Local authority, census and planning issues. Strand organiser: John Hollis, Greater London Authority.
Changes in tenure patterns by ethnic group in London boroughs, 1991 to 2001.
Greater London Authority
This paper will look at changes in the tenure distribution of households in London boroughs between 1991 and 2001. It will look at the extent of the increase in private renting across London and investigate any differences between ethnic groups in particular boroughs. It will also look at changes in levels of owner occupation and whether the traditional patterns of change (moving out of Inner London for areas of cheaper housing to buy) are still apparent.
Samples of anonymised records 2001: Progress and potential.
University of Manchester
Census Microdata enable users to undertake more flexible analyses than is possible using census tables. Users are free to explore individual and household characteristics, produce models, undertake data manipulation and to define their own tables using samples from the Census output database.
A full range of anonymised microdata are now available from the UK 2001 Census. The family of datasets now includes:
The Licensed Individual SAR: a 2% sample of individuals; available under a standard end user license
The Special License Household SAR: a 1% sample of households in England and Wales only; available under a special license
The Small Area Microdata file: a 5% sample of individuals, with less detailed individual information but with local authority geography; available under a standard end user license
The Controlled Access Microdata: more detailed versions of the 2% individual and 1% household files available for use only within an ONS safe setting.
Confidentiality has been achieved by limiting the detail of key variables in the end-user license data files, and by providing more detailed data under more stringent licenses or within safe settings.
Overall, these data now provide a range of data options, including many variables that were previously unavailable such as local authority and religion.
This paper will discuss the range of data available, access routes and the potential of these data for demographic and other social research, including use by local authority researchers. Further information about these data is available on the SARs website at http://www.ccsr.ac.uk/sars .
Cornish ethnicity data from the 2001 Census
Cornwall County Council
About 34,000 people in Cornwall and 3,500 people in the rest of the UK wrote on their census forms in 2001 that they considered their ethnic group to be Cornish. This represented nearly 7% of the population of Cornwall and is therefore a significant phenomenon.
This response followed some publicity before the Census but does not seem to have been co-ordinated tightly.
The numbers in themselves can not be taken as an accurate count of the number of Cornish people in Cornwall. However, the County Council anticipated that the information would be interesting, primarily to test whether there were any significant differences in characteristics between the Cornish and non Cornish populations.
Additional Census tables have been commissioned and their results will be presented and the implications discussed with other relevant data sources.
This issue is significant because planning for the 2011 Census is now well advanced and Cornish ethnicity is under consideration. Many people in Cornwall hope that a strong case for Cornish being a tick box option in 2011 can be made. The County Councils research will hopefully inform this debate.
Study of migration based on the statistics of the 2001 Census.
Stoke on Trent Council
The purpose of this paper is to set out my experience of using statistics from the 2001 Census to study migration and travel-to-work patterns using the City of Stoke-on-Trent as an example. We explore the shortcomings and advantages or disadvantages of the available statistics; discuss what they are measuring, what they meant to measure and whether or not they are fit for purpose. The conclusion is that for the study of migration the Census has the shortcoming of being just a snapshot at one particular point in time. Suggestions are made that might make the Census more dynamic.
The ONS Improving Migration and Population Statistics Project: A progress update.
Roma Chappell and Amanda Blunden
Office for National Statistics
Through the IMPS project ONS are making a substantial investment to improve migration and population statistics. This presentation explains what the IMPS project covers and provides an update on recent progress on the high priority research work IMPS is doing. This includes an update on the work on definitions, the LA Case Studies and most recently the work on the Interdepartmental Task Force on international migration statistics, established in May this year. This presentation complements the other more detailed ONS sessions on improving the geographical distribution of international migrants, short-term migration and the use of alternative data sources.
Email: Roma.Chappell@ons.gsi.gov.uk . Amanda.Blunden@ons.gsi.gov.uk
Improving the distribution of international in-migrants at UK country and Government Office Region (GOR) levels
Office for National Statistics
The International Passenger Survey (IPS) is the main source of data used in ONS estimates of total international migration. The IPS records in-migrants intended area of residence in the UK upon entry. In-migrants could, in fact, settle in a different area of the UK, either directly, or by moving on quickly from their initial destination.
This paper presents research into the use of international migration data derived from the Labour Force Survey (LFS) to distribute IPS in-migrants at UK country and GOR level. The paper describes the method developed to produce geographic distributions in this way. It also considers the impact of this method on international in-migration estimates for UK countries and GORs.
