Strand organisers: John Hollis, Greater London Authority, and Cheng Yong Lee, Kent County Council.
Delving deeper: estimating unknown multi-way relationships at sub-district level
Paul Williamson
University of Liverpool
In the UK, the planned 5% sub-district Sample of Anonymised Census Records appears to offer the potential to estimate previously unknown multi-way relationships. This paper will assess the reliability of estimates derived from this source, and compare it to the reliability of equivalent estimates derived through two alternative methods, one of which is currently used by a consortia of regional and local governments in Australia.
Paul Williamson, Department of Geography, University of Liverpool, Liverpool L69 3BX
Tel: 0151 794 2854
Email: p.williamson@liverpool.ac.uk
Putting shape to the 2011 Census
Peter Benton
Office for National Statistics
The 2011 Census may seem a long way off to some, but for those of us involved in its development, time is passing quickly. We are now in the process of finalising the detailed operational design to take into the 2007 Census test, collating responses to our recent consultation on the content of the 2011 Census questionnaire and starting work on our output strategy, to mention but a few activities. This presentation will give an update on progress to date and outline how we plan to proceed from here.
Peter Benton, Deputy Director, 2011 Census
Email: peter.benton@ons.gov.uk
The development of the 2011 Census in Scotland
Ian Maté
General Register Office, Scotland
Currently GROS are planning a traditional enumerator-centred fieldwork with a local postback option. This contrasts strongly with ONS's current proposals based around mobile field forces and technological solutions. There is less difference once the forms get into the processing hall.
This presentation will highlight what GROS considered were the main problems in 2001, our evaluation of those problems and our response to them. Our response to them includes contingency arrangements if RM falters - for enumerator collection - , in-house pay systems, changes in the pay rewards, combined with a strong programme of community involvement over the next 5 years. It may be considered a traditionalist's view, as opposed to the more radical ONS proposals.
Email: ian.mate@gro-scotland.gov.uk
Andy Bates
Office for National Statistics
An Office for National Statistics (ONS) project was set up to investigate which datasets and methods could be used to produce an authoritative set of small area population estimates for England and Wales. Following on from this, ONS released in April 2005 mid-2001 and mid-2002 ward population estimates as experimental statistics.
The presentation will cover the production of the mid-2001 ward estimates, and an explanation of the Ratio Change methodology and data sources used to produce the mid-2002 estimates. For the mid-2002 estimates there will be case studies of some of the more challenging wards for which estimates were produced.
When the experimental estimates were published there was a period of user consultation with the estimates. The presentation will cover the responses received as part of this user consultation.
ONS will also outline further publication plans for small area population estimates, including the production and publication of estimates for Super Output Areas, and the intended re-evaluation of the three previously identified shortlisted methods.
Email: andrew.g.bates@ons.gov.uk
Kent County Council and the setting of Local Public Service Agreement Targets
Nick Moon
Kent County Council
The agreement of the second generation Local Public Service Agreement in Kent, between the county council and central government, necessitated, in a number of instances, the use of demographic data in order to improve the process of negotiation and the setting of targets. Local PSAs might best be described as voluntary agreements between an individual local authority and the Government, which are designed to stretch performance in the delivery of local public services by focusing on targeted outcomes. The agreements demand that robust, evidence-based targets are identified to act as a basis for the agreement, which should ultimately be something which is of benefit to the local population.
To this end, many of the targets that Kent County Council have agreed with the Government have been based on a variety of demographic data and are supported by an agreed definition of what precisely is being counted. This paper outlines the methodology used by the officers at Kent County Council when setting the LPSA2 targets in general, but will concentrate on one particular target, namely that of increasing the number of long-term Incapacity Benefit customers who are in work. The setting of this target was underpinned by research data supplied by the Social Disadvantage Research Centre at Oxford University and the Department of Work and Pensions, but the use of a wide variety of demographic data has meant that the council and its partners (Jobcentre Plus, the NHS and a variety of agencies in the public, private and voluntary sectors) are able to target resources available to where the needs are greatest.
