Strand organisers: Dr. Julie Jefferies, Office for National Statistics; Professor Ludi Simpson, University of Manchester
Assuring the quality of the 2011 Census population estimates
Paula Guy, Jonathan Wroth-Smith, Luke Aki, Office for National Statistics
The first release of 2011 Census estimates will be published in July 2012. This paper will present an overview of the process that was followed including the key pieces of demographic analysis which were carried out to give confidence in the robustness of the estimates. It will draw on case study examples that demonstrate the work carried out and describe some of the adjustments which were made as a result of what was found.
Email: Luke Aki: firstname.lastname@example.org
Finalising the Census after dual system estimations
Roma Chappel, Office for National Statistics
Census estimates are produced "bottom-up" based on a comprehensive count together with a coverage survey and estimation methods that allow adjustments to be made for those people who inevitably are missed. This presentation shows what was done to quality assure the cumulative picture that emerges when the individual LAs are aggregated. It explains what was done following the quality assurance of the results in this more “top-down” way, using the national adjustment that was part of the methods reviewed and published before the census estimates were released.
Email: Roma Chappell: email@example.com
Delivering the 2011 Census estimates: reflections on quality assurance
David Martin, University of Southampton
Quality assurance of the 2011 estimates has been an integral part of census processing by the Office for National Statistics. QA of a census presents several unique challenges: although it is possible to review and audit the design and operation of procedures and to assess initial values against alternative sources, the true counts for specific groups and areas which the census is attempting to estimate remain the greatest unknown (and ultimately unknowable) elements. The importance of census QA is highlighted by the complex revisions required to 2001 estimates - both the adjustments prior to data publication and the subsequent revision of populations for 15 local authorities and the experience of validating the census potentially has much to inform the statistical community about intercensal population estimation and potential future replacements for a census. ONS invited four external members to participate in its High Level Quality Assurance panels which have met throughout the spring of 2012 and this BSPS paper complements those by ONS staff, presenting an independent view of the QA process and the lessons to be learned.
Email: Professor David Martin: firstname.lastname@example.org
2011 Census: First release, subsequent releases and specialist products
Denise McGregor, Emma White, Office for National Statistics
The first results from the 2011 Census are scheduled for release in July 2012. These include population estimates by age and sex down to local authority level for England and Wales, as well as household estimates. This presentation will highlight some key results and trends. The presentation will also look ahead to the three further releases of census information on the usual resident population, with details of the planned releases of univariate and cross-tabulated data.
Short-term migration estimates from the 2011 Census
Marie-Lisa Burrough, Office for National Statistics
Since the 2011 Census the Office for National Statistics has recognised the increasing significance of short-term migration moves, where individuals move to or from the UK for between 3 and 12 months. Such moves were identified as a particular feature of increased migration flows to the UK following European Union expansion in 2004. Having published estimates of short-term migration moves as part of the Improving Migration and Population Statistics Programme, the 2011 Census included questions to identify stock of such migrants. This paper will present an overview of these census estimates and will consider implications for future statistics.
Email: Marie-Lisa Burrough: email@example.com
Plans and proposals for analysis of the 2011 Census
Jon Gough, Craig Taylor, Office for National Statistics
The ONS are conducting a Census Analysis work programme to provide users with a planned and coordinated work plan that will lead to the production of timely analysis that meets user needs. This requires a coordinated approach across ONS, other Governments Departments, Local Government and external researchers & academics. This work programme ultimately aims to deliver a series of analytical products to be produced by ONS in a timely and coordinated manner, link and coordinate with analysis produced externally to ONS and develop and maintain an Analytical Centre for the Census. This presentation will highlight the plans and proposals for the ONS Census analysis work programme.
Email: Jon Gough: firstname.lastname@example.org
Methods for revising population estimates for mid-2002 to mod-2010 in line with the 2011 census results
Emily Shrosbree, Office for National Statistics
ONS produces mid-year population estimates for England & Wales at national and local authority level on an annual basis using the cohort component method. The estimates for mid-2002 to mid- 2010 have been produced using the 2001 Census as the population base, ageing the population on and taking account of population change due to births, deaths and migration. Components of change are estimated annually from the best available administrative and survey sources. The census provides the most complete source of information about the population every ten years, and the accuracy of the inter-censal population estimates is determined by both the quality of the available data sources and the methods used to estimate population change. Any inaccuracy accumulates as the population estimates become further from the census base. The mid-2011 population estimates will be rebased using the 2011 Census estimates, and the back-series from mid-2002 to mid-2010 will be revised to ensure a consistent time-series, addressing any drift in the inter-censal estimates. This paper focuses on the methods used to create such a time-series. Firstly, the methodological approach chosen by ONS to create a consistent back-series is outlined. Secondly, methods for determining the size and cause of the inter-censal drift including identifying how any drift has accumulated over time are detailed. The methods for implementing the complete approach to create the back-series for publication are also discussed. Finally, the paper outlines the timetable for publishing the revised population estimates.
