High Frequency Phone Survey 2020-2021, Round 2
Papua New Guinea, 2020 - 2021
Get Microdata
Reference ID
PNG_2020_HFPS-W1_v02_EN_M_v02_A_OCS
Producer(s)
World Bank
Collections
Metadata
Created on
Aug 06, 2021
Last modified
Nov 08, 2022
Page views
7161
Downloads
48
Identification
Survey ID Number
PNG_2020_HFPS-W1_v02_EN_M_v02_A_OCS
Title
High Frequency Phone Survey 2020-2021, Round 2
Country
| Name | Country code |
|---|---|
| Papua New Guinea | PNG |
Study type
Other Household Survey [hh/oth]
Abstract
The objective of the second round of the World Bank high frequency mobile phone survey was to measure the continued socioeconomic impacts of COVID-19 in Papua New Guinea, including on livelihoods, food security, and public safety and security. The length of the survey was limited to 15 minutes and the survey instrument consisted of the following modules: Basic Information, Employment and Income Loss, Food Access and Food Security, Health, Public Trust and Security, and Assets and Wellbeing. The questions on employment and income were asked to the respondent and to the household head if different from the respondent. The recall period for current employment was in the previous week.
In addition, retrospective questions were asked for new respondents about the baseline (“the start of this year 2020”) as well about the situation at the time of round 1 in June (“June, around the time of the Queen’s birthday holiday”). The information from the new respondent could then be pooled with the returning respondents to have three consistent points in 2020.
For retrospective questions on employment, the baseline is defined as “the start of this year 2020” and new households were asked both about the baseline as well as the situation in early July, corresponding with the implementation of round 1 of data collection. Three subsequent rounds are planned, with the next in May 2021, though the implementation calendar may be revised to respond to changing conditions on the ground.
In addition, retrospective questions were asked for new respondents about the baseline (“the start of this year 2020”) as well about the situation at the time of round 1 in June (“June, around the time of the Queen’s birthday holiday”). The information from the new respondent could then be pooled with the returning respondents to have three consistent points in 2020.
For retrospective questions on employment, the baseline is defined as “the start of this year 2020” and new households were asked both about the baseline as well as the situation in early July, corresponding with the implementation of round 1 of data collection. Three subsequent rounds are planned, with the next in May 2021, though the implementation calendar may be revised to respond to changing conditions on the ground.
Kind of Data
Sample survey data [ssd]
Unit of Analysis
Households
Scope
Notes
The survey instrument for the second round consisted of the following modules:
-Basic Information,
-Employment and Income Loss,
-Food Access and Food Security,
-Health,
-Public Trust and Security,
-and Assets and Wellbeing.
The questionnaire for the UNICEF survey included sections on:
-Basic Information, Knowledge and Behavior,
-Service Delivery,
-Roster of Children Living in the Household (including schooling status),
-Access to Health,
-Education,
-Child Discipline,
-and Life Perspectives.
-Basic Information,
-Employment and Income Loss,
-Food Access and Food Security,
-Health,
-Public Trust and Security,
-and Assets and Wellbeing.
The questionnaire for the UNICEF survey included sections on:
-Basic Information, Knowledge and Behavior,
-Service Delivery,
-Roster of Children Living in the Household (including schooling status),
-Access to Health,
-Education,
-Child Discipline,
-and Life Perspectives.
Topics
| Topic |
|---|
| pacific-skills, education, training |
| pacific-skills, employment |
Keywords
| Keyword |
|---|
| COVID-19 |
| High Frequency Phone Survey |
| Education |
| Economic activity |
| Business |
| Income |
| Farming |
| Remittances |
| Food |
| Health |
| Public Services |
| Well-being |
Coverage
Geographic Coverage
National coverage
Universe
Over 18 years of age from the Digicel subscriber logs.
Producers and sponsors
Primary investigators
| Name | Affiliation |
|---|---|
| World Bank | World Bank Group |
Producers
| Name | Affiliation | Role |
|---|---|---|
| Development Data Group | World Bank Group | Technical Assistance |
| United Nations Children's Fund | United Nations | Technical Assistance |
| Korea Trust Fund and Peace-Building Fundation | Ministry of Strategy and Finance of Korea and World Bank | Technical Assistance |
| United States Agency for International Development | United States Government | Technical Assistance |
| World Health Organization | United Nations | Technical Assistance |
| Statistics for Development Division | Pacific Community | Technical Assistance |
| Digicuel Papua New Guinea | Advice on implementation |
Funding Agency/Sponsor
| Name | Abbreviation | Role |
|---|---|---|
| World Bank Group | WBG | Funded the survey and analysis |
Sampling
Sampling Procedure
The original objective of round 2 was to re-interview all households and respondents that were interviewed in round 1. There is high turnover of SIM cards in PNG, however, as numbers must be officially registered with a valid government ID within six months of activation or they are disconnected. Overall, of the original 3,115 households and 4,528 individuals interviewed during round 1, only 951 households and 962 individuals were re-interviewed in round 2. Though a small percentage of respondents refused (less than 1 percent), the main reason for failure to re-contact was that the number was no longer working.