The paper also outlines the original need for better and improved data at a local level. In the initial stages of the LPSA there was a paucity of local data, which necessitated the commissioning of the Oxford University study. Simple benefit claimant counts, however, would not have been suitable as a means of defining success or failure of any intervention - more sophisticated analysis was necessary to identify the changes in life chances brought about by such interventions. The coupling of demographic data and that gleaned from the Oxford study has resulted in there being a brighter future for the provision and utilisation of local data such as that used in the setting of this target.
Nick Moon, Supporting Independence Programme, Sessions House 1.69,
County Hall, Maidstone ME14 1XQ, Kent
Telephone: 01622 696932
Email: nick.moon@kent.gov.uk
Simon Adams
Tribal Secta
The Department of Health Green Paper 'Independence, Well-Being and Choice' is the latest in a long line of policy papers that advocates more older people being helped to live at home. While the Green Paper does stress the importance of the prevention and well-being agendas in order to promote independence and active life into old age, it seems to assume that more older people being helped to live at home is automatically a positive outcome. In fact there may be an optimal level that reflects need and dependency, as well as a range of other factors, for each local authority area.
The "more is better" approach is reinforced by the Social Services Performance Assessment Framework (PAF). For AO/C32, which covers eight defined sets of activity recorded through RAP for older people helped to live at home (OPHTLAH), the optimal 5 "blob" performance band is set at 100 per 1000 people aged 65+ and above. In other words, and unlike some of the other performance indicators in the PAF, there is no upper limit beyond which local authorities' performance rating reduces. In 2003/04 some authorities reported very high levels of support, with four authorities reporting that more than one in seven of all older people in their area were provided with Social Services support.
Such levels may be appropriate given the characteristics and circumstances of older people in those authorities, but they do at least raise questions about whether the needs of the population may vary from one authority to another. They do also beg the question whether levels of support as high as this may be fostering dependence, rather than promoting independence.
These concerns led to a study "Predicting the numbers of older people helped to live at home: A needs-based model and the implications". This has been undertaken by an ADSS Eastern Region Performance Network, chaired by Hugh Gault (Cambridge County Council) and supported by Simon Adams of Tribal Secta.
A theoretical model was developed by the project group, seeking to identify the characteristics of those older people that are likely to require the provision of publicly funded social support services to enable them to live at home. This model highlights that four sets of factors are likely to be significant at the population level:
· the health, deprivation, disadvantage and dependency characteristics within the older people population.
· the availability of direct access and preventive
· the availability of alternative forms of support, some of which may be purchased from the private sector
· the affluence of the population and their ability to purchase support from the private sector as required rather than having to rely on public care.
· the age-balance within the older people population, given the association of dependency with increasing age.
A range of data variables were assembled, using Census 2001, DWP, the updated Index of Multiple Deprivation (at local authority level) and Local Authority Social Services activity and budget data. We investigated 48 similar authorities for correlations to the reported levels of the PAF C32 (Older people helped to live at home per 1000 population aged 65+).
This work indicated that there were significant correlations between some of the indicators of deprivation and the numbers of OPHTLAH, prompting us to test the hypothesis further by expanding the dataset across all England authorities. Again the results were clearly indicating the links, so we set about developing a regression analysis model that gave us a credible predictive formula at Local Authority level.
Conclusions
The overall conclusion is that "more does not necessarily equal better" where OP HTLAH are concerned. On the contrary, the number of older people helped to live at home should reflect the level of need and dependency in each community, the availability of direct access and preventive services that reduce risk and dependency, and the requirement for public social care as a result of deprivation and the inability to purchase care.
The assumption that a linear "League Table" for the presentation of Social Services "performance" in helping older people to live at home is essentially flawed, as it takes no account of the levels of local need. In the majority of cases, those authorities that are at the "bottom" of the league table are there because there are less people in their local community requiring social services, not necessarily because they are performing poorly.
