1
Social and ethnic inequalities in
choice available and choices made
at age 16
Rebecca Allen, Meenakshi
Parameshwaran and Dave
Thomson, Education Datalab
December 2016
Social Mobility Commission
Sanctuary Buildings
20 Great Smith Street
London
SW1P 3BT
2
About the Commission
The Social Mobility Commission is an advisory, non-departmental public body
established under the Life Chances Act 2010 as modified by the Welfare Reform and
Work Act 2016. It has a duty to assess progress in improving social mobility in the
United Kingdom and to promote social mobility in England. It currently consists of
four commissioners and is supported by a small secretariat.
The commission board currently comprises:
• Alan Milburn (chair)
• Baroness Gillian Shephard (deputy chair)
• Paul Gregg, Professor of Economic and Social Policy, University of Bath
• David Johnston, Chief Executive of the Social Mobility Foundation
The functions of the commission include:
• Monitoring progress on improving social mobility
• Providing published advice to ministers on matters relating to social
mobility
• Undertaking social mobility advocacy
3
Contents
Executive Summary ................................................................................................... 5
Social inequalities in post-16 routes .................................................................... 5
Gender inequalities in post-16 routes .................................................................. 6
Ethnic inequalities in post-16 routes .................................................................... 6
Post-16 outcomes where there are few school sixth form places available ........ 7
Introduction ................................................................................................................ 8
Previous research ................................................................................................... 8
Data ............................................................................................................................ 9
Institutions included in our analysis ...................................................................... 10
Describing student post-16 outcomes ................................................................... 12
Creating choice sets of institutional availability ..................................................... 14
Defining choice sets by observed choices made............................................... 14
Describing institutional availability within choice sets ........................................... 15
Findings I – Inequalities in post-16 routes where choices and opportunities are
identical .................................................................................................................... 17
Understanding differences in post-16 choices by social background ................... 17
Understanding differences in post-16 choices by gender ..................................... 21
Understanding differences in post-16 choices by ethnic background ................... 24
Findings II – Post-16 outcomes in areas with few school sixth forms ....................... 31
Matching across MSOAs ...................................................................................... 31
MSOA-level regressions ................................................................................... 39
Possible causes .................................................................................................... 41
Appendices .............................................................................................................. 43
Appendix I – Maps of post-16 institutional availability ........................................... 43
Appendix II – Year-on-year stability of MSOA choice set indicators ..................... 46
Appendix III – Descriptive statistics of choice set indicators, institutional choices,
and post-16 outcomes by pupil prior attainment tercile ......................................... 48
Appendix IV – Post-16 route and outcomes – interacting combinations of pupil
characteristics ....................................................................................................... 49
Appendix V – Post-16 routes and outcomes in London ........................................ 64
Appendix VI – Factor analysis of MSOA characteristics ....................................... 73
Appendix VII – Pupil-level matching and regressions of impact of living in an area
with no school sixth forms ..................................................................................... 74
4
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
5
Executive Summary
The transition at the age of 16 marks the first point in most individuals’ educational
lives where opportunities and choice become markedly diverse. Students can
choose whether to attend school sixth forms, sixth form colleges or further education
(FE) colleges. This choice is, in part, geographically shaped by quite idiosyncratic
variation in institutional provision. For example, students in London are able to
choose between large numbers of providers with exceptionally high access to school
sixth forms. The North West has high levels of provision in sixth form colleges, unlike
the North East and East Midlands. Students in the North East and South West have
the lowest levels of choice of provider.
This research provides an up-to-date understanding of post-16 educational choices
and transitions, highlighting the implications of differences in choice sets for students
from different areas and backgrounds. Specifically, our research analysis explores
how the choice sets available to students vary according to their geographic, social
and educational background. It investigates the impact of these institutional and
qualification choices made on students’ educational trajectories at age 16 and into
higher education.
Using three linked databases – the National Pupil Database, Individual Learner
Records, and Higher Education Statistics Authority data – we explore all educational
choices made by students who sat GCSEs in summer 2010. Our modelling strategy
allows us to distinguish between inequalities in post-16 choice and attainment that
arise between groups facing identical institutional availabilities and those that arise
because different groups have access to different types of local post-16 provision.
We do this, in part, by comparing choices made by students who live in the same
neighbourhood (Middle Super Output Area) and who achieved similar GCSEs at age
16.