In addition, there were 67 households in which someone answered at the original mobile number, but they were not a part of the original household. Therefore 1,804 additional households were added for the second round, for a total sample size of 2,820 households and 3,368 individuals in round 2.
To attempt to address some of the issues seen in round 1 in terms of the skew towards the higher deciles of the wealth distribution, a different targeting mechanism was used in round 2 based on subscriber characteristics derived from the Digicel database to try to address some of the skew towards richer households seen in the first round. To proxy poor households, the team targeted subscribers that did not send text messages on the assumption they were less likely to be literate. Similarly, subscribers that received only incoming calls or for whom the majority of credit was not purchased but transferred from other subscribers were thought to be more likely to be poor.
The UNICEF survey of households with children interviewed 2,449 of the 2,820 households interviewed in the second round of the World Bank survey, 86.8 percent of the total sample, and 96.6 percent of the total 2,534 that were targeted as having children under age 15. Using logit econometric model to compare the characteristics of eligible households which attritted between round 2 of the World Bank survey and the UNICEF survey, there are no statistically significant relationships accounting for the sex and education of the respondent, household wealth, and the geographic location (province, urban/ rural), with the exception of a statistically significant higher probability of attrition from those living in East Sepik Province.
In addition, there were 67 households in which someone answered at the original mobile number, but they were not a part of the original household. Therefore 1,804 additional households were added for the second round, for a total sample size of 2,820 households and 3,368 individuals in round 2.
To attempt to address some of the issues seen in round 1 in terms of the skew towards the higher deciles of the wealth distribution, a different targeting mechanism was used in round 2 based on subscriber characteristics derived from the Digicel database to try to address some of the skew towards richer households seen in the first round. To proxy poor households, the team targeted subscribers that did not send text messages on the assumption they were less likely to be literate. Similarly, subscribers that received only incoming calls or for whom the majority of credit was not purchased but transferred from other subscribers were thought to be more likely to be poor.
The UNICEF survey of households with children interviewed 2,449 of the 2,820 households interviewed in the second round of the World Bank survey, 86.8 percent of the total sample, and 96.6 percent of the total 2,534 that were targeted as having children under age 15. Using logit econometric model to compare the characteristics of eligible households which attritted between round 2 of the World Bank survey and the UNICEF survey, there are no statistically significant relationships accounting for the sex and education of the respondent, household wealth, and the geographic location (province, urban/ rural), with the exception of a statistically significant higher probability of attrition from those living in East Sepik Province.
Weighting
In round 1, there were two sets of primary weights calculated: those at the individual level and those at the household level. For the second round, there are four sets of primary weights: cross section weights at the household level, panel weights at the household level, cross section weights at the individual level, and panel weights at the individual level. Given the high levels of attrition, the panel weights were only used for selected analysis and the majority of the analysis relied on cross sectional weights.
The data collected from the mobile phone surveys differed substantially from the characteristics of the population as a whole, the data required reweighting. The process for reweighting was similar to that used in round 1 (see technical appendix to that report for full details).
For more information on weighting, please refer to the "Weighting" section (p.55) of the report provided in the External Resources.
The "weight" variable in the Household dataset is called "weight" whereas that in the Person dataset is called "indweight".
The data collected from the mobile phone surveys differed substantially from the characteristics of the population as a whole, the data required reweighting. The process for reweighting was similar to that used in round 1 (see technical appendix to that report for full details).
For more information on weighting, please refer to the "Weighting" section (p.55) of the report provided in the External Resources.
The "weight" variable in the Household dataset is called "weight" whereas that in the Person dataset is called "indweight".
Data Collection
Dates of Data Collection
| Start | End | Cycle |
|---|---|---|
| 2020-12-09 | 2020-12-31 | Data collection (World Bank survey) |
| 2021-01-06 | 2021-01-24 | Data collection (UNICEF survey) |
Data Collection Mode
Computer Assisted Telephone Interview [cati]
Access policy
Confidentiality
See https://microdata.worldbank.org/index.php/terms-of-use
Access conditions
See https://microdata.worldbank.org/index.php/terms-of-use
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_PNG_2020_HFPS-W2_v02_EN_M_v02_A_OCS_FAO
Producers
| Name | Abbreviation | Affiliation | Role |
|---|---|---|---|
| World Bank | WBG | World Bank Group | Metadata producer |
| Office of Chief Statistican | OCS | Food and Agriculture Organisation | Metadata adapted for FAM |
DDI Document version
PNG_2020_HFPS-W2_v02_EN_M_v02_A_OCS_v02