COVID-19 High Frequency Phone Survey of Households 2020
Ethiopia, 2020
World Bank
Created on
September 15, 2020
Last modified
September 15, 2020
Page views
6135
Download
933
Documentation in PDF
Metadata
DDI/XML
JSON
Identification
Survey ID Number
ETH_2020_HFPS_v03_M
Title
COVID-19 High Frequency Phone Survey of Households 2020
Country
| Name | Country code |
|---|---|
| Ethiopia | ETH |
Study type
Socio-Economic/Monitoring Survey [hh/sems]
Series Information
The World Bank is providing support to countries to help mitigate the spread and impact of the new coronavirus disease (COVID-19). One area of support is for data collection to inform evidence-based policies that may help mitigate the effects of this disease. Towards this end, the World Bank is leveraging the Living Standards Measurement Study - Integrated Survey on Agriculture (LSMS-ISA) program to implement high-frequency phone surveys on COVID-19 in 5 African countries - Nigeria, Ethiopia, Uganda, Tanzania, and Malawi. This effort is part of a broader first wave of World Bank-supported national longitudinal high frequency survey that can be used to help assess the economic and social implications of the COVID-19 pandemic on households and individuals.
Abstract
The potential impacts of the COVID-19 pandemic in Ethiopia are expected to be severe on Ethiopian households' welfare. To monitor these impacts on households, the team selected a subsample of households that had been interviewed for the Living Standards Measurement Study (LSMS) in 2019, covering urban and rural areas in all regions of Ethiopia. The 15-minute questionnaire covers a series of topics, such as knowledge of COVID and mitigation measures, access to routine healthcare as public health systems are increasingly under stress, access to educational activities during school closures, employment dynamics, household income and livelihood, income loss and coping strategies, and external assistance.
The survey is implemented using Computer Assisted Telephone Interviewing, using a modular approach, which allows for modules to be dropped and/or added in different waves of the survey. Survey data collection started at the end of April 2020 and households are called back every three to four weeks for a total of seven survey rounds to track the impact of the pandemic as it unfolds and inform government action. This provides data to the government and development partners in near real-time, supporting an evidence-based response to the crisis.
The sample of households was drawn from the sample of households interviewed in the 2018/2019 round of the Ethiopia Socioeconomic Survey (ESS). The extensive information collected in the ESS, less than one year prior to the pandemic, provides a rich set of background information on the COVID-19 High Frequency Phone Survey of households which can be leveraged to assess the differential impacts of the pandemic in the country.
The survey is implemented using Computer Assisted Telephone Interviewing, using a modular approach, which allows for modules to be dropped and/or added in different waves of the survey. Survey data collection started at the end of April 2020 and households are called back every three to four weeks for a total of seven survey rounds to track the impact of the pandemic as it unfolds and inform government action. This provides data to the government and development partners in near real-time, supporting an evidence-based response to the crisis.
The sample of households was drawn from the sample of households interviewed in the 2018/2019 round of the Ethiopia Socioeconomic Survey (ESS). The extensive information collected in the ESS, less than one year prior to the pandemic, provides a rich set of background information on the COVID-19 High Frequency Phone Survey of households which can be leveraged to assess the differential impacts of the pandemic in the country.
Kind of Data
Sample survey data [ssd]
Unit of Analysis
- Households
- Individuals
- Individuals
Version
Version Description
Version 03: Edited, anonymized dataset for public distribution
Scope
Notes
The Ethiopia - COVID-19 High Frequency Phone Survey of Households covered the following topics:
- Household Roster (Rounds 1, 2, 3)
- Knowledge Regarding the Spread of COVID-19 (Round 1)
- Behaviour and Social Distancing (Rounds 1 and 3)
- Access to Basic Services (Rounds 1, 2, 3)
- Employment (Rounds 1, 2, 3)
- Income Loss and Coping (Rounds 1, 2, 3)
- Food Security (Rounds 1, 2, 3)
- Aid and Support/ Social Safety Nets (Rounds 1, 2, and 3)
- Agriculture (Round 3)
- Household Roster (Rounds 1, 2, 3)
- Knowledge Regarding the Spread of COVID-19 (Round 1)
- Behaviour and Social Distancing (Rounds 1 and 3)
- Access to Basic Services (Rounds 1, 2, 3)
- Employment (Rounds 1, 2, 3)
- Income Loss and Coping (Rounds 1, 2, 3)
- Food Security (Rounds 1, 2, 3)
- Aid and Support/ Social Safety Nets (Rounds 1, 2, and 3)
- Agriculture (Round 3)
Coverage
Geographic Coverage
National coverage - rural and urban
Universe
The survey covered all de jure households excluding prisons, hospitals, military barracks, and school dormitories.