Email: Emily Shrosbree: email@example.com
Matching of administrative data to validate the 2011 census in England and Wales
Andrew Charlesworth, Nicola Rogers, Louisa Blackwell, David Fallows, Office for National Statistics
This presentation describes the role of administrative data matching in the quality assurance process for the 2011 Census in England and Wales. A range of administrative sources were matched both to the Census and to each other with the aim of assessing relative coverage. The sources will be described. The innovative data matching methods and systems designed for this task will be presented. These systems allowed a flexible, reactive approach to the application of analytic methods that were appropriate to the research questions that arose during the census quality assurance process. On their own and in combination, the administrative sources provided new insights, both empirical and theoretical, into population enumeration. We will present some examples of these findings.
Email: Andrew Charlesworth: firstname.lastname@example.org
On the first results of the last 2011 Population Census of Greece
Cleon Tsimbos, University of Piraeus
In this work some demographic issues are raised concerning possible enumeration deficiencies of the 2011 population census of Greece. The analysis is based on the provisional census figures released in July 2011 by the Hellenic Statistical Authority (ELSTAT). The application of demographic techniques reveals a possible under-enumeration of the country\'s population raging from 90 to 441 thousand persons (1% - 4%) depending on the assumptions made concerning migration; the latter estimate is considered more robust but the undercounting error may well be above that figure (a reasonable guess would be around 5%) as the actual documented and ndocumented migration inflows are eventually higher than the official estimates that this study relies upon. Undercounting seems higher for males than females and affects the sex ratio of the population. It is suspected that omissions are larger for specific population segments such as foreign citizens but in several instances enumeration difficulties, particularly in big cities and small towns, have been reported for both migrants and natives; these refer mainly to a refusal of people to answer the census questionnaire as well as to the inability of the census enumerators to find members of households at the place of their usual residence. A census post-enumeration survey of 20,000 households has been carried out by the ELSTAT to assess the quality of the statistical information collected, but the results are not available, yet. Obviously, more census material, such as the age-sex composition of the population and the distribution by ethnic group, is needed to produce conclusive results. In addition, matching techniques could be performed to compare census counts with administrative sources of data.
Email: Professor Cleon Tsimbos: email@example.com
Beyond 2011: Tackling the challenges of using administrative data
Victoria Staples, Pamela Dent, Office for National Statistics
The UK Statistics Authority formally established the Beyond 2011 Programme on 1st April 2011 to assess options for meeting future user needs for population and small area sociodemographic statistics in England and Wales. This presentation will present a brief introduction to the purpose of the programme and then focus on the challenges associated with using administrative data to estimate the population. For example, in a conventional census, coverage adjustment is primarily about controlling for under coverage, while over coverage is a minor issue. With administrative sources the key issues are likely to be over coverage and the recording of individuals in the wrong place due to lags in the update cycle for administrative records. We will present our early thinking on how best to address these challenges.
Email: Victoria Staples: firstname.lastname@example.org
Beyond 2011: Using administrative data to produce population estimates - some early results
Jennifer Wall, Office for National Statistics
The UK Statistics Authority formally established the Beyond 2011 Programme on 1st April 2011 to assess options for meeting future user needs for population and small area sociodemographic statistics in England and Wales. This presentation will focus on some of the results from our early investigations into the models based on administrative data. Our focus so far has been on developing options using aggregate data, and we will present our early research findings including an analysis of some first cut trial population estimates generated using these approaches.
Email: Jennifer Wall: email@example.com
Beyond 2011: Meeting the Challenge of Linking Data
Peter Jones, Office for National Statistics
The UK Statistics Authority formally established the Beyond 2011 Programme on 1st April 2011 to assess options for meeting future user needs for population and small area sociodemographic
statistics in England and Wales. Two of the options being researched in the Beyond 2011 programme are underpinned by the use of record level data. Whilst the availability of sources from government departments may support the necessary data infrastructure to achieve this, there are a number of obstacles that need to be overcome to ensure that population statistics derived from administrative sources are reliable and accurate. Key issues to be presented will highlight the following:
Record Linkage and Privacy: A record level model is entirely dependent on the linkage of multiple administrative sources. An approach that preserves the privacy of personal data is a major requirement for the programme.
Data Matching Methods: The absence of common identifiers across the administrative sources will necessitate the use of refined matching techniques that can resolve inconsistencies between the administrative sources and survey data.
Methodological Development: A record level model can be constructed using differing approaches. An address centred approach that identifies people at the household level will produce differing results from a person centred approach that matches people nationally.
Email: Peter Jones: firstname.lastname@example.org