Samples of Anonymised Records a resource for ethnicity

Samples of Anonymised Records: a resource for ethnicity research Ed Fieldhouse Director, SARs Support team http: //www. ccsr. ac. uk/sars

Key features of the SARs • • Coverage Full range of census variables Size Ethnicity and religion Geography Multivariate/flexibility Household structure Comparability with other census products • Easy access

1991 Census microdata File Sample type 1991 2% sample Individual of individuals SAR Geography Availability GB and NI available separately. Online registration and access via CCSR. Divided into a total of 288 SAR areas 1991 Househol d SAR 1% sample of households GB and NI available separately. Regional geography Data can be downloaded in SPSS, tab or Stata. Nesstar tool available for online data exploration.

The SARs family 2001 File Sample type Geography Availability Individual 3% sample of licenced individuals UK EUL GOR (+ Wales, Scot, CCSR NI, Inner/Outer London) Small area 5% sample of microdata individuals UK: LA (or consituency in NI) EUL CCSR Household licensed 1% hierarchical file None: England & Wales only Special licence UKDA Individual CAMS Same sample as Individual licenced SAR LA (GB) or Constituency (NI) IMD info for SOA In house at ONS Household CAMS 1% hierarchical file All of UK In house at ONS

Individual SAR • • • 3% sample of individuals from UK 1. 84 million records All census variables present Lowest geography – GOR Access – End User Licence (via Athens for academics – Do not attempt to identify anyone – Do not pass on data to unregistered individual

Controlled Access Individual SAR • 3% sample of individuals from UK • 1. 84 million records • All census variables present with very great detail • Lowest geography – LA • Access – Only within 4 ONS offices – Rigorous application procedure by ONS – Careful vetting of outputs by ONS

Household SAR • • • 1% sample of households, E & W only 225 K Households; 525 K individuals All census variables present No geography Individuals linked within household • Access – Special Licence - administered by UKDA – More restrictive than EUL, eg cannot use on laptop – Applications need approval by ONS

Controlled Access Household SAR • Same sample as SL-Household SAR, but also contains Scotland NI • All census variables present with very great detail • Lowest geography – LA • Access – Only within 4 ONS offices – Rigorous application procedure by ONS – Careful vetting of outputs by ONS

Small Area Microdata (SAM) • 5% sample of Individuals from UK • 2. 96 million records • Most census variables present – restricted detail • Lowest geography – LA • Access – End User Licence (via Athens for academics) – Do not attempt to identify anyone – Do not pass on data to unregistered individual

The 1991 Census ethnic group question asked in England, Wales and Scotland

The 2001 Census ethnic group question asked in England Wales

The 2001 Census ethnic group question asked in Scotland

The 2001 Census ethnic group question asked in Northern Ireland

2007 Test – England Wales

File No of ethnic group categories Country of birth Religion 1991 Individual SAR 10 42 No Individual licensed 2001 16 England Wales 14 Scotland 2 Northern Ireland 16 9 E/W 11 Scotland 7 NI Small area microdata 13 E/W 8 Scotland 2 NI 5 9 E/W 10 Scotland 7 NI 1991 Household SAR 10 42 No SL Household licensed 16 E/W 16 9 E/W Individual CAMS 16 E/W 14 Scotland 12 NI 49 9 E/W 10 Scotland 7 NI Household CAMS 16 E/W 49 9 E/W 10 Scotland 7 NI

1991 2% Individual SAR

Ethnic Group for England Wales – 2001 3% Individual Licensed SAR

SAM: Sample size by ethnicity (England Wales)

SAM: Northern Ireland

SAM in Scotland

Comparison of ethnic group categories in 1991 -2001 Census From Simpson and Akinwale, 2006

Ethnic group imputation in 2001

Comparing with the 100% data: England

Comparing with the 100% data: Scotland

Key research areas on ethincity using SARs • Ethnic differences in unemployment • Ethnic differences in educational attainment and participation • Ethnic differences in class attainment • Ethnic composition of families • Ethnic differences in health

Research findings: 1991 SARs: Unemployment • In contrast to the small sample size of survey data which forces many researchers to over-collapse the categories and call minority ethnic groups 'black', the SARs allows for detailed analysis of ethnic groups. • Differences both between and within major ethnic groups can be explored in various aspects of their socio-economic lives while at the same time controlling for other important characteristics. • Blackburn, Dale and Jarman (1997) showed striking differences between ethnic groups in the vulnerability to unemployment, even among people with the same level of educational qualifications. One in five (20 per cent) of UK-born Black. African men and women with higher qualifications were unemployed, but the rate for similarly qualified UK-born Whites was only one fifth as many (3 to 4 per cent). This is a case of what might be called 'ethnic penalty' • See also Fieldhouse and Gould, 1998 on how ethnic penalties are affected by local labour market conditions using subregional geography in SAR • Similar work by Simpson et al (for DWP) and Heath et al using 2001 SARs

Mixed couples – SL-HSAR

. . . and UK born

Impact of age, sex, qualifications and country of birth on economic activity (source DWP Research Report 333)

Impact of age, sex, qualifications and country of birth on economic activity (source DWP Research Report 333

‘Ethnic penalties’ on women's economic activity (source DWP Research Report 333

‘Ethnic penalties’ on men’s unemployment (source DWP Research Report 333

Accessing the files • EUL files available online following standard registration with the Census Registration System – Licensed Individual file 2001 – Small Area Microdata file 2001 – 1991 1% household and 2% individual SARs • Special licence Household SAR (via UKDA) • CAMS – controlled access allowed via ONS at specified sites