26
Developing Alternatives to Census General Register Office for SCOTLAND information about Scotland’s people Finding the Population Statistics You Need, Glasgow 23 February 2010 Alternative Sources branch, GROS

National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

  • Upload
    others

  • View
    0

  • Download
    0

Embed Size (px)

Citation preview

Page 1: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Developing Alternatives to Census

General Register Officefor

SCOTLANDinformation about Scotland’s

people

Finding the Population Statistics You Need, Glasgow 23 February 2010

Alternative Sources branch, GROS

Page 2: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Overview

• Background• Beyond 2011• Development Work in Scotland

Page 3: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Background 1

• Key outputs– Census: decennial count, rich detail at small area

level– Mid-year population estimates: multiple data sources

• Challenges– Social change: migration– User needs– Traditional census difficulties: response, cost– Data availability: administrative sources

Page 4: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Background 2• Short and medium term response

– Migration Statistics Task Force (2006)– Improving Migration and Population Statistics

Programme (2007)

• Key developments– Increased Use of Administrative data – Statistics and Registration Services Act 2007(Office for National Statistics (ONS): English School

Census, Higher Education (HE) Student Record, Department of Work and Pensions (DWP) Benefit Recipients Records)

Page 5: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Beyond 2011 Project 1

• Longer term framework• Statistics User Forum, May 2009 • Scope: UK-wide, ONS-led, England and Wales

development work • General Register Office for Scotland

involvement: Steering Group and Project Board

• http://www.ons.gov.uk/about-statistics/methodology-and- quality/imps/beyond-2011/index.html

Page 6: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Beyond 2011 Project 2Aims:• To investigate the feasibility of improving

population statistics in the UK by making use of integrated data sources to replace or complement existing approaches

• To investigate whether alternative data sources can provide the priority statistics on the characteristics of small populations, typically provided by census

Page 7: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Beyond 2011 Project Structure

Page 8: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Beyond 2011 Project Structure

Work Stream 1: User Requirements

• What outputs are needed: stocks and flows, individuals/households, dwellingscounts and structures, characteristics

• Quality Aspects: relevance, frequency , timeliness, accuracy/geography, accessibility/clarity, comparability, coherence

Page 9: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Beyond 2011 Project Structure

Work Stream 2: Administrative Data

• Identify Key Sources: DWP/Her Majesty’s Revenue and Customs (HMRC), National Health Service Central Register (NHSCR)

• Demonstrate Feasibility: proof of concept

Page 10: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Beyond 2011 Project Structure

Work Stream 3: Survey/Census Design

• Extended and integrated household surveys

• Alternative census designs – e.g. French (rolling census) or United States (US) (short and long form) models

• Modelling: Small area estimation

Page 11: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Beyond 2011 Project Structure

Work Stream 4: Enablers/Barriers Key developments underpinning the work

• IT infrastructure

• Data security/confidentiality

• Legal issues

• Harmonisation of definitions, concepts, identifiers

• Public acceptability http://www.ons.gov.uk/about- statistics/methodology-and-quality/imps/beyond-2011/data-sharing-between- government-departments---report-on-public-acceptability.pdf

Page 12: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Beyond 2011 Project Stages

• March 2010 – HMT Strategic Case and work plan

• Spring 2011 – Census 2011 and key sources snapshots

• 2011 – 2013 – Analysis, assessment

• 2013/14 – Decisions on a future approach

Page 13: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Work in Scotland

• GROS focus on administrative sources :

Small team to cover this area since October 2004

• SG-led harmonisation and integration of population surveys:

Long term strategy for population surveys 2009-2019http://www.scotland.gov.uk/Topics/Statistics/About/SurvStrat

Page 14: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

GROS: Administrative Sources Research 1

• Aiming to arrive at a coherent UK approach

but

• Development work for Scotland where

• Sources, or

• Legal framework for data sharing may be different

(e.g. No Scottish equivalent of the 2007 SRSA which gives ONS enabling powers to access OGD data; easier access to NHSCR data)

Page 15: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

GROS: Administrative Sources Research 2

• Data sharing

• Record linkage capability

• Data quality assessment

• Data integration

Page 16: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Demonstrating feasibility: data integration 1

Page 17: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Demonstrating feasibility: data integration 2

Coverage Survey

Census

Coverage Survey

Census

Administrative records census

3rdlist: TSE

QA

Imputation

Demographic SpineUpdates/

MaintenanceStatistical Population

Register?