Social inequalities in post-16 routes
Students eligible for free school meals (FSM) live in more urban areas, which gives
them greater choice of post-16 institution and marginally greater access to 6
th
form
colleges. We find that FSM pupils living in the same neighbourhood and with similar
GCSE attainment as non-FSM pupils have marked differences in the post-16
choices they make that cannot be explained by inequalities in institutional availability.
In this cohort, 9% of non-FSM students drop-out of the education system at age 16
compared to 16% of FSM students. This is, of course, largely related to GCSE
attainment. But we estimate over a third of this participation gap results from
differences in choices made by students with identical opportunities open to them.
We estimate these choices are also responsible for about a third of the gap in entry
to a sixth form environment (i.e. school sixth form, sixth form college or independent
school).
There are FSM gaps in the proportions studying at least level 2 qualifications (58%
versus 78%), level 3 qualifications (40% versus 66%) and 3 A levels (47% versus
21%). We estimate that 24%, 26% and 35% of these gaps, respectively, arises from
choices made by students living in the same neighbourhoods with approximately the
same GCSE attainment.
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
6
There are also marked FSM differences in higher education routes. Just 24% of FSM
pupils attend HE versus 42% of non-FSM pupils, with over a quarter of this
participation gap arising from students within the same neighbourhood with the same
GCSE attainment. We estimate that as much as half the FSM gap in Russell Group
attendance (at 2% for FSM versus 10% for non-FSM) could be removed if we were
able to equalise university access between pupils living in the same neighbourhoods
and with similar GCSE attainment.
Gender inequalities in post-16 routes
Male and female pupils generally face similar choice sets but take slightly different
post-16 routes as a result of both differences in age 16 achievement and choices
made by students with the same achievement levels. Boys are more likely to drop
out of education altogether at age 16 (9% versus 11%), but about half of this
participation gap arises from choices by gender of students with the same GCSE
attainment.
Girls are more likely to take at least level 2 qualifications (81% versus 70%), level 3
qualifications (68% versus 58%) and 3 A levels (48% versus 39%). But while
differences in rates studying for at least level 2 qualifications are largely explained by
choice, the differences in rates studying A levels are almost all explained by gender
differences in GCSE attainment.
Girls are more likely than boys to attend HE (44% versus 36%) both because they
have higher GCSE attainment and because they have higher attendance for a given
level of GCSE attainment. The latter contributes about a third of this gender gap. The
pattern for elite Russell Group attendance is quite different because, although girls
are slightly more likely to attend over (10% versus 9%), they are less likely (with 17%
lower odds) than boys to attend for a given level of GCSE attainment.
Ethnic inequalities in post-16 routes
Students of minority ethnic backgrounds – measured here using six broad ethnic
groupings – have lower propensities to drop out of education at age 16, compared to
the White British population (for whom 10% leave). This rate is the smallest for the
Indian group at just 3%, both because they are higher attaining at GCSE and
because they make different choices compared to White British students living in the
same neighbourhoods with similar attainment. It is lower at just 7% for black
students and 8% for Pakistani and Bangladeshi students, despite having lower
GCSE attainment.
The white other and other ethnic groups are more similar to the White British group,
most likely because these categorisations mask wide variation in educational
choices. Nevertheless, all ethnic minority groups are much more likely than the
White British group to attend a school sixth form or sixth form college, even
accounting for prior attainment and neighbourhood. They are also more likely to take
level 2 qualifications, level 3 qualifications and 3 A levels.
All ethnic minority groups have much higher university participation: the rates for the
black, Indian and Pakistani/Bangladeshi group are 57%, 72% and 53%, respectively,
compared to 36% for the White British population. However, only the Indian ethnic
group is markedly more likely to attend a Russell Group university, compared to
White British pupils with similar GCSE attainment living in the same neighbourhoods.
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
7
Post-16 outcomes where there are few school sixth form places available
There are 20 places in England where there is almost no opportunity for students to
attend school sixth forms because they do not exist due to historic local authority
level organisational decisions. We match the neighbourhoods in these areas to
similar neighbourhoods elsewhere in England that have school sixth forms. This
allows us to estimate the impact of living in an area where all students must proceed
to a sixth form college or further education college. We find that living in an area with
no sixth form college has no overall effect on levels of post-16 participation, but it
does affect the types of courses that students study. There are significantly lower
percentages of pupils studying at least 2 level qualifications (73.5% versus 75.5%),
at least level 3 qualifications (59.9% versus 63.0% and at least level 3 academic
qualifications (40.0 versus 43.2%).
