Large-scale, cross-sectional government datasets; research - - PowerPoint PPT Presentation

Large-scale, cross-sectional government datasets; research published and recent developments.

Jo Wathan Data Support Economic and Social Data Service (Government) Cathie Marsh Centre for Census and Survey Research University of Manchester UK Jo.wathan@manchester.ac.uk

http://www.esds.ac.uk/government

ESDS Government

• Economic and Social Data Service is the leading

data dissemination and support service for social data in the UK. Started Jan ‘03

• Distributed service, involving 4 organisations at 2

sites in the UK

• ESDS Government is one of 4 specialist services. It

deals with large-scale cross-sectional continuous surveys

– Academic service funded by Economic and Social Research Council – Data archiving and dissemination done by the UK Data Archive (download/Nesstar) – User support and outreach done by the Centre for Census and Survey Research – Data comes from the Office for National Statistics or National Centre for Social Research

Which surveys?

• General Household Survey
• Labour Force Survey
• Family Resources Survey
• Expenditure and Food Survey (previously the

National Food Survey and Family Expenditure Survey)

• ONS Omnibus Survey
• National Travel Survey
• Time Use Survey
• British Crime Survey/Scottish Crime Survey
• British Social Attitudes/Scottish Social

Attitudes/Northern Ireland Life & Times/Young People’s Social Attitudes

• Health Survey for England/Wales/Scotland
• Survey of English Housing (England only)

What is the data like?

• Survey microdata
• Large sample sizes (but smaller than the

SARs)

• Continuous surveys – always up-to-date
• Cross-sectional (although the LFS has a 5-

quarter panel element, GHS goes longitudinal this year)

• Specialist topic surveys – more depth than

the Census

• Freely available to academics via ESDS

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 2002(1) 2002(2) 2003(1) 2003(2) 2004(1) 2004(2)

Other FES/NFS/EFS BCS HSE GHS LFS

Increased data use: Jan 02-now (ESDS started Jan 03)

How are the data used?

To provide nationally representative results, with the flexibility of microdata:

• Analyses looking at change over time

– Repeated cross-sections – Pseudo-cohort studies

• To look at sub-populations

– Large sample sizes – Many datasets can be pooled due to relative consistency in content and method over time

• Hierarchical data for household analyses
• To provide users with the scope to
• perationalise concepts differently and

use more sophisticated modelling

SMOKING AND SOCIAL CLASS - MEN

5 10 15 20 25 30 35 40 45 1994 1995 1996 1997 1998 1999 2000 2001 year % all sc I&II sc IV&V

Source:HSE

Change over time

• Data well suited – emphasis
• n stability and comparability
• Straightforward trends; e.g.

– Smoking by class (Marmot 2003) – Smoking prevalence by month (Jarvis 2003) – Attitudes to homosexuality (Crocket and Voas 2003)

• Pseudo cohort analyses

– Women’s access to pensions by partnership (Ginn 2003) – Alcohol consumption by age (Kemm 2003) – Changing risk factors for divorce by year of marriage (Chan & Halpin 2005)

Representative samples of subpopulations

• Samples vary but can be large:

– Labour Force Survey: c.60k households – Family Resources survey: c.27k households – British Crime Survey: c. 33k individuals

• Pool to increase sample size of small subpopulations

– e.g. individual ethnic groups

– 12 years pooled of LFS to look at women’s employment by ethnic group (Dale et. Al 2005) – 3 years of FRS to look at women’s pension chances by ethnic group (Ginn and Arber 2001)

• Health Survey contains subgroup boosts

– 1999 Ethnic minority boost allow analysis of obesity by ethnic group (Saxena et. Al. 2004)

Using hierarchy to look within households

• Workless

Households (Dickens et.al, 2000)

• Educational

homogamy in Britain & Ireland (Halpin 2003)

• Impact of

parenthood/ partnership

Workless households (source FES, various years 1968-1996)

5 10 15 20 25

68 70 72 74 76 78 80 82 84 86 88 90 92 94 96

Year Percentage (of present working age hoh) workless households children in workless households

Household ID Household type Person 1 40 years Female Working Person 2 42 Male Unemployed Person 3 14 Male Student

ESDS Government Outreach and Support

• Help-desk
• User Groups
• Annual Research Conference
• Newsletter
• Vital stats GIS interface
• Interface between Government and Users
• Themed materials and training

– Training workshops and publicity events – Production of teaching datasets – Web-based materials

Workshop and Events

Introductory workshops (inc. those joint with SARs workshops) Joint workshops with

• ther ESDS services

Methods Research conferences Dataset user groups Other presentations Posters

Data - value added

• Annual teaching dataset produced

– With documentation – Based around annual theme

• Information about comparing data over

time

– Derived variables produced where necessary – User-generated derived variables QAd and made available

Resources to support users

• Thematic Guides
• FAQs & Starting Analysis

Guides

• Methodological Guides

– SPSS, Stata – Guide to weighting

• Publications database
• Biannual newsletter
• Links to other web resources

CommonGIS Interface for Vital Statistics

Help-desk

• 174 Queries in the last year
• Range of different ranges of complexity:

– How do I get access to the HSE? – What are social classes written as roman numerals? – Why do my sample sizes increase if I weight? – Should I apply weights if I’m modelling? – The sample size for matched LFS data seem wrong – The household identifiers are missing from an early LFS

• Help desk involves collaboration with UKDA

and ONS

Interface between data depositors and users

• Organise User Groups for data producers

to meet users

• Awareness and consultation role in

developments

– General Household Survey (Longitudinal) – Continuous Population Survey

• Data access e.g. Special Conditions

– Response to increased concern about confidentiality and withdrawal of some data – Negotiation of tighter conditions to permit access to data which poses greater confidentiality concerns – Access to information on date of birth, local authority district etc.