Methods for distributing international in-migrants at local authority district level.
Office for National Statistics
In addition to researching methods for producing in-migration estimates at regional level, ONS is researching methods for distributing regional in-migration totals to local authority districts. In-migration estimates at local authority district level are required by single year of age and sex for use in both the mid-year population estimates and sub-national population projections. This talk will cover three aspects of the methods:
1. The sub-regional geography on which International Passenger Survey data should be used.
2. Smoothing methods for producing stable sub-regional estimates
3. Age distributions applied to local authority district totals
Using alternative sources to improve the estimation of population and migration statistics.
Briony Eckstein, Helen Evans and Folkert van Galen
Office for National Statistics
Our rapidly changing society is providing ever increasing challenges to those attempting to measure and estimate the population. Changes such as increased population mobility, lower fertility and mortality rates, more complex living arrangements and shifting societal norms in the areas of partnership and family formation all add to the problems of producing accurate population estimates on a usual residence basis.
Many of the statistics produced by the Office for National Statistics are calculated through the use of administrative data sources. For example, the mid-year estimates (MYEs) are calculated using such data sources as birth and death registrations from the Registrar Generals Office, GP patient register and National Health Service Central Register data and the International Passenger Survey. This paper considers whether the inclusion of additional administrative data sources (such as data on the allocation of National Insurance numbers to migrant workers, data on the transfer of state pensions overseas and the Workers Registration Scheme) might help to improve the quality of migration and population statistics.
Email: email@example.com , firstname.lastname@example.org , email@example.com
The General Register Office for Scotland (GROS) publishes population projections for Scottish local authorities. These are a potential source of information for local authority strategic planning. In addition, each local authority holds administrative data on the current client group for key services. However, the quantity and quality of this information varies according to the service in question.
Both sources of information on the local population have strengths and weaknesses.
The GROS population projections are particularly sensitive to the migration and fertility assumptions made, but provide comprehensive information on the future population at a local authority level. Variant assumptions can be made, reflecting the assumptions applied to variant national projections. These reveal a high degree of uncertainty over the school age and younger working age population, but much greater certainty over the future numbers of older people.
By contrast, the local authoritys own administrative data can provide current information at a more localised geographical level. It can also provide a near-comprehensive profile of the current school age population. However it provides only partial information for most other client groups.
This paper: (i) reviews the strengths and weaknesses of GROS sub-national projections and administrative data; (ii) summarises recent attempts to quantify the quality of information provided by each source for each client group; and (iii) discusses ongoing work to improve the quality and utility of information used for strategic planning, by synthesising these two sources.
People, projections and projects.
Essex County Council
Population projections are used extensively during the preparation of development plans. But, usually plans make proposals which result in adjustment of past demographic trends. There is therefore need to assess the potential demographic outcome of the plan in order to inform implementation and funding programmes. The paper will assess the issues raised by this process. It will also examine the implications for justifying local and sub-regional initiatives within national funding programmes.
Demographic estimates for London: The truth?
Greater London Authority
A review of population and household estimates for London boroughs and Greater London since the 2001 Census and a critique of the ONS and DCLG subnational population and household projections. The presentation will cover the conversion of 2001 Census data to form the 2001 population and household bases, the use of national and international migration data in creating estimates up to 2005 and the conversion of migration estimates to form the population projections upon which the household projections are based.
The presentation will look at ways of improving the estimates and therefore developing a more secure base for projections.
Preparations for the 2007 Census Test in Camden.
Camden is one of 5 local authorities to be selected to take part in the 2007 Census test and is the only LA in London taking part. ONS is going to use the Census Test to evaluate several different aspects of its enumeration procedures, but a crucial role is to assess the value of close liaison with local authorities in maximising enumeration in the census.
Why Camden? In the 2001 census, according to One Number Census (ONC) analysis, Camden faired badly, with 23% of population and 24% of households undercounted. ONS quality indicators further showed that 16 out of 18 Camden wards were in their category 4 (over 20% under enumerated) and that the other 2 were category 3 (10-20% under enumerated). At the lower geography, 75% of Camdens Output Areas (OAs) were category 4 and 24% were category 3. Of 733 OAs, just 3 were category 2 (5-10% under enumerated) and only 1 in category 1.