Authorities are therefore being set unrealistic targets, which provide unnecessary concentration of effort and may lead to wasteful allocation of scarce resources.
Simon Adams, Senior Consultant, Tribal Secta, 97-91 Newman Street, London W1 3EY.
07968 616285
Email: simon.adams@tribalsecta.co.uk
Integrated approach to local level planning by local government, considering population as the weighing factor
Jagannath Ojha, Ministry of Local Development, Jawalakhel, Nepal
This paper examines the prevailing planning process in local level institutions in Nepal and suggests the appropriate method to identify real needs of people using the objective tools. The study is based on a field survey of one of the district of Nepal, consisting of 18 VDC, and 151 Settlements. The study tries to analyze the variables to be considered to examine the real needs of people as part of local level planning with the concept of people-centred development. The variables considered in the study were sum of time taken to get the desired services, quality factor of service stations and local priority that people perceived, with population as the weighting factor. The study attempts to establish the empirical relationship between these variables with population as weighing factor. The empirical relation was established by sharing the experiences with concerned stakeholders. There are two schools of thought: one favouring the access right of people, and the other is the economic viability of the service. This study tries to set a relationship between the combination of variables and size of population which is considered as the weighing factor. The level of accessibility is the most important factor for the identification of real needs, mathematically, which is based on accessibility index (AI).
The detailed method of using the tool will be explained in the presentation.
Jagannath Ojha, Department of Local Infrastructure Development and Agricultural Roads, Ministry of Local Development, Jawalakhel, Nepal
Email: ojhajagan@yahoo.com
The policy Delphi: a tool to shape the future
Maura Misiti and Adele Menniti
Istituto di Ricerche sulla Popolazione e le Politiche Sociali, Rome
The paper presents the results of a research based on a policy-Delphi survey carried out in Italy in 2003, whose aim was to design population and society scenarios up to 2030. The study is part of a Europe-wide research project conducted in 15 European countries. The research follows an ad-hoc methodology that considers both quantitative and qualitative perspectives. The method integrates Delphi, Appreciative Inquiry and SWOT Analysis. The overall approach - from the methodology, to the expected results - is designed to provide tools and inputs to shape policy interventions at national and local level. Scenarios were designed following the comments and suggestions indicated by a panel of experts.
Three main issues were dealt with in terms of policy developments: the elderly, family and fertility and gender roles. All the issues are strictly interrelated and animate the Italian scientific, political and social debate.
The paper firstly describes the applied methodology then highlights the main findings of the study. Suggestions are presented on how to set the Italian policy agenda for the next thirty years: ageing should not only be dealt with by economic interventions but also through actions aiming at increasing the social inclusion of the elderly and improving their quality of life. Gender roles might change supporting the presence of women in the labour market via work-life balance measures, women's empowerment, and abating gender stereotypes. Finally, family issues will be challenged through a number of interventions aiming at creating a more family friendly society.
Maura Misiti and Adele Menniti
Istituto di Ricerche sulla Popolazione e le Politiche Sociali
Consiglio Nazionale delle Ricerche
128, via Nizza-00198 Roma Italy
+39 06 4993 2868 (direct)
+39 06 858 34506 (fax)
Email: m.misiti@irpps.cnr.it
Enumerating the Republic of Moldova Census, 2004
Malcolm Brown and Maia Godonoga
Cornwall County Council & University of Plymouth
Malcolm Brown had the good fortune to be selected as one of the International Observers for the Census in Moldova in October 2004. Maia Godonoga was one of the Moldovan interpreters and we worked together with our driver, her brother in law, Sergiu, monitoring how the Census went in the Orhei region.
Moldova was part of the USSR and is now a very small republic. Although the poorest country in Europe, it is attractive, there is a strong sense of community and the people are very hospitable. We will explain our respective perspectives and impressions of the country including its main demographic and socio-economic characteristics.