The largest impacts are for those in the middle prior attainment group who are much
less likely to study for 3 A levels. This may be because sixth form colleges serve a
smaller community of students than do school sixth forms, leaving greater numbers
of middle attainers to attend general FE colleges.
Higher prior attainment students have a five percentage point lower take-up of at
least 2 facilitating A-levels and lower rates studying for at least one science or maths
A level. One possible explanation is that sixth form colleges are able to provide a
wider range of A-levels, increasing the chances that they study less traditional
subjects. We also find this higher attaining group is slightly less likely to attend
higher education (2 percentage points) or a Russell Group university (1 percentage
points).
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
8
Introduction
The transition at the age of 16 marks the first point in most individuals’ educational
lives where opportunities and choice become markedly diverse. Students can
choose whether to attend school sixth forms, sixth form colleges and Further
Education (FE) colleges. This choice is, in part, geographically shaped by quite
idiosyncratic variation in institutional provision. This research provides an up-to-date
understanding of post-16 educational choices and transitions, highlighting the
implications of differences in choices available for students from different areas and
backgrounds. Specifically, our research analysis explores how the choice sets
available to students vary according to their geographic, social and educational
background. It investigates the impact of these institutional, subject and qualification
choices made on students’ educational trajectories, including their subsequent
educational attainment and their access to higher education.
Our study uses three linked databases – the National Pupil Database, Individual
Learner Records, and Higher Education Statistics Authority data – to explore all
choices made by all individuals, rather than those appearing the Key Stage Five
attainment tables. This allows a greater understanding of the role of both academic
and vocational pathways in producing inequalities in higher education enrolment. We
create a unique post-16 institutional choice set for each prospective student, which
allows us to distinguish between inequalities in post-16 choice and attainment that
arise between groups facing identical institutional availabilities and those that arise
because different groups have access to different types of local post-16 provision.
Previous research
The existing literature has tended to focus on the decision to participate in education
post-16, rather than on the type of institution chosen.
1
The result is a limited
evidence base on post-16 choice patterns and their implications for students. The
literature does, however, suggest that attendance at different types of post-16
institutions has a social and educational gradient to it. For example, students are
more likely to attend sixth form based provision if they are from more advantaged
social backgrounds or attend a school with a lower proportion of pupils in receipt of
free school meals, whilst they are more likely to enrol in an FE college if they are
from a non-professional background.
2
Achieving at least five GCSEs at A*-C grade
and attending a school with a sixth form are also important influences on pupils’
decisions to stay on in post-16 education, particularly in sixth-form based provision.
34
1
Clark, D., Conlon, G. and Galindo-Rueda, F. (2005) “Post-Compulsory Education and Qualification
Attainment”, in Machin, S. and A. Vignoles (eds.) What’s the Good of Education? The Economics of
Education in the United Kingdom, Princeton University Press
2
Crawford, C., Meschi, E., and Vignoles, A. (2011) Post-16 Educational Choices and Institutional
Value Added at Key Stage 5. CEE DP 124. Centre for the Economics of Education.
http://cee.lse.ac.uk/ceedps/ceedp124.pdf
3
Ibid.
4
Foskett, N., Dyke, M., and Maringe, F. (2004) The Influence of the School in the Decision to
Participate in Learning Post-16. Research Report No. 538. Department for Education and Skills.
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
9
There are also differences in pupil choices and outcomes at different types of
institutions. Pupils attending schools with sixth forms are more likely to choose
academic subjects for post-16 study, whilst those without in-school sixth forms have
a greater chance for choosing vocational subjects.
5
Sixth form colleges have been
found to add the most value to higher achieving pupils taking A-levels, and FE
colleges adding the least.
6
Students attending FE colleges are also less likely to go
to university, with those that do attend being less likely to gain acceptance to a highly
selective university.
7
This is likely due to FE colleges serving students with lower
prior attainment, making the finding that pupils are more likely to choose to enrol in
FE colleges when they live in poorer areas all the more pertinent.
8
One explanation
for different post-16 choices is local provision, with pupils living in poorer areas
having more access to FE colleges than their peers in wealthier areas.
9
Another
explanation is the behaviour of schools, which are more likely to promote
participation in post-16 education in their own sixth forms when they are based in
metropolitan or urban local authorities where other post-16 providers are also
available.