Camden is high on the latest ONS index of hard to count areas: it is home to both established and recent migrant communities; a high proportion of students in both halls of residence and in private households; a high proportion of accommodation in public and private estate blocks with gated/secure access; houses in multiple occupation and addresses with unmeasured living spaces; a highly mobile population and one with a younger demographic which tends to defy normal working/waking hours. At the same time, Camden has expressed a desire to help improve the Census and, with the emphasis on post-out questionnaires and the need to better understand local address geographies, Camden is also recognised as having a well established and good quality local land and property gazetteer (LLPG).
Local authority liaison could well prove be a key part of improving enumeration performance in the next Census. Camden will be liaising with ONS to give a better understanding our local communities and geography, better prepare census staff and our residents for the arrival of the census and to build the importance of the Census into the psyche of the borough.
We are conscious that any lessons learned will set us in good stead for real thing in 2011 and the knowledge gained should benefit all London boroughs in the run up to the 2011 Census. The modest input of resources into liaising with ONS now could be nothing in comparison with the potential costs associated with a poor enumeration.
Developing a questionnaire for the 2011 Census
Office for National Statistics
The 2011 Census will collect a range of information about every resident of the UK, and the Office for National Statistics has a significant programme of work underway to determine what this will include. This presentation will review progress so far which includes analysis of the 2001 Census, consultation with users and stakeholders and the development of questions for the 2007 Census Test. The process and timetable for continuing question development will be discussed and ONS' current judgement on the likely content of the 2011 Census will be outlined.
Methods of geographical perturbation for disclosure control
University of Southampton
Disclosure Control methods are used to protect the confidentiality of individuals and households in published census data. This paper will describe research into some methods designed to protect against disclosure which might arise when geographies are published that overlap and can be differenced to produce slivers. These slivers or small areas usually contain small numbers of people, particularly so in rural areas, which increases the occurrence of unique records. Unique records exist when an individual or household is the only one in an area with a particular combination of characteristics. Disclosure control is applied to the data to minimise the probability of identification of these unique records.
Some new ideas for geographical perturbation of household records for disclosure control will be described, performed on a synthetic census dataset. Random record swapping was carried out on the 2001 UK census and is used as a benchmark for assessing the new methods. Results of two new swapping approaches will be shown; firstly perturbation of households irrespective of geographical boundaries and secondly a spatially-sensitive approach, perturbing households according to local population density.
The aim is to create uncertainty in the data, measured by the conditional probability that a published value of one in a small area is actually the true record. At the same time, we try to preserve the statistical properties of the data.
Review of statistical disclosure control methods for census frequency tables
University of Southampton
This paper provides a review of statistical disclosure control (SDC) methods for standard tabular outputs containing whole population counts from a Census (either enumerated or based on a register). SDC methods implemented at Statistical Agencies for protecting Census tables include both pre-tabular and post-tabular methods or combinations of both. Pre-tabular methods are implemented on the microdata prior to the tabulation of the tables and typically include forms of record swapping between a pair of households matching on some control variables. This method has been used for protecting Census tables at the United States Bureau of the Census and the Office for National Statistics (ONS) in the United Kingdom. Record swapping can be generalized into a pre-tabular method called PRAM (the Post-Randomization Method). This method adds noise to categorical variables by changing values of categories for a small number of records according to a prescribed probability matrix and a stochastic process based on the outcome of a random multinomial draw. Post-tabular methods are implemented on the entries of the tables after they are computed and typically take the form of random rounding, either on the small cells of the tables or on all entries of the tables. Small cell adjustments (rounding) have been carried out on the Census tables at the Australian Bureau of Statistics (ABS) and the UK ONS, and full random rounding has been carried out at Statistics Canada and Statistics New Zealand. A fully controlled rounding option which uses linear programming techniques to round entries up or down and in addition ensures that all rounded entries add up to the rounded totals may be a viable solution in the future but at the moment is not able to cope with the size, scope and magnitude of Census tabular outputs. Other post-tabular methods include cell suppression or some form of random perturbation on the internal cells of the Census tables.
Few evaluation studies have been carried out on the impact of SDC methods on disclosure risk and the resulting utility and quality of Census tables. The approach for assessing SDC methods is based on a disclosure risk-data utility framework and the need to find the balance between managing disclosure risk while maximizing the amount of information that can be released to users and ensuring high quality outputs. To carry out this analysis, quantitative measures of disclosure risk and data utility are defined which then determine optimal SDC methods and their parameters. We present the analysis on common methods for SDC on Census tables, and in addition examine utility measures which assess not only distortions to distributions but the impact on statistical inference.