The Census was delayed several times and there were practical difficulties organising it because of internal political problems, ethnic tensions and a shortage of money. However, it eventually took place with a heavily devolved structure and high calibre enumerators who conducted interviews and completed the forms themselves.
The topics had some similarities with British experience but questions were
more sophisticated in respect of residence definitions, duration of residence and languages. Differences with the British Census will be discussed.
We will explain how we came to work on the Census, our respective roles, how we worked together and our impressions of how the Census was undertaken and how the monitoring process was implemented.
We were asked to watch very closely how the enumeration was undertaken. Generally it was impressive but there were problems. There were inconsistencies with how students and people working abroad were treated and answers did not all seem accurate. Some answers to the languages question were exaggerated.
Moldova is a fascinating country and there are also lessons we can learn from how they conducted their Census.
Email. mbrown@cornwall.gov.uk
Making use of the 2001 Census in Cambridgeshire PCTs.
Vicky Head
Cambridgeshire County Council and Anglia Support Service
At a local level, Primary Care Trusts (PCTs) are becoming important users of Census and other socio-economic and demographic data. The current emphasis on PCTs' public health function, including health promotion initiatives, has led PCTs to adopt an increasingly population-based approach to their work. This presentation outlines how Cambridgeshire PCTs have been making use of the 2001 Census and describes the benefits of a joined-up approach to Census analysis. In Cambridgeshire, joint NHS/County Council work has helped PCTs access and interpret data, through publications such as area-based profiles and topic-based reports. The presentation focuses in particular on using the Census to identify local health inequality - by geography and by ethnic group - and the importance of considering age-structure variation when doing so.
Vicky Head, The Research Group, Cambridgeshire County Council, RES1203,
Shire Hall, Castle Hill, Cambridge, CB3 0AP
Telephone: 01223 717208
Email: vicky.head@cambridgeshire.gov.uk
Vicky holds a joint post between Cambridgeshire County Council and Anglia Support Service (which provides support services, including public health information, to the PCTs in Cambridgeshire and Peterborough).
Using the 2001 Census microdata to explore household formation and ethnicity:
a multilevel modelling approach.
Abdelouahid Tajar and Ludi Simpson
University of Manchester
We examine how household formation varies between ethnic groups and whether and why there is additional geographic variation. We provide headship rates by ethnic group, local authority district (LAD), age and sex, to be used in a household projection.
This paper studies the household structure in England and Wales using the CAMS (The Controlled Access Microdata Sample), a 3% sample from the UK 2001 census data.
We review the existing literature in this area of social research. We consider in a first step standard multiple regression models in a single level analysis to explore the headship rate in England and Wales for both males and females, taking into account the age and the ethnic dimensions, marital status and country of birth.
We investigate further the relationship between these variables by introducing a geographical factor using multilevel modelling to allow for hierarchical structures in the data. These techniques are employed to estimate simultaneously variations between individuals and between geographical areas after allowing for the variation of age, ethnicity and sex.
Using the CAMS and the random intercept multilevel logistic models, it has been possible to show how headship varies between areas. More specifically we test and explore how the headship rate curves differ among LADs in England and Wales, for the white population and each ethnic minority. We investigate whether the major distinctive characteristics of an ethnic group's rates are repeated in different parts of UK, by considering the regions of Britain and the major cities which contain the largest number of ethnic groups.
We investigate whether the household is more or less likely to be headed by women for ethnic minorities; we examine the headship at different ages between ethnic group and sex in different areas. The methodology is also applied to model concealed family rate. Illustrations, using some local authority districts in England and Wales compared to the national rate, are provided.
Abdelouahid Tajar and Ludi Simpson, The Cathie Marsh Centre for Census and Survey Research, School of Social Sciences, University of Manchester, Crawford House, Booth Street East , Manchester M13 9PL
Tel: 0161 275 0277 Fax: 0161 275 4722
Email: abdelouahid.tajar@manchester.ac.uk
From households to population?
