The Routledge Handbook of Census Resources, Methods and Applications

Unlocking the UK 2011 Census

John Stillwell, John Stillwell

• Editorial: Routledge
• Año: 2017
• ISBN:
  9781317188018
• Edición: 1
• Categoría: Naturwissenschaften
• Categoría: Geographie

The collection of reliable and comprehensive data on the magnitude, composition and distribution of a country's population is essential in order for governments to provide services, administer effectively and guide a country's development. The primary source of basic demographic statistics is frequently a population census, which provides hugely important data sets for policy makers, practitioners and researchers working in a wide range of different socio-demographic contexts.

The Routledge Handbook of Census Resources, Methods and Applications provides a comprehensive and authoritative guide to the collection, processing, quality assessment and delivery of the different data products that constitute the results of the population censuses conducted across the United Kingdom in 2011. It provides those interested in using census data with an introduction to the collection, processing and quality assessment of the 2011 Census, together with guidance on the various types of data resources that are available and how they can be accessed. It demonstrates how new methods and technologies, such as interactive infographics and web-based mapping, are now being used to visualise census data in new and exciting ways. Perhaps most importantly, it presents a collection of applications of census data in different social and health science research contexts that reveal key messages about the characteristics of the UK population and the ways in which society is changing. The operation of the 2011 Census and the use of its results are set in the context of census-taking around the world and its historical development in the UK over the last 200 years.

The results of the UK 2011 Census are a unique and reliable source of detailed information that are immensely important for users from a wide range of public and private sector organisations, as well as those working in Population Studies, Human Geography, Migration Studies and the Social Sciences more generally.

Part I
Introducing the census

1
The 2011 Census in the United Kingdom

John Stillwell

1.1 Introduction

Most countries around the world recognise that, in order to provide services, administer effectively and guide development, it is essential to collect reliable and comprehensive data on the magnitude, composition and distribution of their populations. The primary source of basic demographic statistics is a population census which involves “the total process of collecting, compiling, evaluating, analysing and publishing or otherwise disseminating demographic, economic and social data pertaining, at a specified time, to all persons in a country or in a well-delimited part of the country” (United Nations Statistics Division (UNSD), 1997, p. 3). The socio-demographic data that are collected by this process not only support needs assessment, policy formulation and strategic planning, but they also underpin a wide range of research studies and are often used in combination with data from other sources or as a benchmark, a gold standard, against which data from elsewhere may be compared.

On 27 March 2011, censuses were taken by each of the national statistical offices (NSOs) in the United Kingdom (UK) – the Office for National Statistics (ONS) in England and Wales, National Records of Scotland (NRS) and the Northern Ireland Statistics and Research Agency (NISRA) – continuing a tradition that began in 1801 in England, Wales and Scotland and in 1821 in Northern Ireland. In a ‘traditional’ census, questionnaire forms are delivered to and collected from every household by designated enumerators, but the UK is one of a number of countries whose census forms in 2011 were posted out to 25 million addresses with the option of respondents being able to complete and submit their answers to the census questionnaire online or by post and with the additional phase of identification and follow-up of non-respondents.

This introductory chapter of the handbook has three aims. First, in Section 1.2, it provides some global context for the UK Census held in 2011 by reviewing the ways in which populations have been counted in countries across the world in what the UNSD (2013) refers to as the 2010 Census round. Second, given the evidence that many countries, particularly in Europe, have decided to replace traditional census collection methods with approaches that collect demographic data from administrative registers and sample surveys (Valente, 2010), Section 1.3 considers the arguments for and against the continuation of the census in the UK in its current form. This debate, leading to the National Statistician’s recommendation in 2014 for a census in 2021, has underpinned ONS’ Beyond 2011 programme¹ and the subsequent Census Transformation Programme.²

Due to the volume and detail of the data collected for processing, various statistical products have been released by the NSOs in stages commencing in July 2012, with ONS producing all statistics for the UK. In Section 1.4, the different channels of access to these statistics are summarised, and the role and structure of the UK Data Service-Census Support in providing a ‘one stop shop’ for users to access and download aggregate statistics, origin–destination flows, boundary data and cross-sectional microdata from the 2011 Census is explained. In the final section of the chapter, the rationale for the structure and content of the chapters that follow is outlined.

