UN Banner
HomeAboutExperiences and LessonsCase StudiesRecommendationsResourcesChecklist
Home
About▶
BackgroundPurposeRationaleStructureTarget Audience
Experiences and Lessons▶
1. Project Planning and Management2. Geospatial Mapping and EA Database3. Enumeration Instruments and Tools4. Data Capture and Management5. Census Testing and Pilots6. Recruitment and Training7. Deployment and Supervision8. Enumeration and Logistics9. Quality Assurance and Risk Management10. Census Analysis and Products11. Partnerships and Collaboration12. Advocacy and Publicity13. Procurement and Financial Management14. Post-Enumeration Survey15. Alternative Approaches16. Planning Checklist
Case Studies▶
1. Project Planning and Management2. Geospatial Mapping3. Enumeration Instruments4. Data Capture5. Census Testing6. Recruitment and Training7. Deployment and Supervision8. Enumeration and Logistics9. Quality Assurance10. Census Analysis11. Partnerships12. Advocacy13. Procurement14. Post-Enumeration Survey15. Alternative Approaches16. Planning Checklist
Recommendations
Resources
Checklist
Skip to main content

Recommendations​

  • The PES should be planned and budgeted for as an integral part of the census operations. A PES is unlikely to yield useful results if it is a last-minute addition only considered once the census operation has been completed. As such, unless the PES is properly planned, funded and executed, it is unlikely to yield value for money as the results may not be valid. Consider whether or not carrying out a PES is a good use of resource. Furthermore, the budget for the PES should be independent from the principal budget and should be ring-fenced so that the PES is adequately funded and an independent PES team, separate from the main census team can be properly staffed.

  • If there is no administrative data available in the country, then the importance of the PES is magnified since there is no alternative to quality assure the coverage and content of the census. Therefore, countries with poor administrative data should attempt to carry out a PES as part of their census operation. However, bear recommendation 1 in mind. If there is insufficient resource to carry out the PES, then concentrate instead on improving census enumeration.

  • The PES enumeration should be carried out within a few months of completing the census enumeration (UN P&R, Rev 4, para. 4.123). Ideally the PES enumeration should start as soon as possible after census enumeration has completed. However, PES enumeration should not start until the census enumeration is complete in all areas in order to maintain independence.

  • A different team should be responsible for running and management of the PES, not made up from people who are responsible for running and management of the census.

  • The necessity to match the PES data to the census data with high accuracy should be borne in mind when designing the PES questionnaire, the census and PES CAPI applications, and the census and PES enumerator training, so that everything is done to make automatic matching as successful (fast and accurate) as possible.

  • Harmonise the PES CAPI with the census CAPI for the questions that are asked in both surveys to gather the variables that will be used for matching (e.g., first name, last name, date of birth, relationship to head of household, marital status).

  • When training census and PES enumerators, ensure they understand how the data will be used for matching and the importance of spelling names correctly. Give clear instructions on what to do when information such as a name or date of birth is unknown. This will prevent enumerators using their own versions of ‘unknown’ which all have to be cleaned from the data prior to matching.

  • Collect geolocation codes for each household enumerated on both the census and PES. This will assist with automatic household matching.

  • Add verification steps into the CAPI to minimise the risk of incorrect data being collected which makes the automatic matching harder. For example, if date of birth is collected, the CAPI can automatically calculate the age and require the enumerator to confirm that this is correct. If the relationship to head of household is impossible e.g., son older than their father, the CAPI should flag this and require the enumerator to confirm the correct relationship.

  • Ensure field force are trained to correct errors in enumeration by editing the record on the CAPI rather than adding a new record and thus causing duplicates.

  • Identify in advance the key variables for estimation – at a minimum these will be age and sex, and programme the CAPI so that these are compulsory questions.

  • Carrying out data processing steps such as Resolving Multiple Responses (RMR) and Removing False Persons (RFP) before attempting to match the census and PES data will improve the quality of the data and the matching.

    • RMR means finding and then marking as out-of-scope, any duplicate records. These may appear as within household duplicates, outside household duplicates, or entire duplicate households.
    • RFP means finding any records where there is not sufficient evidence of this record representing a true person e.g., name, age/date of birth and sex are all missing. This can be caused by CAPI error, data processing or transmission error, or enumerator error.

  • Ensure that the codes used to identify EAs are used consistently in census and PES data so that the correct EAs can be selected from the census data for matching with the PES records.

  • Ensure that household data collected at listing can easily be linked with person data collected during enumeration. There should be a unique identification number that allows for a simple one-to-one join.

  • If it is intended to carry out a PES, then a pilot PES should also be conducted soon after the census pilot (see Chapter 5: Census Testing and Pilots. This will enable any problems with the PES data collection, PES questionnaire, data processing etc. to be resolved before the final PES enumeration. A PES pilot will also provide data that can be matched to the census pilot data thereby allowing the automatic matching algorithms to be coded, tuned and tested for accuracy prior to the live census to PES matching operation. The clerical matching system can also be tested during this time and instructions for clerical matchers can be developed and tested. It is therefore important that the pilot PES is carried out in a sample of EAs used for the pilot census so that the same persons are enumerated in both.

  • The post-enumeration survey itself needs to be widely publicised so that respondents understand the purpose and do not mistakenly think that this is a duplicate census enumeration. PES publicity should start only after census enumeration is completed and may be focused on the parishes that contain PES EAs.

  • Reconciliation visits may not make a significant difference to the results of the PES analysis. Therefore, countries should consider the possibility of only carrying out reconciliation in a sample of PES EAs only – in particular those EAs where there is an unexplained low match rate i.e., a large number of unmatched census and/or PES persons, and should have a plan to use the results of the reconciliation.

  • A description of the PES methodology and the results of the PES analysis should be published so that users of the census data understand the limitations in the census results and can adjust accordingly.