secondary analysis research on ethnicity using government data & sars reza afkhami esds...
TRANSCRIPT
Secondary Analysis
Research on Ethnicity
Using Government Data & SARs
Reza Afkhami
ESDS Government & SARs
1st November University of Bristol
Outline
• Issues and Problems in Secondary Analysis of Ethnic Minority Data
• Some Research Exemplars
• Comparability of 1991/2001 SARs ethnicity Classifications
• ESDS Resources for ethnicity
Issues in Secondary Analysis of Ethnic Minority Data (1)
• Sample size • Boosted samples and weighting • Aggregation of surveys
• Country of birth • Geography
Issues in Secondary Analysis of Ethnic Minority Data (2)
• Change of question wording • Consistency over time
• Mixed and other Categories• Age
Age profile of ethnic groups Source: Vivienne Avery
Distribution of Ethnic Minority Groups by Age, 2001/2% Under 16 % 16-64 % 65 and over
White 19 65 16Black Caribbean 25 67 9Black African 33 66 2Indian 22 71 6Pakistani 35 61 4Bangladeshi 38 58 3
Annual Local Area Labour Force Survey, 2001/2, ONS
Research Exemplars
• Labour Market• Pension• Identity• Health• Women• Income• Household formation
Research Exemplars: Labour Market
Questions• How is the role of human capital
(education) in the labour market attainment for ethnic minorities?
• How is the ethnic minority disadvantage in the labour market?
Unemployment rates in UK by ethnic group and sex, ALALFS 2001/02
Source: Vivienne Avery
0% 5% 10% 15% 20% 25%
White
Mixed
Indian
Pakistani
Bangladeshi
Other Asian
Black Caribbean
Black African
Other
Men
Women
Research Exemplars: Pension
Questions:• How do private pension prospects vary among
ethnic groups?
• Is women’s disadvantage the same in each ethnic group?
• What explains minority ethnic disadvantage in pensions?
Percentage contributing to a private pension by ethnic group. Men and women 20-59
Source: Jay Ginn Family Resource Survey 1994-96
0 10 20 30 40 50 60 70
White
Indian
Black
Chinese/oth
Pakistani
Bangladeshi
Men Women
Research Exemplars: Economic Activity for Women
Source: Dale, et al
0102030405060708090
100
Eco
no
mic
Act
ivit
y Not seeking, would notlike to work
Not seeking, would like towork
Seeking unavailable
ILO Unemployed
Employed or self-employed
Economic activity and employment
women aged 19-60 Spring QLFS, 2000-2, excl. full-time students
01020304050607080
UnemplPart-timeFull-time
Source: Angela Dale and Jo Lindley
Economic activity by degree level qualifications
women 19-60, excl.FT students, 2000-2
0102030405060708090
100
White BlackCar/Oth
BlackAfrican
Indian P & B
no degreedegree
Key points about ethnic minority women
– Unemployment particularly high amongst minority ethnic women
– Part-time working highest for white women
– Black Caribbean women have highest levels of economic activity and highest levels of full-time working
– Pakistani and Bangladeshi women have lowest levels of economic activity
Research Exemplars: National Identity
• Meaning of ‘British’ in LFS questionnaire– Pertaining to UK state– Pertaining to mainland British state (excl Ireland)– Pertaining to territorial area of Britain– Pertaining to nation inclusive of others– Pertaining to nation exclusive of others– White– Pertaining to ‘cultural’ or other background qualifying
group defined by skin colour
Percentage reporting British National Identity by ethnic group LFS
N
Pakistani 69.4 447
Bangladeshi 67.0 176
Black Caribbean 66.4 440
Indian 65.0 758
Other Black 61.5 39
White and Asian 56.5 69
Other Mixed 48.9 45
White and Black Caribbean 48.3 89
White and Black African 45.2 31
Chinese 40.9 154
Other Asian 39.7 224
Black African 38.4 359
British 36.9 40396
Other 30.6 333
Other White 10.2 1747
Source: LFS 2000Q1 – 2004Q1. Excl. Northern Ireland. Ns are weighted John MacInnes
Research Exemplars: Health
• To compare the percentage of obese or overweight children within different ethnic and socio-economic groups
Interrelationships
Diabetes
Cardiovascular disease
Black Caribbean
S. Asian ethnicity
Obesity
Risk of obesity in boys by ethnic group Source: Saxena, et al.