10
This suggests that institutional choice sets can affect the behaviour of
both students and institutions.
The issue of geographical variation in post-16 provision is a key gap in the existing
literature, which lacks explicit consideration of issues such as distance to nearest
provider or differences in institutional choice sets across areas. One study, using a
similar approach to the one proposed here, identifies common patterns of
institutional provision across geographical areas and assesses the impact of
provision on post-16 participation and attainment, finding that patterns in provision
does not seriously affect outcomes.
11
Another study looks at models of choice made
within local authorities, using local authority fixed effects. We expand on previous
research by explicitly considering geographically based institutional choice sets for
students.
Data
In this research we use three linked administrative datasets – the National Pupil
Database (NPD), the Individualised Learner Record (ILR) and the Higher Education
Statistics Authority dataset (HESA). Our study population is the cohort of 15 year
olds who sat their GCSEs in 2009/10, whether in the state or independent sector,
and we attempt to identify them in a post-16 institution in 2010/11 academic year. If
http://webarchive.nationalarchives.gov.uk/20130401151715/http://www.education.gov.uk/publications/
eOrderingDownload/RR538.pdf
5
Foskett, Dyke and Maringe; Thomson, D. (2015) What you study after your GCSEs depends on
where you live, Education Datalab blog at http://www.educationdatalab.org.uk/Blog/August-
2015/What-you-study-after-your-GCSEs-depends-on-where-y.aspx#
6
Crawford, Meschi and Vignoles
7
Ibid.
8
Ibid.
9
Ibid.
10
Foskett, Dyke and Maringe
11
Schagen, I., Lopes, J., Rutt, S., Savory, C., and Styles, B. (2006) Do post-16 structures matter?
Evaluating the impact of local patterns of provision. Learning and Skills Network.
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
10
they are in a school sixth form we find them in the spring school census and
subsequently find their qualifications and subjects pursued in the Key Stage 5
records in 2010/11 and 2012/13. If they are in an FE college or sixth form college we
find them via the Individualised Learner Records learning aims, which gives us their
qualifications and subjects pursued. This includes apprenticeship completions and
information on training for learners who drop-out. We find those in independent
schools in the Key Stage 5 tables, with the proviso that if they begin in this sector
and are not entered for a qualification then we will not identify their post-16 route.
HESA allows us to explore post-16 progression to higher education. The earliest we
find this cohort in the HESA data is 2012/13 and we give them three opportunities to
appear in the higher education sector by cross-checking the 2013/14 and 2014/15
data.
For each individual in the study, we aim to attach the following background
information:
Student gender
Student ethnicity, grouped into white British; white other; black; Indian;
Pakistani or Bangladeshi; other including east Asian
Student free school meals eligibility at age 15/16
12
Student home postcode at age 15/16 (or as close in age as possible)
We apply the principle that the highest data quality lies in the school census, but
draw on other sources where it is missing. For the tiny number of individuals who are
educated in the independent sector for pre-16 and post-16 study and who do not
attend university, we are missing some key background characteristics.
Institutions included in our analysis
The students in our analysis attend a number of different types of post-16
institutions: schools with sixth forms (n=1935); FE colleges (n=473); sixth form
colleges (n=97); independent schools (n=791) and other institutions such as special
schools and private training organisations (n=524). Since our analysis is interested in
the relationship between the location of the student’s home and institution, we split
large, multi-site FE colleges into multiple pseudo-institutions based on the ILR
learning aims postcode.
13
Each individual is assigned to only one institution; this is
their first institution if they have dropped-out and re-started elsewhere.
We retain individuals in independent schools and in other institutions in our analysis,
but do not include these institutions in our choice set analysis. Figure 1 shows the
cohort size of the post-16 institutions we analyse (split into pseudo-institutions as
described above, where appropriate). It shows that school sixth forms are typically
12
This measure of benefits entitlement is used to proxy income disadvantage; POLAR and NS-Sec
are available in HESA, but we do not have them for those who do not attend university so cannot use
them in this study.
13
For each student we identified the modal postcode of their location of study in the ILR learning
aims. We split large institutions into multiple pseudo-institutions where they had at least 70 students
going to one of the combinations of the provider ID and the delivery postcode.
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
11
small, with between 50 and 150 students in a cohort. FE colleges and sixth form
colleges are much larger, typically with 500-1000 students.