Briony Eckstein
Office for National Statistics
Administrative datasets usually do not exist for the explicit purpose of measuring the population, and thus it is likely that they are not suitable to be used to make population estimates purely by aggregating counts. This paper presents work in progress. It considers whether information from administrative data sets relating to housing and property might be sufficient in detail and quality to be used in the estimation of household and population numbers down to Local Authority level.
Briony Eckstein, 2005 National and Local Studies, Office for National Statistics
Segensworth Road, Titchfield, Hampshire PO15 5RR
Tel. 01329 813672
Email: briony.eckstein@ons.gsi.gov.uk
The use of administrative data sources in local population estimates
Piers Elias
Tees Valley Joint Strategy Unit
This presentation will look at some of the different data sources that are available from either central Government Departments or from your own Local Authority or Health Authority and how these can be used to enhance other core data sources, such as Census data, in the pursuit of a population estimate that has some weight and meaning in the context of providing services at local level.
Local Authority demographers are constantly being asked to provide good quality data for ever-decreasing and non-standard geographies in shorter time-scales; having access to individual records, either at household level or at person level, does allow this flexibility, provided you have the resources and expertise to maintain and analyse the datasets, and can act as more than just a quality check.
So, what is available? Where do you start? What obstacles might you encounter? How much does it cost? (Financial, human) What are the pros and cons of each dataset? Is it all worth it? These are just some of the questions I hope to be able to answer.
Piers Elias, Tees Valley Joint Strategy Unit, PO Box 199, Melrose House, Melrose Street, Middlesbrough TS1 2XF
Telephone: 01642-264818
Email: piers.elias@teesvalley-jsu.gov.uk
Julie Jefferies
Office for National Statistics
In October 2004, ONS published a revised back series of population estimates for 1992 to 2000 at both the national and local authority levels. These final estimates were produced after a long period of research into the best methodology to use. For the first time the back series also incorporated both the revised 1991 population estimates published in February 2003 and the revisions to international migration estimates for the 1990s.
After explaining the background - both the intercensal difference remaining after the adjustments to the 2001 estimates and the need to improve on the methodology used in the interim revised back series - the presentation will focus on the research undertaken.
The key decision was how to split the remaining intercensal difference (209,000) by single year of age and sex back over the ten years from 1992 to 2000. Two questions were investigated in depth:
Should ONS use a period approach or a cohort approach?
Should the difference be split evenly over the ten years (on the assumption that the difference had built up in a linear fashion) or should some form of weighting be applied?
The presentation will illustrate the different methods explored and their advantages and disadvantages. The final solution chosen was a cohort method weighted for migration by age and time with a period adjustment, combining the best features from various methods. Finally we will look at the quality assurance of the revised estimates and the wider impacts of the improvements to these population estimates.
Julie Jefferies, Population Estimates Unit, ONS, Segensworth Road, Titchfield,
Hampshire PO15 5RR
Email: Julie.Jefferies@ons.gsi.gov.uk
Residents and staff in communal establishments: data quality issues in the 2001 Census.
Levin Wheller and Madhavi Bajekal
Office for National Statistics
In the 2001 Census, when forms for residents of communal establishments were completed on a proxy basis by staff from the establishment, information about the resident's position in establishment was sometimes misrecorded. This has resulted in an incorrect split between numbers of residents and staff in a number of communal establishment tables. These problems are most apparent for medical and care establishments and particularly affected older age groups.
ONS are currently working on developing a robust methodology to provide more accurate figures on staff and resident numbers in 2001. This methodology will then be used to produce a set of revised tables. Guidance will be provided to users of both aggregate and unit-level Census data (e.g. Sample of Anonymised Records).
The presentation will outline the main data quality issues identified, discuss potential explanations and set out ONS' plan of action to produce and publish revised tables.
Email: madhavi.bajekal@ons.gov.uk