Risk of obesity in girls by ethnic group Source: Saxena, et al.
Reported fair or bad health by ethnic group and age
0%
20%
40%
60%
80%
100%
2-5 5-9 10-12 13-15 16-20 21-25 26-30 31-35 36-40 41-45 46-50 51-60 61-70 > 70
Caribbean Indian Pakistani Bangladeshi Chinese White minority White English
Source: James Nazroo Health Survey for England 1999
Ethnic differences in equivalised household income
48% 45%
69%
90%
41%
27% 31%
0%
20%
40%
60%
80%
100%
Caribbean Indian Pakistani Bangladeshi Chinese Whiteminority
WhiteEnglish
Bottom tertile Middle tertile Top tertile
Source: James Nazroo Health Survey for England 1999 Health Survey for England 1999
How household formation varies between ethnic groups and LADs ?
Tajar & Simpson (2005)
• Data: CAMS (The Controlled Access Microdata Sample). This is a 3% individual sample from the UK 2001 census data.
• One of the most common ways of summarising household patterns is by the use of headship rates – the proportion of those in a particular age, sex or other group who are designated as ‘head of household’.
Headship rate: England and Wales, 2001
0
0.2
0.4
0.6
0.8
1
15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75+
Male
Female
Headship rates for England and Wales - 2001 All Households: Male
00.10.20.30.40.50.60.70.80.9
1
15-19 20-24 25-29 30-44 45-59 60-74 75+
Age
White
Caribbean
Pakistani
Indian
Bangladeshi
Headship rates for England and Wales - 2001 All Households: Female
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
15-19 20-24 25-29 30-44 45-59 60-74 75+
Age
Caribbean
White
Pakistani
Bangladeshi
Indian
Geographical effect
Ethnic groups Variance of the LAD effectWhite 0.032Indian 0.151
Pakistani 0.064Bangladeshi 0.036Caribbean 0.054
African 0.044Other Black 0.044
Chinese 0.129Other 0.083
Significant values in Bold
Conclusion
o Higher headship rate of Caribbean women in their thirties and forties compared to other ethnic groups for England and Wales and also for LADs: high number of lone parent Caribbean women.
o There are geographical variations within ethnic groups.
Ethnic group Comparison in 1991/2001 Census
• Ethnic group questions in 1991 and 2001
• Changes to ethnic identification
• Recommendations for comparing 1991 and 2001 Census ethnic groups
Ethnic group questions in 1991 and 2001
• Ethnicity question asked in 1991 census for England, Wales and Scotland
• All UK countries included ethnic question in 2001 census but questions in Scotland and Northern Ireland each differed from E&W
The 1991 Census ethnic group question asked in England, Wales and Scotland
The 2001 Census ethnic group question asked in England and Wales
Changes to ethnic identification
• The most stable categories; White, Indian, Pakistani, Bangladeshi and Chinese.
• Less stable categories; the Black African and Black Caribbean groups
• The least stable; the ‘Other’ ethnic groups
Recommendations for comparing 1991 and 2001 Census ethnic
groups
• a ten-category classification; an eight category classification; a five-category classification; and a two-category classification.
• More groups combined; greater stability at the cost of losing the detail
• Balance between reliability and ethnic group detail depends on the specific analysis
Ten category classification
Eight category classification
Five category classification
Two category classification
ESDS Resources on Ethnicity
– LFS-Ethnicity Consistent over time
– GHS-Ethnicity Consistent over time
– Ethnicity Theme Page