Surveys

Page 18: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Data Sharing• Key requirement: person record level

• Currently available:

GROS held: Census, Registration, NHSCR

Other: Electoral Roll

• Work in progress (short to medium term):

With ONS: NINO (National Insurance Number) Allocations (Migrant Worker Scan), HE Student record (E&W), DWP benefit recipients

With Scottish Government: HE student record for Scotland

With Local Authorities (LA): School census and other local data

Page 19: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Data Quality Assessment

• Small area aggregate data (comparison to Census and population and migration estimates):

child benefit, state-pension-age benefits, school census (stocks and flows), electoral roll, NHSCR, HE student record (flows)

• Person level data (record linkage)

community health index, electoral roll, NHSCR

Page 20: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Record Linkage Capability

• Reviewed tools and methodologies for combining data

• Identified key software tools

• Carried out pilot studies

Page 21: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Key Conclusions (so far)• Different sources tell different stories/measure

different things (and patterns vary across the country)

• Aggregate data useful as a start but of limited value: raises more questions

• Access to identified person records essential to progress– comprehensive population coverage– ability to cross check sources and resolve

discrepancies (which are numerous)

Page 22: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

NHS Central Register Compared to Population Estimates: Percent Over-Coverage by Gender and

Age, Scotland 2008-2009

16%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

Age

Perc

enta

ge

AverageMenWomen

Page 23: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Over-Coverage of NHSCR against Mid Year Estimates (MYE),

9%

16%

-5%

0%

5%

10%

15%

20%

25%

Aberdeen C

ity

Edinburgh, C

ity of

Glasgow

City

Stirling

Inverclyde

Renfrew

shire

West D

unbartonshire

South A

yrshire

East D

unbartonshire

North A

yrshire

North Lanarkshire

Aberdeenshire

South Lanarkshire

East R

enfrewshire

Midlothian

Argyll &

Bute

Highland

Fife

East A

yrshire

West Lothian

Dum

fries & G

alloway

East Lothian

Perth &

Kinross

Eilean S

iar

Clackm

annanshire

Moray

Orkney Islands

Falkirk

Dundee C

ity

Angus

Scottish B

orders

Shetland Islands

Local Authorties

Perc

enta

ge (%

)

Local Authorities Scotland who are assigned valid postcode Scotland - alive and residing in a scottish Health Board

by Local Authority,2008

Page 24: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Proportion of Census Records Linked to Community Health Index (CHI) by Age and Gender, 2001

84%

86%

88%

90%

92%

94%

96%

98%

00-0

4

05-

09

10-

14

15-1

9

20-2

4

25-2

9

30-3

4

35-3

9

40-4

4

45-4

9

50-5

4

55-5

9

60-6

4

65-6

9

70-7

4

75-7

9

80-8

4

85-8

9

90+

A ge Group

Men Women

Page 25: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Key Challenges

• Data sharing: Insufficient population coverage secured so far

• IT infrastructures and tools

Page 26: National Archives of Scotland - General Register Office for … · 2010. 3. 12. · Over-Coverage of NHSCR against Mid Year Estimates (MYE), 9% 16%-5% 0% 5% 10% 15% 20% 25%. Aberdeen

Timeline

• Comparison of our ‘statistical register’ with 2011 Census– Much of the work to be done retrospectively

• Inform decisions for 2021 around 2013/14– The quality of conclusions depends on future

outcomes of data sharing initiatives