Transcript Slide 1

Opportunities and Challenges of the
2011 Census: A view from academia
Tony Champion
[email protected]
Paper for the TWRI Policy & Research Conference ‘Making best use of
the 2011 Census’, St William’s College, York, 5 October 2012
Introduction 1
Academic users of the Population Census cover a wide
range of interests, but these can be summarised under
three main sets of purposes:
1) Deriving descriptive contextual information on
populations of interest;
2) Analysing data to improve our understanding of UK
society and how it is evolving, often using advanced
quantitative techniques;
3) Benchmarking other surveys to ensure that they
represent the total population of an area.
And these are not just for ‘academic purposes’: much work
is also carried out under contract for local and central
government, research foundations, and the private and
third sectors
Introduction 2
Features of the census that are most prized by academics
(according to submissions to the Science & Technology
Committee Inquiry into the Census and Social Science and
to the ONS’s Beyond 2011 Public Consultation on User
Requirements) include:
- opportunities for multivariate analysis
- fine-grained geography
- high degree of accuracy even at small-area level
- ability to study change between censuses
This is reflected in their high usage of other census datasets
as well as Area Tables: e.g. Origin-Destination Tables,
Samples of Anonymised Records, Longitudinal Studies
The rest of this paper presents examples of benefiting from
these census qualities in my previous research on these
datasets, then looks forward to using the 2011 outputs
Examples from Area Tables
Example 1: from Key Statistics
The GB local authorities with the 10 highest and lowest
proportions of migrant residents, 2001
Rank
1
2
3
4
5
6
7
8
9
10
Highest
Oxford
Cambridge
City of London
Westminster
Wandsworth
Camden
Hammersmith and Fulham
Richmondshire
Kensington and Chelsea
Manchester
%
25.6
24.9
23.0
22.9
21.0
20.6
20.1
19.7
19.4
19.3
Rank
408
407
406
405
404
403
402
401
400
399
Lowest
East Dunbartonshire
Havering
East Renfrewshire
Knowsley
Rochford
North East Derbyshire
Dudley
South Staffordshire
Castle Point
Ellesmere Port and Neston
%
7.3
7.7
7.9
8.0
8.0
8.0
8.0
8.1
8.1
8.1
Example 2: from Area Table
Net within-UK migration rate, age groups,
for Rural Towns and Other Rural, England
4
Rural Towns
Villages & scattered
% residents in age group
2
0
0-15
-2
-4
-6
-8
-10
16-19
20-24
25-29
30-44
45-59
60-74
75+
Example 3: from
Commissioned
Area Table
Districts most
affected by the outmigration of fulltime students (%
out-migrants aged
16-74)
Darkest = 20-44%
Example 4: From several censuses
Population change by local labour market area types,
1901-11 to 1981-91
15
% for period (NB. 1931-51=20 yrs)
10
5
0
1901-11
1911-21
1921-31
1931-51
1951-61
1961-71
-5
Large City
-10
Middle-sized City
Small City
Rural Areas
1971-81
1981-91
Examples from the SARs
The Samples of Anonymised Records 2001 comprise a
suite of datasets: 1% Household SAR & CAMS, 3%
Individual SAR & CAMS, and 5% Small Area Microdata
[CAMS = Controlled Access Microdata Sample]
The SARs allows any combination of crosstabulations
including some not available anywhere in Area Tables:
e.g. it is the only standard census dataset where
commuting and migration variables can be related
CAMS is extremely rich in detail on the characteristics and
geography (e.g. LAD of address currently and one year
ago and LAD of workplace)
Examples: (1) Age distribution (single-year) of one-year
migrants by distance of move; (2) Length of commute by
type of LAD and one-year-migrant status …
Example from the SARs 1:
x
Age profile of within-UK migrants moving <3 km and 50+ km
8.0
7.0
<3 km
50+ km
6.0
% all ages
5.0
4.0
3.0
2.0
1.0
0.0
0
5
10
15
20
25
30
35
40
45
50
55
Age at census (years)
60
65
70
75
80
85
90
95+
Example from the SARs 2:
Proportion commuting 20km+, by urban/rural district type
comparing recent in-migrants with non-migrants
% commuting 20km+
0.0
England
5.0
10.0
15.0
20.0
Non-migrant
Migrant
Major Urban
Large Urban
Other Urban
Significant Rural
Rural-50
Rural-80
25.0
Examples from O/D Statistics
Origin & Destination Statistics comprised two main sets in
2001: Special Workplace Statistics and Special Migration
Statistics (plus Special Travel Statistics in Scotland)
One major use of SWS has been for Travel to Work Areas
(TTWA) & other regionalisations like City Regions (CRs)
by Mike Coombes et al (Newcastle University)
In recent work Mike and I compared commuting patterns
across 3 censuses to look for any tendency for ‘Pennine
England’ to become a more integrated polycentric region
One use of the SMS has been to track attraction and
retention of migrants by skill group, 2001, seeing how
other large Cities compare with and relate to London …
Pennine England:
5 Cities & City Regions
Cities are defined as Core City admin. areas; in total the City Regions include
44 other local admin. areas (identified by meta-analysis of other definitions)
Commuting in Pennine England:
Is there increasing integration between the 5 City Regions?