Producers and sponsors
Primary investigators
| Name |
|---|
| World Bank |
Producers
| Name | Role |
|---|---|
| Central Statistical Agency | Collaborator |
Funding Agency/Sponsor
| Name | Abbreviation | Role |
|---|---|---|
| United States Agency for International Development | USAID | Funded the study |
| The World Bank Group | WB | Funded the study |
Sampling
Sampling Procedure
The sample of the HFPS-HH is a subsample of the 2018/19 Ethiopia Socioeconomic Survey (ESS). The ESS is built on a nationally and regionally representative sample of households in Ethiopia. ESS 2018/19 interviewed 6,770 households in urban and rural areas. In the ESS interview, households were asked to provide phone numbers either their own or that of a reference household (i.e. friends or neighbors) so that they can be contacted in the follow-up ESS surveys should they move from their sampled location. At least one valid phone number was obtained for 5,374 households (4,626 owning a phone and 995 with a reference phone number). These households established the sampling frame for the HFPS-HH.
To obtain representative strata at the national, urban, and rural level, the target sample size for the HFPS-HH is 3,300 households; 1,300 in rural and 2,000 households in urban areas. In rural areas, we attempt to call all phone numbers included in the ESS as only 1,413 households owned phones and another 771 households provided reference phone numbers. In urban areas, 3,213 households owned a phone and 224 households provided reference phone numbers. To account for non-response and attrition all the 5,374 households were called in round 1 of the HFPS-HH.
The total number of completed interviews in round one is 3,249 households (978 in rural areas, 2,271 in urban areas).
The total number of completed interviews in round two is 3,107 households (940 in rural areas, 2,167 in urban areas).
The total number of completed interviews in round three is 3,058 households (934 in rural areas, 2,124 in urban areas).
To obtain representative strata at the national, urban, and rural level, the target sample size for the HFPS-HH is 3,300 households; 1,300 in rural and 2,000 households in urban areas. In rural areas, we attempt to call all phone numbers included in the ESS as only 1,413 households owned phones and another 771 households provided reference phone numbers. In urban areas, 3,213 households owned a phone and 224 households provided reference phone numbers. To account for non-response and attrition all the 5,374 households were called in round 1 of the HFPS-HH.
The total number of completed interviews in round one is 3,249 households (978 in rural areas, 2,271 in urban areas).
The total number of completed interviews in round two is 3,107 households (940 in rural areas, 2,167 in urban areas).
The total number of completed interviews in round three is 3,058 households (934 in rural areas, 2,124 in urban areas).
Weighting
To obtain unbiased estimates from the sample, the information reported by households needs to be adjusted by a sampling weight (or raising factor) w_h. To construct the sampling weights, we follow the steps outlined in Himelein, K. (2014), which outlines eight steps, of which we follow six, to construct the sampling weights for the HFPS-HH:
1. Begin with base weights from the Ethiopia Socioeconomic Survey ESS 2018/19 for each household
2. Incorporate probability of sub-selection of round 1 unit for each of the phone survey households. We calculate the probability of selection for each of the 20 strata in the ESS (urban and rural in each of the 11 regions except for Addis Ababa where we only have an urban stratum) by creating the numerators as the number of completed phone interviews and the denominator as the number of households in the ESS for each stratum.
3. Pool the weights in Steps 1 and 2.
4. Derive attrition-adjusted weights for all individuals by running a logistic response propensity model based on characteristics of the household head (i.e. education, labor force status, demographic characteristics), characteristics of the household (consumption, assets, financial characteristics), and characteristics of the dwelling (house ownership, overcrowding).
5. Trim weights by replacing the top two percent of observations with the 98th percentile cut-off point; and
6. Post-stratify weights to known population totals to correct for the imbalances across our urban and rural sample. In doing so, we ensure that the distribution in the survey matches the distribution in the ESS.
* Additional technical details and explanations on each of the steps briefly outlined above can be found in Himelein, K. (2014).
