Transcript Labour Market Positions of Minority Ethnic Groups in

Origin, education and destination
– an analysis of social mobility in Britain (1991-2005)
For presentation at
Education and Employers Taskforce research conference 2011
exploring social mobility and delivery mechanisms in international
perspective
12 October: University of Warwick
Yaojun Li and Fiona Devine
School of Social Sciences
Manchester University, UK
[email protected]
[email protected]
Aims of this presentation
 To explore patterns of intergenerational
educational and class mobility in Britain
1991 - 2005;
 To assess the extent of class differences in
the two areas and the direction of change;
 To bridge the differences in academic and
political understanding of mobility;
 To see where the government and society
could do better to improve social equality
and social inclusion
2
Why should we be concerned?
• A fair society is the life-blood of democracy and
our shared future;
• A fair society has equal life chances for all
regardless of family origin, gender and other
ascribed factors;
• A fair society is characterised by meritocracy
(ability and effort) but ascribed meritocracy may
be a better description of social reality?
• How much inequality is there in our society, and
• In what direction is it changing?
3
An ideal scenario
Education
Origin
Destination
Notes
1. The dotted lines indicate weakening association and the solid line indicates
strengthening association.
4
Why shall we be concerned with
OED
• Because family classes play a very
important role in E and D
• We might not be able to change E-D very
much but changing the O-E relations, we
might hopefully bring about more mobility
in E and then in D
5
Debates in mobility research
• Academics talk about absolute and relative mobility but
considerable differences among themselves
• Policy-makers, the media and the general public tend to
view mobility as upward mobility only
• A truly fair and equal society includes both upward
mobility for working class children (where policy can
have a difference) and downward mobility for middleclass children (where policy cannot do much)
• Greater upward mobility will increase social fluidity even
though downward mobility remains at a similar level
6
Competing theses of mobility
• Constant social fluidity (Goldthorpe 1987;
Goldthorpe and Mills 2004, 2008; Goldthorpe and
Jackson 2007)
t1
tn
• Declining social mobility (Blanden et al 2005)
t1
tn
• Signs of increasing social fluidity (Heath and
Payne 2000; Lambert et al 2007; Li and Devine
2011)
t1
tn
7
Why the differences
• Economists use family income data put in quantiles
and thus relativise the analysis from the start,
unable to differentiate absolute and relative mobility
• Sociologists use class-based models with attention
to both absolute and relative mobility (but with the
main focus on relative mobility expressed in odds
ratios, however, those models do not allow multiple
continuous covariates to be included--loglinear and
log multiplicative layer-effects or UNIDIFF)
• Data consistency also affects conclusions reached
among sociologists
8
Data for this analysis
 The British Household Panel Survey (BHPS)
1991
 The General Household Survey (GHS) 2005
 These are the best data sources currently available:
 Nationally representative social surveys with large samples
 Consistent coding for father’s, mother’s and respondent’s
class in National Statistics Social-economic Classification
(NS-SeC) allowing for theoretically-guided recoding:
dominance approach to parental class and putting lowergrade routine-non-manual (IIIb) to unskilled manual routine
class (VIIab) for both parental and respondent’s classes
 BHPS has parental and respondent’s class for all aged 16+
 GHS has only parental class data for respondents aged 25-65
 We therefore use data for men 25-65 and women 25-63
The O-E-D links
Education
Origin
Destination
Notes
1. The dotted lines indicate weakening association and the solid line indicates
strengthening association.
10
The O-E link
1991
Men
Higher salariat
Lower salariat
Intermediate
Small employer/own account
Lower supervisory/technical
Semi-routine
Routine
All
N
Tert
Sec
Prim
68
49
49
36
28
21
18
34
992
23
33
30
27
32
28
27
29
871
10
18
21
36
40
51
54
37
1,063
11
The O-E link
2005
Men
Higher salariat
Lower salariat
Intermediate
Small employer/own account
Lower supervisory/technical
Semi-routine
Routine
All
N
Tert
Sec
Prim
64
47
38
27
26
22
21
35
1,510
29
38
48
41
43
46
40
40
1,762
7
15
14
33
30
32
39
24
1,054
12
The O-E link
1991
Women
Higher salariat
Lower salariat
Intermediate
Small employer/own account
Lower supervisory/technical
Semi-routine
Routine
All
N
Tert
Sec
Prim
61
43
32
21
19
14
13
25
803
26
33
38
34
31
30
24
31
980
13
24
30
45
49
56
63
44
1,351
13
The O-E link
2005
Women
Higher salariat
Lower salariat
Intermediate
Small employer/own account
Lower supervisory/technical
Semi-routine
Routine
All
N
Tert
Sec
Prim
64
49
41
27
23
22
19
35
1,633
29
37
46
43
42
42
39
39
1,865
7
14
13
30
35
36
41
26
1,216
14
Summary of the O-E link
• Great class differences in education for both
men and women
• There are also signs of class differences on the
decline. Class I and VII differences in tertiary
education fell from 50 to 43 percentage points
(68% v 18%; 64% v 21%) for men, and from 48