Between-CR commuters
1981
1991
2001
2001-1981
121,150
158,610
222,141
+100,991
3.0
4.1
5.4
+2.4
12.5
15.0
17.6
+5.1
Between-CR as % of:
All workers
Inter-zonal commuters
People commuting between the 5 City Regions are becoming a:
(1) larger share of all those who live & work in Pennine England
(2) larger share of those commuting between any of the 49 zone pairs
Modelling is then used to see whether the deterrence effect of CR
boundaries has reduced (results not shown here)
Migration for Larger Cities 1:
27 cities ranked by in/out ratio for high-status workers
27 cities ranked by in/out ratios for migration beyond the
City Region: Higher Managerial & Professional MGRPs
in/out ratio
0.4
London
Brighton
Derby
Northampton
Edinburgh
Reading
Manchester
Bristol
Leeds
Norwich
Southampton
Preston
Portsmouth
Bradford
Birmingham
Glasgow
Plymouth
Cardiff
Newcastle
Liverpool
Leicester
Middlesbrou
Nottingham
Hull
Sheffield
Stoke
Coventry
0.6
0.8
1.0
1.2
1.4
1.6
Migration for Larger Cities 2:
In/out ratio for exchanges with London by skill group
PUAs of 8 Core Cities: in/out ratio migration exchanges with London PUA,
2000-2001, for classified MGRPs by NS-SeC grouping at Census
1.2
Higher M&P
Lower M&P
Intermediate
Low Skill
1.0
in/out ratio
0.8
0.6
0.4
0.2
0.0
Birmingham Manchester
Liverpool
Newcastle
Nottingham
Sheffield
Leeds
Bristol
Examples from the ONS-LS
The ONS Longitudinal Study was set up in the 1970s
mainly for relating death events and notifiable diseases
to Census-derived socio-demographic characteristics
It comprises a ca-1% sample of people’s anonymised
records linked across the 1971 & following censuses,
allowing tracking of people’s social & spatial mobility
Extremely powerful source for following people’s life
courses and relating their several ‘biographies’ to each
other, e.g. re family/household, occupation, whereabouts
Examples: (1) Working career progression, e.g. chance of
rising from White Collar Non-core to Core 1991-2001, by
place of residence; (2) Length of stay in the ‘escalator
region’ of SE England after migrating there 1966-71 …
Example from ONS-LS 1 (a):
Probability of WC Non-core 1991 becoming WC Core 2001:
all stayers of 10 City Regions (ranked)
x
Probability
(%)
6.0
8.0
10.0
12.0
14.0
16.0
London
Manchester
Leicester
Sheffield
Leeds
Birmingham
London = 15.8
Nottingham
Weighted average for 9
non-London CRs = 12.2
Bristol
All zone stayers in
England & Wales = 13.1
Newcastle
Liverpool
Liverpool is 68% of London
18.0
Example from ONS-LS 1 (b):
Probability of WC Non-core starters becoming WC Core by
end of decade (out of all those still in work, stayers only)
16.0
London
9 other CRs
Rest of E&W
14.0
Probability (%)
12.0
10.0
8.0
6.0
4.0
2.0
0.0
1971-1981
1981-1991
1991-2001
Example from ONS-LS 2:
Movements between Rest of England & Wales (REW) and
South East (SE) of the 1966-71 REW-to-SE migration cohort
Period
In REW at
start of
period
From
REW to
SE
From SE
to REW
In REW at
end of
period
In SE at end of
period
%
1966-1971
3136
3136
n/a
0
3136
100.0
1971-1981
0
n/a
1098
1098
2038
65.0
1981-1991
1098
104
287
1281
1855
59.2
1991-2001
1281
61
195
1415
1721
54.9
The ‘escalator region’ hypothesis has it that people moving to the South East early in
their working lives (most of the 3,136) will stay there for most of their careers to
benefit from its better prospects, but many had departed within 15 years of arriving.
Plans for 2011 Census analysis 1
Once again, merely a personal research agenda, which will no
doubt be much multiplied across the academic community
Plans include:
• Using the Area Tables to monitor the ‘state of British cities’,
building on the demographic and economic analyses carried
out for the English SOCR
• Using the ODS to update and extend the previous work on
the changing nature of polycentric urban regions
• Using the ONS LS to probe further the relationships between
social and spatial mobility, with a particular focus on the
experience of immigrants (by year of arrival in UK)
• Using the Area Tables and SARs to explore the results
flowing from other ‘new’ questions on, e.g., visitors, second
addresses and alternative population bases
Plans for 2011 Census analysis 2
• Using SARs and LSs to test for post-1970 change in withinUK migration rates by population sub-group
• Responding to opportunities for providing info to central and
local government, etc., for planning & policy purposes
• Trying to ensure as much benefit as possible from the
2011 Census outputs to help make the case for a 2021
Census