracted from the updated sample frames developed by the UBOS as part of the 2012 Uganda Population and Housing Census. In addition, the UNPS will fit within the Long-Term Census and Household Survey Program and therefore both the questionnaires and the timing of data collection will be coordinated with the current surveys and census implemented by UBOS.

Note: Detailed description of sampling procedure and calculation of panel weights is presented in "Basic Information Document". The document is provided along with this metadata.

Mode of data collection

Face-to-face [f2f]

Cleaning operations

To ensure good quality of data, a system of double entry was used for data capture. A manual system of editing questionnaires was set-up and two office editors were recruited to further assess the consistency of the data collected. A computer program (hot-deck scrutiny) for verification and validation was developed and operated during data processing. Range and consistency checks were included in the data-entry program. More intensive and thorough checks were carried out using MS-ACCESS by the data processing team.

Sampling error estimates

The estimates presented in the reports of this study were derived from a scientifically selected sample and analysis of survey data was undertaken at national and regional levels. Standard Errors (SE) and Coefficients of Variations (CVs) of some of the variables have been presented in Appendix 2 to show the precision levels.

List of explanatory variables by importance in predicting housing rental values using bagging regression in Uganda.

Abstract

The Integrated Household Survey is one of the primary instruments implemented by the Government of Malawi through the National Statistical Office (NSO) roughly every 5 years to monitor and evaluate the changing conditions of Malawian households. The IHS data have, among other insights, provided benchmark poverty and vulnerability indicators to foster evidence-based policy formulation and monitor the progress of meeting the Millennium Development Goals (MDGs) as well as the goals listed as part of the Malawi Growth and Development Strategy (MGDS).

Geographic coverage

National

Analysis unit

• Households
• Individuals
• Children under 5 years
• Consumption expenditure commodities/items
• Communities
• Agricultural household/ Holder/ Crop

Universe

Members of the following households are not eligible for inclusion in the survey: • All people who live outside the selected EAs, whether in urban or rural areas. • All residents of dwellings other than private dwellings, such as prisons, hospitals and army barracks. • Members of the Malawian armed forces who reside within a military base. (If such individuals reside in private dwellings off the base, however, they should be included among the households eligible for random selection for the survey.) • Non-Malawian diplomats, diplomatic staff, and members of their households. (However, note that non-Malawian residents who are not diplomats or diplomatic staff and are resident in private dwellings are eligible for inclusion in the survey. The survey is not restricted to Malawian citizens alone.) • Non-Malawian tourists and others on vacation in Malawi.

Kind of data

Sample survey data [ssd]

Sampling procedure

The IHS3 sampling frame is based on the listing information and cartography from the 2008 Malawi Population and Housing Census (PHC); includes the three major regions of Malawi, namely North, Center and South; and is stratified into rural and urban strata. The urban strata include the four major urban areas: Lilongwe City, Blantyre City, Mzuzu City, and the Municipality of Zomba. All other areas are considered as rural areas, and each of the 27 districts were considered as a separate sub-stratum as part of the main rural stratum. It was decided to exclude the island district of Likoma from the IHS3 sampling frame, since it only represents about 0.1% of the population of Malawi, and the corresponding cost of enumeration would be relatively high. The sampling frame further excludes the population living in institutions, such as hospitals, prisons and military barracks. Hence, the IHS3 strata are composed of 31 districts in Malawi.

A stratified two-stage sample design was used for the IHS3.

Note: Detailed sample design information is presented in the "Third Integrated Household Survey 2010-2011, Basic Information Document" document.

Mode of data collection

Face-to-face [f2f]

Research instrument

The survey was collectd using four questionnaires: 1) Household Questionnaire 2) Agriculture Questionnaire 3) Fishery Questionnaire 4) Community Questionnaire

Cleaning operations

Data Entry Clerks Each IHS3 field team was assigned 1 data entry clerk to process completed questionnaires at the teams field based residence. Each data entry clerk was issued a laptop with the CSPro based data entry application, a printer to produce error reports on entered questionnaire, and flash disks for transferring files. The field based data entry clerk's primary responsibilities included: (1) receiving the completed questionnaires following the field supervisor's initial screening, (2) organizing and entering completed questionnaire in a timely manner, (3) generating and printing error reports for supervisor review, (4) modifying data after errors were resolved and authorized by the field supervisor, and (5) managing data files and local data back-ups. The data entry clerk was responsible for beginning initial data entry upon receipt of questionnaires from the field and generating error reports as quickly as possible after interviews were complete in the EA. When long distance travel to an enumeration area by the field team was required and the field team was required to spend multiple days away from their field residence the data entry clerk was required to travel with the team in order to maintain data processing schedules.

Field Based Data Entry and CAFE To better facilitate higher quality data and increase timely availability of data during the data capture process IHS3 utilized computer assisted field entry (CAFE). First data entry was conducted by field based data entry clerks immediately following completion of the team's daily field activities. Each team was equipped with 1 laptop computer for field based data entry using a CSPro-based application. The range and consistency checks built into the CSPro application was informed by the LSMS-ISA experience in Tanzania and Uganda, and the review of the IHS2 data. Prior programming of the data entry application allowed for a wide variety of range and consistency checks to be conducted and reported and potential issues investigated and corrected before closing the assigned enumeration area. Completed data was frequently relayed to the NSO central office in Zomba via email and tracked and processed upon receipt.

Double Data Entry Double data entry was implemented by a team of data entry clerks based at the NSO central office. Electronic data and questionnaires received from the field were cataloged by the Data Manager and electronic data loaded onto a central server to enable data entry verification on networked computers. To increase quality, the Data Entry Manager monitored the data verification staff and conducted quality assessments by randomly selecting processed questionnaires and comparing physical questionnaires to the result of double data entry. Data verification clerks were coached on inconsistencies when required.

Data Cleaning The data cleaning process was done in several stages over the course of field work and through preliminary analysis. The first stage of data cleaning was conducted in the field by the field based field teams utilizing error reports produced by the data entry applications. Field supervisors collected reports for each enumeration area and household and in coordination with the enumerators reviewed, investigated, and collected errors. Due to the quick turn-around in error reporting, it was possible to conduct call backs while the team was still operating in the enumeration area when required. Corrections to the data were entered by the field based data entry clerk before transmitting data to the NSO central office.

Upon receipt of the data from the field, module and cross module checks were performed using Stata to identify systematic issues and, where applicable, field teams were asked to investigate, revise and resend data for questionnaires still in their possession. Revised data files were cataloged and then replaced previous version of the data.

After data verification by the headquarters' double data entry team, data from the first data entry and second data entry were compared. Cases that revealed large inconsistencies between the first and second data entry, specifically large amounts of missing case level data in the second data entry relative to the first data entry were completely reentered. Further, variable specific inconsistency reports were generated and investigated and corrected by the double data entry team. Additional cleaning was performed after the double data entry team cleaning activities where appropriate to resolve systematic errors and organize data modules for consistency and efficient use. Case by case cleaning was also performed during the preliminary analysis specifically pertaining to out of range and outlier variables.

All cleaning activities were conducted in collaboration with the WB staff providing technical assistance to the NSO in the design and implementation of the IHS3.