1. Begin with base weights from the Ethiopia Socioeconomic Survey ESS 2018/19 for each household
2. Incorporate probability of sub-selection of round 1 unit for each of the phone survey households. We calculate the probability of selection for each of the 20 strata in the ESS (urban and rural in each of the 11 regions except for Addis Ababa where we only have an urban stratum) by creating the numerators as the number of completed phone interviews and the denominator as the number of households in the ESS for each stratum.
3. Pool the weights in Steps 1 and 2.
4. Derive attrition-adjusted weights for all individuals by running a logistic response propensity model based on characteristics of the household head (i.e. education, labor force status, demographic characteristics), characteristics of the household (consumption, assets, financial characteristics), and characteristics of the dwelling (house ownership, overcrowding).
5. Trim weights by replacing the top two percent of observations with the 98th percentile cut-off point; and
6. Post-stratify weights to known population totals to correct for the imbalances across our urban and rural sample. In doing so, we ensure that the distribution in the survey matches the distribution in the ESS.
* Additional technical details and explanations on each of the steps briefly outlined above can be found in Himelein, K. (2014).
Data Collection
Dates of Data Collection
| Start | End | Cycle |
|---|---|---|
| 2020-04-22 | 2020-05-13 | Round 1 |
| 2020-05-14 | 2020-06-03 | Round 2 |
| 2020-06-04 | 2020-06-26 | Round 3 |
Data Collection Mode
Computer Assisted Telephone Interview [cati]
Data Collection Notes
The Ethiopia- COVID-19 High Frequency Phone Survey of Households (HFPS) was conducted using Computer Assisted Telephone Interview (CATI) techniques. The household questionnaire was implemented using the CATI software, SurveyCTO. Each enumerator was given a tablet which they used to implement the interviews, along with data bundles to be used on their own mobile phone devices.
DATA COMMUNICATION SYSTEM: SurveyCTO's built-in data monitoring functions are used. Each enumerator was provided with a data bundle, allowing for internet connectivity and daily synchronization of their tablet. Data was sent to the server daily. Senior Field Supervisors served as the first step in ensuring data quality. Senior Field Supervisors reviewed the survey with enumerators twice daily via one-on-one calls and were always available to address any concerns that arose while performing an interview. At the same time, a Research Analyst was in charge of checking the uploaded data daily to correct errors and work to prevent them in future surveys. The following data quality checks were completed:
• Daily SurveyCTO monitoring: This included outlier checks, skipped questions, a review of “Other, specify”, other text responses, and enumerator comments. Enumerator comments were used to suggest new response options or to highlight situations where existing options should be used instead. Monitoring also included a review of variable relationship logic checks and checks of the logic of answers. Finally, outliers in phone variables such as survey duration or the percentage of time audio was at a conversational level were monitored. A survey duration of close to 15 minutes and a conversation-level audio percentage of around 40% was considered normal.
• Dashboard review: This included monitoring individual enumerator performance, such as the number of calls logged, duration of calls, percentage of calls responded to and percentage of non-consents. Non-consent reason rates and attempts per household were monitored as well. Duration analysis using R was used to monitor each module's duration and estimate the time required for subsequent rounds. The dashboard was also used to track overall survey completion and preview the results of key questions.
• Daily Data Team reporting: The Field Supervisors and the Data Manager reported daily feedback on call progress, enumerator feedback on the survey, and any suggestions to improve the instrument, such as adding options to multiple choice questions or adjusting translations.
• Audio audits: Audio recordings were captured during the consent portion of the interview for all completed interviews, for the enumerators' side of the conversation only. The recordings were reviewed for any surveys flagged by enumerators as having data quality concerns and for an additional random sample of 2% of respondents. A range of lengths were selected to observe edge cases. Most consent readings took around one minute, with some longer recordings due to questions on the survey or holding for the respondent. All reviewed audio recordings were completed satisfactorily.
• Back-check survey: Field Supervisors made back-check calls to a random sample of 5% of the households that completed a survey in Round 1. Field Supervisors called these households and administered a short survey, including (i) identifying the same respondent; (ii) determining the respondent's position within the household; (iii) confirming that a member of the the data collection team had completed the interview; and (iv) a few questions from the original survey.