to 45 points for women (61% v 13%; 64% v
19%)
• Reduced gender differences (34% to 35% for
men; 25% to 35% for women)
15
The O-D link
1991
Salariat Intermediate
Men
Higher salariat
Lower salariat
Intermediate
Small employer/own account
Lower supervisory/technical
Semi-routine
Routine
All
N
70
56
50
33
32
25
23
37
1,089
19
28
30
44
35
34
38
34
1,009
Working class
11
15
20
23
32
41
39
29
828
16
The O-D link
Men
Higher salariat
Lower salariat
Intermediate
Small employer/own account
Lower supervisory/technical
Semi-routine
Routine
All
N
Salariat
2005
Intermediate
Working class
68
54
52
36
38
33
31
44
1,948
18
26
29
36
32
32
32
29
1,247
14
20
19
28
30
35
37
27
1,131
17
The O-D link
1991
Salariat Intermediate
Women
Higher salariat
Lower salariat
Intermediate
Small employer/own account
Lower supervisory/technical
Semi-routine
Routine
All
N
61
42
31
23
24
19
20
29
901
25
32
40
37
35
28
28
32
1,011
Working class
14
26
29
40
41
53
52
39
1,222
18
The O-D link
Women
Higher salariat
Lower salariat
Intermediate
Small employer/own account
Lower supervisory/technical
Semi-routine
Routine
All
N
Salariat
2005
Intermediate
Working class
68
54
52
36
38
33
31
44
1,948
18
26
29
36
32
32
32
29
1,247
14
20
19
28
30
35
37
27
1,131
19
Summary of the O-D link
• Again great class differences for both men and
women, in class attainment
• There are also signs of class differences on the
decline. Class I and VII differences in access to
salariat positions fell from 47 to 37 percentage
points (70% v 23%; 68% v 31%) for men, and
from 41 to 31 points for women (61% v 20%;
56% v 25%)
• Reduced gender differences (37% to 44% for
men; 29% to 39% for women)
20
The E-D link
1991
Salariat Intermediate
Men
Degree+
Sub-degree
A Levels
O Levels or equivalent
Primary
No qualifications
All
89
58
44
31
11
9
37
9
29
41
38
47
40
34
Working class
2
13
15
32
41
51
29
21
The E-D link
2005
Salariat Intermediate
Men
Degree+
Sub-degree
A Levels
O Levels or equivalent
Primary
No qualifications
All
81
64
44
30
23
10
44
12
24
37
38
33
35
29
Working class
7
12
19
32
44
56
27
22
The E-D link
1991
Salariat Intermediate
Women
Degree+
Sub-degree
A Levels
O Levels or equivalent
Primary
No qualifications
(All)
84
63
26
21
11
8
29
11
22
48
45
40
27
32
Working class
5
15
26
34
49
65
39
23
The E-D link
2005
Salariat Intermediate
Women
Degree+
Sub-degree
A Levels
O Levels or equivalent
Primary
No qualifications
(All)
80
62
34
24
18
8
39
12
20
33
31
36
21
24
Working class
8
18
33
46
45
71
37
24
Summary of the E-D link
• Expected educational differences for both men
and women in class attainment
• There are also signs of educational differences
on the decline. Differences between degreeholders and those with no qualifications in
access to salariat positions fell from 80 to 71
percentage points for men, and from 76 to 72
points for women
25
Figure 2
Changes in the overall O-E-D associations (1991-2005)
Men
E
-4.1%
Women
E
-2.4%
-1.8%
O
D
-2.3%
-1.7%
O
D
-1.6%
26
Summary of trends in
relative mobilities
• Weakening O-E and O-D links for men
and women
• Unchanging E-D links
• The meritocratic ideal is not substantiated
27
Which part do the changes
occur?
• Men
 O-E: Classes II-IV
 O-D: Class II
 E-D: Degree holders
 O-E-D: Classes I-II, Degree-holder, but
non-significant
• Women
O-D: Class I;
E-D: Degree holder
28
An example of O-E
Men 1991
Higher salariat
Lower salariat
Intermediate
Small employer
Supervisory/technical
Semi-routine
Routine
Degree
Subdegree
A
O Primary
41
22
20
11
5
5
2
27
28
29
25
23
17
16
11
16
14
11
12
8
9
12
17
15
16
19
20
18
4
6
8
7
11
12
12
None
6
11
13
29
29
39
43
29
An example of O-E
Men 1991
Higher salariat
Lower salariat
Intermediate
Small employer
Supervisory/technical
Semi-routine
Routine
Degree
Subdegree
A
O Primary
55
36
28
17
16
13
11
9
11
11
10
10
9
10
14
17
20
16
16
16
17
14
21
28
24
27
30
23
3
5
5
8
9
9
7
None
4
10
9
25
21
23
32
30
Conclusion
• Great origin class differences in people’s
education and occupational attainment
• The origin effects have reduced slightly, giving
some hopes that progress could be made
• Government can do more to help the most
disadvantaged to have better education by
reducing fees, getting rid of NEETs
• Employers could provide or fund more on-job
training to help the poorly-educated to have
better skills, to give them a second chance
31
Many thanks and suggestions
warmly welcome
[email protected]
[email protected]
32
33
Loglinear and UNIDIFF Models
1 Baseline model (conditional independence)
logFijk = µ + λiO + λjD + λkY + λikOY + λjkDY
2 Constant Social Fluidity model (CSF)
logFijk = µ +
O
λi
+
D
λj
+ λk + λik
Y
OY
+ λjk +
DY
OD
λij
3 Log multiplicative or uniform difference (Unidiff) model
logFijk = µ + λiO + λjD + λkY + λikOY + λjkDY + λijOD+ βkXij
34
How to account for the lack of
progress in social mobility?
–AR A T
for a S A
ational
ocial
ction
heory
dvantage and
explanation
D
story
isadvantage
‘In terms of individuals of differing class origins
pursuing ‘mobility strategies’ that, while rationally
adaptive to the constraints typical of their class
situations, tend in their aggregate outcome to
maintain relative rates unaltered, at all events in
the absence of any external modification of these
constraints that would constitute a reduction – or
an increase – in class-lined inequalities of
condition’ (Goldthorpe and Mills, 2004: 223). 35
36