Figure 1 Cohort size of institutions
Figure 2 shows that state school sixth forms and sixth form colleges both principally
cater for students wishing to take the traditional level 3 route of studying for A-levels,
although a substantial minority do study vocational courses and level 2 courses (e.g.
retaking GCSE qualifications). By contrast, almost half the 16 year olds in FE
colleges are studying for level 2 qualifications, with the next largest group studying
for level 3 vocational qualifications.
Figure 2 Proportions studying different types of course by institution type
0
100
200
300
400
500
600
700
0-50 51-100 101-150 151-200 201-300 301-500 501-1000 1001-2500
Number of institutions
Number of pupils
State school FE college Sixth form college
6.9%
8.5%
47.9%
23.0%
12.7%
27.6%
68.3%
78.2%
7.2%
0.2%
0.7%
17.3%
0% 20% 40% 60% 80% 100%
State school
Sixth form college
FE college
L2 any L3 vocational L3 academic Apprenticeships/foundation learning
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
12
Given these differences in course provision, clearly state school sixth forms and sixth
form colleges require higher entry grades. Figure 3 shows that they most frequently
have a student cohort with a mean GCSE grade B.
Figure 3 Average GCSE grade of student intake by institution
Describing student post-16 outcomes
We assess the impact of the post-16 institutional choices made in two way. First, we
identify post-16 course choice, grouping courses into the following categories:
Level 3 academic (3+ A-levels) (N = 270,629)
Level 3 other, which includes 1 or 2 A-levels and mixtures of academic
and vocational qualifications (N = 45,652)
Level 3 large vocational (N = 73,508)
Level 2, which includes those taking GCSEs and any vocational
qualifications (N = 78,784)
Below level 2 (N = 34,861)
Foundation learning (N = 33,402)
Apprenticeship (N = 17,036)
Unknown qualification (N = 5,150)
Not in learning at school/college (N = 60,654)
For those studying A-levels we identify those studying two or more Russell Group
‘facilitating’ A-levels that are frequently required for university courses: biology,
chemistry, English literature, geography, history, physics, modern/classical
languages, maths and further maths. We also identify those studying two or more
science or maths A-levels, which we classify in Figure 4 as ‘SEM’.
14
14
We include a number of A-levels here that are not considered facilitating, such as human biology,
environmental science and pure maths with statistics.
0
200
400
600
800
1000
1200
G F E D C B A A*
Number of institutions
Mean GCSE grade
State school FE college Sixth form college
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
13
Figure 4 Post-16 courses studied
Our second measure of post-16 outcomes is university entry at ages 18, 19 and 20.
Figure 5 summarises this information, showing that 60% of the cohort do not attend
university in these three years. 24% attend at age 18 in 2013, 12% attend in 2014,
and 3% first attend in 2015. We also identify those who are studying at a Russell
Group university and those who are studying a science, engineering or maths
subject. Almost all the cohort are studying for a full degree (as opposed to a diploma
or certificate) so we do not explore this further in our analysis.
Figure 5 University entry rates
- 50,000 100,000 150,000 200,000 250,000 300,000
Not in learning
Unknown
Apprenticeship
Foundation learning
Below level 2
Level 2
Level 3 large vocational
Level 3 other
3+ A levels
Non-facilitating non-SEM Non-facilitating SEM Facilitating non-SEM Facilitating SEM
Did not attend
60%
Did not attend
60%
Did not attend
60%
Did not attend
60%
2013
24%
No
[VALUE]
No
[VALUE]
No
[VALUE]
2014
12%
DK
[VALUE]
2015
3%
Yes
[VALUE]
Yes
[VALUE]
Yes
[VALUE]
0% 20% 40% 60% 80% 100%
Date attending HE
Russell Group
Full degree
Science/maths degree
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
14
We do not include information on qualifications achieved at the age of 18 in our
analysis, as others have. To include A-level grades would restrict our analysis to the
271,000 who study the conventional route of 3 or more A-levels. This would be a
valid analysis only if we believed that post-16 institutional availability had no impact
on the chances of studying A-levels, a hypothesis we will challenge in this report.
Creating choice sets of institutional availability
Our analysis explores the extent to which social and ethnic differences in post-16
choices and outcomes arise from the opportunities open to them at age 16, or
whether individuals are making different choices given a similar post-16 institutional
environment.
15
To do this we generate local choice sets of post-16 options and then
explore how choices made are related to these choice sets.