DATA COMMUNICATION SYSTEM: SurveyCTO's built-in data monitoring functions are used. Each enumerator was provided with a data bundle, allowing for internet connectivity and daily synchronization of their tablet. Data was sent to the server daily. Senior Field Supervisors served as the first step in ensuring data quality. Senior Field Supervisors reviewed the survey with enumerators twice daily via one-on-one calls and were always available to address any concerns that arose while performing an interview. At the same time, a Research Analyst was in charge of checking the uploaded data daily to correct errors and work to prevent them in future surveys. The following data quality checks were completed:
• Daily SurveyCTO monitoring: This included outlier checks, skipped questions, a review of “Other, specify”, other text responses, and enumerator comments. Enumerator comments were used to suggest new response options or to highlight situations where existing options should be used instead. Monitoring also included a review of variable relationship logic checks and checks of the logic of answers. Finally, outliers in phone variables such as survey duration or the percentage of time audio was at a conversational level were monitored. A survey duration of close to 15 minutes and a conversation-level audio percentage of around 40% was considered normal.
• Dashboard review: This included monitoring individual enumerator performance, such as the number of calls logged, duration of calls, percentage of calls responded to and percentage of non-consents. Non-consent reason rates and attempts per household were monitored as well. Duration analysis using R was used to monitor each module's duration and estimate the time required for subsequent rounds. The dashboard was also used to track overall survey completion and preview the results of key questions.
• Daily Data Team reporting: The Field Supervisors and the Data Manager reported daily feedback on call progress, enumerator feedback on the survey, and any suggestions to improve the instrument, such as adding options to multiple choice questions or adjusting translations.
• Audio audits: Audio recordings were captured during the consent portion of the interview for all completed interviews, for the enumerators' side of the conversation only. The recordings were reviewed for any surveys flagged by enumerators as having data quality concerns and for an additional random sample of 2% of respondents. A range of lengths were selected to observe edge cases. Most consent readings took around one minute, with some longer recordings due to questions on the survey or holding for the respondent. All reviewed audio recordings were completed satisfactorily.
• Back-check survey: Field Supervisors made back-check calls to a random sample of 5% of the households that completed a survey in Round 1. Field Supervisors called these households and administered a short survey, including (i) identifying the same respondent; (ii) determining the respondent's position within the household; (iii) confirming that a member of the the data collection team had completed the interview; and (iv) a few questions from the original survey.
Data Collectors
| Name |
|---|
| Laterite BV |
Questionnaires
Questionnaires
The survey questionnaires (Round 1, Round 2, and Round 3) were administered to all the households in the sample. The questionnaires consisted of the following sections:
Baseline (Round 1)
- Household Identification
- Interview Information
- Household Roster
- Knowledge Regarding the Spread of Coronavirus
- Behavior and Social Distancing
- Access to Basic Services
- Employment
- Income Loss and Coping
- Food Security
- Aid and Support/ Social Safety Nets
Round 2
- Household Identification
- Household Roster
- Access to Basic Services
- Employment
- Income Loss and Coping
- Food Security
- Aid and Support/ Social Safety Nets
Round 3
- Household Identification
- Household Roster
- Behavior and social distancing
- Access to Basic Services
- Employment
- Income Loss and Coping
- Food Security
- Agriculture
- Aid and Support/ Social Safety Nets
Baseline (Round 1)
- Household Identification
- Interview Information
- Household Roster
- Knowledge Regarding the Spread of Coronavirus
- Behavior and Social Distancing
- Access to Basic Services
- Employment
- Income Loss and Coping
- Food Security
- Aid and Support/ Social Safety Nets
Round 2
- Household Identification
- Household Roster
- Access to Basic Services
- Employment
- Income Loss and Coping
- Food Security
- Aid and Support/ Social Safety Nets
Round 3
- Household Identification
- Household Roster
- Behavior and social distancing
- Access to Basic Services
- Employment
- Income Loss and Coping
- Food Security
- Agriculture
- Aid and Support/ Social Safety Nets
Data Processing
Data Editing
DATA CLEANING
At the end of data collection, the raw dataset was cleaned by the Research team. This included formatting, and correcting results based on monitoring issues, enumerator feedback and survey changes. The details are as follows.
Variable naming and labeling:
• Variable names were changed to reflect the lowercase question name in the paper survey copy, and a word or two related to the question.
• Variables were labeled with longer descriptions of their contents and the full question text was stored in Notes for each variable.
• “Other, specify” variables were named similarly to their related question, with “_other” appended to the name.
• Value labels were assigned where relevant, with options shown in English for all variables, unless preloaded from the roster in Amharic.
Variable formatting:
• Variables were formatted as their object type (string, integer, decimal, time, date, or datetime).
• Multi-select variables were saved both in space-separated single-variables and as multiple binary variables showing the yes/no value of each possible response.
