EGM – Population & Housing Censuses Eurostat / UNECE - Geneva - 24/25 May 2012 Building the address register for the 2011 Census (England & Wales) Alistair Calder – Office for National Statistics
Role of addresses in Census 2011 Match Field check Evidence base & rules Summary Address Register for 2011
THE CENSUS FIELD OPERATION Post Out Special & CEs Post Back Internet Back CCS Follow Up Enforce -ment Questionnaire Tracking MATCH & check etc ADDRESS REGISTER
THE CENSUS FIELD OPERATION Post Out Post Back Follow Up Internet Back Special & CEs MATCH & check etc ADDRESS REGISTER Enforce -ment CCS Questionnaire Tracking Addresses
Why so important for 2011? Post-out to deliver 95% of questionnaires At the centre of questionnaire tracking –focus effort on non response –target follow-up resources more effectively Key source for use in QA of census estimates –Address frame feeds in to estimation
Address Register for 2011 Aims To create: best possible register of households from available sources and fieldwork a register that stakeholders were confident in a register that is best value for money > 99% complete <1% duplication A list designed specifically for Census
So what’s the problem ? just use the national address list !
Postcode Address File PAF Valuation Office National Land & Property Gazetteer NLPG Address Layer 2 AL2 LLPGs x 348 Council Tax TV License Utilities Emerg. Services etc
PAF & AL2 NLPG
PAF NLPG 95 % ? ? ?
PAF NLPG 95 % ? ? ? ? new CEs
Three main streams of work Match and validate Match of 3 national address files Royal Mail – PAF (& Ordnance Survey – AL2) Local Government Information House - NLPG (and VOA data) Unresolved address anomalies sent to suppliers and Local Authorities (LAs) ONS field checks 15% of E&W
Targeting the Address Check finding the suspect addresses
Where to check ??
Electricity - Multiple meters
Electoral register - Multiple names
Anomalies between sources (10%) - proxy for address complexity Multiple meters / Multiple names from electoral role within addresses (4%) – proxy for complex structures + Stratified random sample (1%) Criteria for checking
The Evidence Base and evidence codes
Evidence base ….. Evidence IARMFLA3rd? 9 High Street 1L1F1R1L1L1G 10 High Street0C1F1R0C1L1G 11 High Street0S1F0X1L1L0X 11a High Street0X1F0X0X0X0X 11b High Street0X1F0X0X1D0X Flat 1/11 High Street1L0X1R1L0X1G Flat 2/11 High Street1L0X1R1L0X1G
Evidence base ….. Evidence IARMFLA3rd? 9 High Street 1L1F1R1L1L1G 10 High Street0C1F1R0C1L1G 11 High Street0S1F0X1L1L0X 11a High Street0X1F0X0X0X0X 11b High Street0X1F0X0X1D0X Flat 1/11 High Street1L0X1R1L0X1G Flat 2/11 High Street1L0X1R1L0X1G
Evidence base ….. Evidence IARMFLA3rd? 9 High Street 1L1F1R1L1L1G 10 High Street0C1F1R0C1L1G 11 High Street0S1F0X1L1L0X 11a High Street0X1F0X0X0X0X 11b High Street0X1F0X0X1D0X Flat 1/11 High Street1L0X1R1L0X1G Flat 2/11 High Street1L0X1R1L0X1G
Evidence base ….. Evidence IARMFLA3rd? 9 High Street 1L1F1R1L1L1G 10 High Street0C1F1R0C1L1G 11 High Street0S1F0X1L1L0X 11a High Street0X1F0X0X0X0X 11b High Street0X1F0X0X1D0X Flat 1/11 High Street1L0X1R1L0X1G Flat 2/11 High Street1L0X1R1L0X1G
Evidence base ….. Evidence IARMFLA3rd? 9 High Street 1L1F1R1L1L1G 10 High Street0C1F1R0C1L1G 11 High Street0S1F0X1L1L0X 11a High Street0X1F0X0X0X0X 11b High Street0X1F0X0X1D0X Flat 1/11 High Street1L0X1R1L0X1G Flat 2/11 High Street1L0X1R1L0X1G
Keep the data up-to-date ! Constantly match in Change Pick up change late Research, Data Cleaning & Other Sources Clerical processes Desktop comparison, GIS, Internet, Aerial photography Other Sources Vacant houses, electoral register, utility meters Three main streams of work
Royal Mail vs. ONS Address Check Non-Residential vs. Residential Flats above shops, Complex addresses 221 Cannock Road, Cannock, WS11 5DD Recommendation: send a questionnaire
Royal Mail vs. ONS Address Check Non-Residential vs. Residential Flats above shops, Complex addresses 221 Cannock Road, Cannock, WS11 5DD Recommendation: send a questionnaire
Communal Establishments
Strong prioritisation Population Individual size of units (so local relevance) Target populations (hard to count) Split of types Complex – Universities, prisons (100% field) Standard – Hospitals, hostels etc (100% field) Simple – Hotels, B&Bs care homes (phone) Targeted field check Used local input / QA Communals
New approach to Census Building the address register Matching several datasets Targeted checking in field Use of evidence base and rules Communal establishments Dealt with separately Students done differently Result – considered highly successful Hit targets for over and under coverage Summary