Defining choice sets by observed choices made
We summarise the post-16 institutional options open to each individual by assuming
that all students living in the same Middle Super Output Area (MSOA) have the same
choices available to them. An MSOA is a small geographical area that is designed to
contain between 2,000 and 6,000 households. They divide England into 6,791 small
areas
16
and in our dataset they contain an average of 91 post-16 students (with a
minimum and maximum of 4 and 269, respectively). We believe it is reasonable to
constrain all students living in a MSOA to having the same choice set, even though
these areas are relatively large, because older students are less constrained
geographically and more able to travel further to the institution of their choice
compared to younger students.
We include a post-16 institution in the choice set of an MSOA if two or more students
in the cohort attended that institution. This follows an approach often used to
generate primary and secondary school choice sets.
17
We exclude independent
schools (and other minor institutions) from the choice sets and ignore the problem
that a tiny number of post-16 institutions are single sex. This approach to generating
choice sets assumes that choices made by students in the past reflect all feasible
choices. This has the advantage that we do not need to know admission policies or
transport links to generate the choice sets and we need not assume it is equally
straightforward to travel in all directions from the MSOA. However, the choice sets
are necessarily endogenous to the nature of the individuals who live in the MSOA. If,
for example, there are two contiguous MSOAs with completely different demographic
profiles and only in the wealthier MSOA do students ever attend a particular post-16
institution (e.g. if the journey to the institution is expensive) then it will be excluded
from the choice set of the MSOA with lower income families, even though it might be
15
Crawford et al. (2011) use two measures to capture patterns of post-16 provision: the proportion of
the total number of pupils in the previous cohort who are enrolled in an FE college for full-time post-16
education, and the percentage of secondary schools in each local authority that had a sixth form in
the previous academic year.
16
We use 6776 MSOAs in our analysis, dropping those that have zero institutions in their choice sets
17
For discussion of these methods see: Singleton, A.D., Longley, P.A., Allen, R. and O’Brien, O.
(2011) Estimating secondary school catchment areas and the spatial equity of access, Computers,
Environment and Urban Systems, 35(3)241-249.
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
15
quite close to them. We implement a procedure to check there is relative stability in
the estimation of choice sets in different student cohorts (see Appendix II).
Figure 6 below shows the distribution of the number of institutions across the 6776
MSOAs included in the analysis. It was not possible to define choice sets for 14
MSOAs because there were no institutions chosen by at least two students living in
these areas.
Figure 6 Distribution of number of institutions across MSOAs
Describing institutional availability within choice sets
We create a number of indicators to describe the institutional choices available to
students living in the MSOA. They are summarised in Table 1 and are as follows:
The number of institutions, which measures the overall amount of choice
available and will clearly be relatively highly correlated with population
density. Thus, it is highest in London and lowest in the South West and the
North East.
The percentage of pupils attending 6
th
form colleges and percentage of
pupils attending school sixth forms. The former varies a great deal across
regions where there are very few in the South West, the East Midlands
and North East. London has by far the most students studying in school
sixth forms.
The percentage of pupils attending the most popular institution is used as
a secondary measure of the diversity of choices exercised, and will again
by related to population density. So it is under 20%, on average, in
London.
The intake diversity, which is an index that measures the standard
deviation in the average institutional GCSE scores at the MSOA-level,
excluding pupils who completed their GCSEs in the independent sector. In
other words, it captures the level of post-16 institutional segregation by
KS4 attainment.
0
100
200
300
400
500
600
700
800
900
1000
1 2 3 4 5 6 7 8 9 10 11121314151617181920212223242526272829
Number of MSOAs
Number of institutions
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
16
Table 1 Choice set availability by region
Number of
institutions
% sixth form
colleges
% school sixth
forms
% attending most
popular institution
Intake diversity
mean
(s.d.)
mean
(s.d.)
mean
(s.d.)
mean
(s.d.)
mean
(s.d.)