• Time and date variables were stored as POSIX timestamp values and formatted to show Gregorian dates.
• Location information was left in separate ID and Name variables, following the format of the incoming roster. IDs were formatted to include only the variable level digits, and not the higher-level prefixes (2-3 digits only.)
• Full Household and Enumeration Area ID variables were given leading 0s to match incoming roster format.
Observation and variable arrangement:
• Only consented surveys were kept in the dataset, and all personal information and internal survey variables were dropped from the clean dataset.
• Roster data is separated from the main data set and kept in long-form but can be merged on the key variable (key can also be used to merge with the raw data).
• In the main dataset, ii4_resp_id and cs7_hhh_id are the roster IDs of the respondent and household head respectively, and can be merged with individual_id in the roster.
• The variables were arranged in the same order as the paper instrument, with observations arranged according to their submission time.
Backcheck data review: Results of the backcheck survey are compared against the originally captured survey results using the bcstats command in Stata. This function delivers a comparison of variables and identifies any discrepancies. Any discrepancies identified are then examined individually to determine if they are within reason.
At the end of data collection, the raw dataset was cleaned by the Research team. This included formatting, and correcting results based on monitoring issues, enumerator feedback and survey changes. The details are as follows.
Variable naming and labeling:
• Variable names were changed to reflect the lowercase question name in the paper survey copy, and a word or two related to the question.
• Variables were labeled with longer descriptions of their contents and the full question text was stored in Notes for each variable.
• “Other, specify” variables were named similarly to their related question, with “_other” appended to the name.
• Value labels were assigned where relevant, with options shown in English for all variables, unless preloaded from the roster in Amharic.
Variable formatting:
• Variables were formatted as their object type (string, integer, decimal, time, date, or datetime).
• Multi-select variables were saved both in space-separated single-variables and as multiple binary variables showing the yes/no value of each possible response.
• Time and date variables were stored as POSIX timestamp values and formatted to show Gregorian dates.
• Location information was left in separate ID and Name variables, following the format of the incoming roster. IDs were formatted to include only the variable level digits, and not the higher-level prefixes (2-3 digits only.)
• Full Household and Enumeration Area ID variables were given leading 0s to match incoming roster format.
Observation and variable arrangement:
• Only consented surveys were kept in the dataset, and all personal information and internal survey variables were dropped from the clean dataset.
• Roster data is separated from the main data set and kept in long-form but can be merged on the key variable (key can also be used to merge with the raw data).
• In the main dataset, ii4_resp_id and cs7_hhh_id are the roster IDs of the respondent and household head respectively, and can be merged with individual_id in the roster.
• The variables were arranged in the same order as the paper instrument, with observations arranged according to their submission time.
Backcheck data review: Results of the backcheck survey are compared against the originally captured survey results using the bcstats command in Stata. This function delivers a comparison of variables and identifies any discrepancies. Any discrepancies identified are then examined individually to determine if they are within reason.
Access policy
Contacts
| Name | Affiliation | |
|---|---|---|
| LSMS | The World Bank Group | lsms@worldbank.org |
Citation requirements
Use of the dataset must be acknowledged using a citation which would include:
- the Identification of the Primary Investigator
- the title of the survey (including country, acronym and year of implementation)
- the survey reference number
- the source and date of download
World Bank. Ethiopia - COVID-19 High Frequency Phone Survey of Households 2020. Dataset downloaded from www.microdata.worldbank.org on [date].
- the Identification of the Primary Investigator
- the title of the survey (including country, acronym and year of implementation)
- the survey reference number
- the source and date of download
World Bank. Ethiopia - COVID-19 High Frequency Phone Survey of Households 2020. Dataset downloaded from www.microdata.worldbank.org on [date].
Disclaimer and copyrights
Disclaimer
The user of the data acknowledges that the original collector of the data, the authorized distributor of the data, and the relevant funding agency bear no responsibility for use of the data or for interpretations or inferences based upon such uses.
Metadata production
DDI Document ID
DDI_ETH_2020_HFPS_v03_M
Producers
| Name | Abbreviation | Affiliation | Role |
|---|---|---|---|
| Development Economics Data Group | DECDG | The World Bank | Documentation of the DDI |
Date of Metadata Production
2020-09-14
DDI Document version
Version 03 (September 2020). This metadata contains Baseline (Round 1), Round 2, and Round 3 survey information.