1.2 Counting populations around the world

The 2010 World Population and Housing Census Programme³ of the UNSD reports that 214 out of 235 countries (or areas) conducted some form of population and housing census at least once during the 2010 Census round (between 2005 and 2014) where the term ‘census’ is used in its broadest sense to include traditional censuses, population registers, the use of administrative records, sample surveys and data from other surveys. The UK is one of 60 countries whose census was taken in 2011, in close temporal proximity to population censuses carried out in Portugal (21 March), the Czech Republic (25 March), Curaçao (26 March), Croatia and Poland (31 March). Subsequently, the UNSD (2013) reported the results of two surveys, the first based on a sample of 138 responding countries on how they have implemented their censuses, and a second of 126 countries on what lessons can be learned to inform the formulation of the Third Revision of The Principles and Recommendations for Population and Housing Censuses (UNSD, 2014). The results of the first survey (UNSD, 2011), summarised in Table 1.1, suggest that whilst 83 per cent of the 138 countries responding carried out a traditional census with full enumeration as their main methodology, many countries had developed alternative methods when compared to previous rounds.

Table 1.1 Main census methodology for the 2010 Census round, by geographical region

All the countries in Africa, Oceania, Latin America and the Caribbean responding to the survey took a traditional census whilst all but four in Asia did so; Bahrain and Singapore used a register-based census, whilst Turkey and Israel, together with five European countries (Estonia, Italy, Latvia, Lithuania and Poland), used a combination of data sources. France is the only country in the world which counts its population using a rolling census, whereas nearly one-third of the responding countries in Europe now use administrative registers. The results of the UNSD survey align with those obtained from other studies. An inventory of internal migration data collections among the 193 UN member states built as part of the IMAGE project (Bell et al., 2015) indicates that 88 per cent of the 179 countries collecting migration data did so using a census, although this study confirms an increasing number of countries, particularly in Europe, are using ‘register-based censuses’ or ‘combined censuses’ which link data from registers and surveys. Similarly, the Ethnicity Counts? project (Kukutai et al., 2015), investigating how different countries count their ethnic populations between 1984 and 2014, created a database which provides evidence of an absolute decline in countries undertaking a traditional census between the 1990 and 2010 rounds, the growing use of administrative data, and the wider use of survey data.

In summary, whilst census-taking retains its importance in many countries, particularly in the less developed world, full enumeration using a traditional census is becoming much less popular in more developed countries, particularly in Europe where many countries use population registers as the source of their demographic statistics (Poulain and Herm, 2013) and where opportunities to generate population counts that combine data from different administrative sources using electronic data linkage techniques has gained ground in recent years. In view of these developments, Coleman (2013) refers to the ‘twilight of the census’ as many NSOs consider the future of their census-taking methodologies and explore alternative options that are available. In the UK, the debate on the future of the census is captured in the consultation and research activities involved in the ONS’ Beyond 2011 which, since January 2015, has become the ‘Census Transformation Programme’.

1.3 Census-taking in the UK: benefits and concerns

The question we address in this section is: what are the reasons that are driving the consideration of alternative approaches to a traditional census which has served the country pretty well for over 200 years and which generates a range of products that provide essential data for public administration, governance and research as well as for strategic planning by private enterprises, community groups and voluntary sector agencies?

In England and Wales, the aggregate data products derived from the 2011 Census and released subsequently by the ONS include: population and household estimates plus headcounts for post-codes; univariate Key Statistics (KS) presented in 35 tables and Quick Statistics (QS) which are more detailed data about a single topic available from 74 tables; Detailed Characteristics (DC) with more detailed multivariate data (218 tables) that are not available for the smallest areas (output areas) and Local Characteristics (LC) which are less detailed (184 tables) but are available for small areas; and populations with a base alternative to usual residence that include those with second addresses and short-term residents as well as workplace and workday populations. Whilst each of these products, after statistical disclosure controls have been applied, provides estimates of single area attributes, Origin–Destination Statistics (ODS), also known as flow data, include the counts of flows of migrants, those commuting to work, students and second home owners between two geographical areas. In addition, there are the cross-sectional samples of microdata derived from the full census and the longitudinal microdata that link a sample of individuals between censuses.