National
8.07
(3.52)
14.05
(15.26)
43.45
(22.81)
30.95
(13.46)
0.88
(0.20)
East Midlands
7.93
(2.52)
8.32
(10.01)
45.00
(18.05)
31.51
(11.76)
0.90
(0.17)
East of England
7.23
(2.59)
9.84
(14.17)
49.98
(19.92)
33.66
(11.40)
0.92
(0.17)
London
11.58
(4.66)
14.35
(13.21)
55.12
(17.89)
18.95
(9.09)
0.91
(0.22)
North East
6.80
(2.21)
9.57
(12.81)
38.15
(21.05)
34.05
(11.87)
0.79
(0.15)
North West
7.44
(3.15)
22.24
(17.05)
32.74
(23.24)
32.24
(12.47)
0.87
(0.19)
South East
7.59
(2.85)
19.09
(19.67)
43.48
(27.96)
32.65
(13.65)
0.89
(0.21)
South West
6.86
(2.81)
10.42
(12.00)
43.10
(20.84)
37.00
(15.39)
0.80
(0.25)
West Midlands
8.40
(3.29)
10.62
(10.64)
40.81
(20.16)
29.40
(11.11)
0.89
(0.16)
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
17
Findings I – Inequalities in post-16 routes where
choices and opportunities are identical
Students from different social backgrounds, gender and ethnicity make different
post-16 choices. Here we show the extent to which these different choices arise from
differences in institutional availability rather than differences in preferences. We
present descriptive statistics of institutional availability, choice set indicators and
institutions attended. We then compare pupils with the same institutional choice set
because they have the same GCSE prior attainment (measured in deciles) and live
in the same MSOA neighbourhood. If we see systematic social background, gender
and ethnic differences in post-16 choices made after comparing pupils who have the
same institutions available to them, we can reasonably conclude that these
differences are due to pupil or family preferences rather than inequalities in
institutional availability.
Understanding differences in post-16 choices by social
background
Student of different social backgrounds – measured here via eligibility for free school
meals (FSM) - are unevenly distributed across the country. This results in FSM and
non-FSM students having different types of institutions available to them at age 16.
Table 2 shows that FSM students are more likely to live in areas that give them a
greater choice of post-16 institution (an average of 9.4 versus 8.3 institutions in the
MSOA choice set for FSM and non-FSM respectively; also fewer attending the most
popular institution, on average). The mixture of institutional options open to them is
slightly different with fewer in their choice sets tending to attend school 6
th
forms and
only marginally more attending 6
th
form colleges. The amount of variation in
institutional selectivity is the same.
Table 2 Institutional characteristics of choice sets by FSM status
FSM
Non-FSM
Mean
S.D
Mean
S.D.
Number of institutions
9.45
4.21
8.31
3.59
% sixth form colleges
14.09
13.23
13.54
14.74
% school sixth forms
41.17
21.89
43.44
22.50
% attending most popular institution
27.78
12.50
31.69
13.43
Intake diversity
0.87
0.19
0.88
0.19
These differences in the nature of the choice sets available to FSM and non-FSM
students may, in part, explain why there are differences in the proportions attending
different types of institutions, distances travelled to these institutions and the
selectivity of the institution attended.
Figure 7 shows far greater proportions of non-FSM than FSM pupils choosing school
sixth forms than sixth form colleges. For both types of institution, there is a gap in
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
18
attendance by FSM status. However, this gap is marginally in favour of FSM pupils in
the first four attainment deciles, but reverses in favour of non-FSM pupils at higher
attainment deciles.
Figure 7 Post-16 institutional attendance by KS4 decile and FSM status
Figure 8 shows non-FSM pupils attend slightly more selective institutions than FSM
pupils at each KS4 decile. The gap widens as the KS4 attainment decile increases,
indicating that higher attaining FSM pupils attend post-16 institutions with lower
attaining intakes than their similar attaining non-FSM peers.
Figure 8 Selectivity of institution attended by KS4 decile and FSM status
At higher KS4 attainment deciles, FSM pupils tend to attend institutions with more
pupils compared to non-FSM pupils (Figure 9). However, the pattern is reversed at
lower attainment deciles.
0%
10%
20%
30%
40%
50%
60%
70%
80%
1 2 3 4 5 6 7 8 9 10
Percentage of pupils
KS4 attainment decile (mean grade)
Non-FSM school sixth form FSM school sixth form
Non-FSM sixth form college FSM sixth form college
3.5
4
4.5
5
5.5
6
6.5
1 2 3 4 5 6 7 8 9 10
Selectivity of institution attended
(mean GCSE grade)
KS4 attainment decile (mean grade)
Non-FSM FSM
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
19
Figure 9 Size of institution attended by KS4 decile and FSM status
We next try to disentangle whether FSM differences in choices made at age 16
result from differences in institutional availability or differences in choices made
given the same institutional availability. We do this by running fixed effect models
that estimate the chances (i.e. odds) of an outcome for pupils who live in the same
MSOA and are in the same prior attainment decile.