As well as the 100 per cent census aggregate and flow estimates and the various unadjusted sample census datasets, there are also data products that are derived from the primary data which are of value to different user communities. Among the examples of these products are the area classifications that are produced by the NSO, by researchers in collaboration with an NSO or by independent researchers in the public or private sectors. For example, the first 2011 Census geodemographic classification was produced for output areas (2011 Output Area Classification (OAC)) in partnership with University College London (Gale, 2014). This is a three-tiered hierarchical geodemographic classification of the whole of the UK consisting of eight supergroups, 26 groups and 76 sub-groups.⁴ Figure 1.1 illustrates the distribution of supergroups in the central and inner suburbs of Leeds, showing the ‘Cosmopolitan’ demographic of much of City and Hunslet, Hyde Park and Woodhouse and Headingley, all census wards with large student populations. To the east and south are found suburbs classified as ‘Multicultural Metropolitan’ (Chapel Allerton, Gipton and Harehills and Beeston and Holbeck), interwoven with areas classified as ‘Hard-Pressed Living’ (such as Burmantofts and Richmond Hill) and ‘Constrained City Dwellers’ (Armley) with increasing numbers of ‘Suburbanites’ and ‘Urbanites’ (e.g. Weetwood, Moortown, Roundhay) with distance from the city centre. More recently, a Classification Of Workplace Zones (WZs) for England and Wales (COWZ-EW) has been constructed at the University of Southampton in collaboration with ONS using k-means clustering based on 48 census variables selected for their ability to differentiate types of workers and workplaces to produce a two-tier WZ classification of seven supergroups and 29 groups⁵ (Cockings et al., 2015).

Figure 1.1 Distribution of 2011 geodemographic supergroups in inner city Leeds, by output area, 2011

The 2011 Census is therefore an unrivalled source of data that are exploited by researchers from a wide range of academic backgrounds. Geographers, in particular, benefit from the high level of coverage that makes it possible to analyse and understand the socio-demographic characteristics of populations in very small areas as shown in Figure 1.1. Output areas (OAs) were introduced across the UK in 2001 and are the lowest level of spatial unit for which census estimates are provided. In 2011, there were 171,372 OAs in England, 10,036 in Wales, 46,351 in Scotland and 5,022 in Northern Ireland, although in the case of the latter, the OAs defined initially in 2001 were aggregated into 4,537 small areas (SAs) in 2011. The requirement to maintain consistency between 2001 and 2011 has meant that only 2.6 per cent of OAs in England and Wales, for example, have changed. Since these are the building blocks for higher-level geographies, this has facilitated spatial analysis of change between the two censuses. Comparisons of small areas between 2001 and 1991 were problematic because data in 1991 were released for enumeration districts, the census collection areas, with OAs being used for the first time in 2001.

In response to a public consultation organised by the ONS (2013) as part of the Beyond 2011 programme, the Royal Geographical Society with the Institute of British Geographers (2014) provided a submission consisting of a number of case studies that highlighted the value of small area census estimates and the importance of their flexibility in answering important societal questions. When detailed geography is less important, the cross-sectional and longitudinal microdata come into their own, providing users with detailed attributes for individuals or households that can be cross-classified according to the researcher’s requirements, unconstrained by the limited combination of variables imposed on the aggregate and flow data to preserve confidentiality.

One of the underlying motivations for the ONS Beyond 2011 review was a concern over the rising costs of delivery aligned with tighter fiscal constraints, a view aired publicly by Francis Maude, Minister for the Cabinet Office in the Coalition Government, and reported in the popular press (Hope, 2010). Maude considered that the decennial census was an expensive and inaccurate method of measuring the population of the UK and revealed that the Government was looking at alternative and less expensive ways to count the population more regularly and make use of existing registers and administrative data sources. The debate on the future form of the census in the UK has been paralleled by similar debates reported by the press in other countries around the world including, for example, the USA (Singer, 2010), Canada (Underhay, 2011) and Australia (Hutchens and Martin, 2015).

The need to control cost is by no means the only concern of those who consider the days of the traditional census to be numbered. The list of concerns of those commenting on the census debate around the world includes issues of privacy, frequency, accuracy and apathy, as well as the pressure to make better use of information available elsewhere (Valente, 2010; Fienberg and Prewitt, 2010; Coleman, 2013).

In the UK, as in other countries like the USA and Canada, there is tension between the need for the Government, through its NSOs, to collect information on which to formulate policy and to base legislation and the fundamental human rights of individual privacy and freedom from discrimination. Fienberg and Prewitt (2010) indicate that privacy was a major issue in the 2010 Census in the USA, with questions relating to age, gender and ...