Table 3 shows FSM pupils are more likely to have lower post-16 routes and
outcomes than non-FSM pupils with similar levels of prior attainment and who face
the same post-16 institutional choice sets because they live in the same area.
In this cohort, 9% of FSM students drop-out of the education system at age 16
compared to 16% of non-FSM students. This is, of course, largely related to GCSE
attainment. But we estimate over a third of this participation gap results from
differences in choices made by students with identical opportunities open to them
(we calculate this by converting the estimated odds ratio of 1.34 back to an impact
on participation rates).
We estimate 26% lower odds of staying on in a school sixth form, sixth form college
or independent school, which accounts for about a third of the overall FSM gap in
entry to a sixth form environment. We also find FSM students attend less selective
institutions that are closer to home than non-FSM students with similar attainment in
the same neighbourhood.
There are FSM gaps in the proportions studying at least level 2 qualifications (58%
versus 78%), level 3 qualifications (40% versus 66%) and 3 A levels (47% versus
21%). We estimate that 24%, 26% and 35% of these gaps, respectively, arises from
choices made by students living in the same neighbourhoods with approximately the
same GCSE attainment. They also have lower odds of studying for facilitating A
levels of science and maths A levels.
There are also marked FSM differences in higher education routes. Just 24% of FSM
pupils attend HE versus 42% of non-FSM pupils, with over a quarter of this
participation gap arising from students within the same neighbourhood with the same
0
100
200
300
400
500
600
700
1 2 3 4 5 6 7 8 9 10
Size of institution attended (pupils)
KS4 attainment decile (mean grade)
Non-FSM FSM
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
20
GCSE attainment. We estimate that as much as half the FSM gap in Russell Group
attendance (at 2% for FSM versus 10% for non-FSM) could be removed if we were
able to equalise university access between pupils living in the same neighbourhoods
and with similar GCSE attainment.
The findings in Table 3 suggest that the FSM differences in choices made at age 16
are not a result of differences in institutional availability (choice sets) but instead a
result of differences in choices made by pupils given the same choice set and level
of prior attainment.
Table 3 Post-16 routes and outcomes by social background
Effects for FSM pupils compared
to non-FSM pupils on…
OR/beta
s.e.
Number
of
pupils
Number
of
groups
with
variatio
n
Dropping out post-16
1.34
0.02
619,253
32,265
Staying on in a sixth form
0.74
0.01
619,253
32,265
Distance to institution (km)
-0.95
0.09
540,906
29,655
Selectivity of institution (mean GCSE
grade)
-0.15
0.00
541,074
29,658
Cohort size
9.42
1.83
541,094
29,659
Studying at least level 2
0.77
0.01
619,253
32,265
Studying at least level 3
0.74
0.01
619,253
32,265
Studying at least level 3 academic
0.68
0.01
619,253
32,265
Studying 2+ facilitating A-levels
0.71
0.01
619,253
32,265
Studying 1+ SEM A-levels
0.75
0.01
619,253
32,265
Attending HE
0.81
0.01
619,253
32,265
Attending a Russell Group HEI
0.53
0.02
619,253
32,265
Studying a SEM degree
0.78
0.01
619,253
32,265
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
21
Understanding differences in post-16 choices by gender
The choice sets faced by male and female students are almost identical, as can be
seen in Table 4.
Table 4 Institutional characteristics of choice sets by gender
Female
Male
Mean
S.D
Mean
S.D.
Number of institutions
8.46
3.70
8.46
3.69
% sixth form colleges
13.62
14.56
13.60
14.55
% school sixth forms
43.13
22.43
43.17
22.45
% attending most popular institution
31.16
13.38
31.19
13.38
Intake diversity
0.88
0.19
0.88
0.19
Figure 10 shows us that males and females both attend school sixth forms in
increasing proportions as their GCSE attainment increases. However, greater
proportions of males attend school sixth forms across all ten attainment deciles, and
the attendance gap widens as attainment increases. Very similar proportions of
males and females attend sixth form colleges across all deciles up until the eighth
decile, after which more females attend than males.
Figure 10 Post-16 institutional attendance by KS4 decile and gender
Females in the bottom three attainment deciles attend institutions with slightly lower
mean GCSE grades than do males (Figure 11). However, males attend more
selective post-16 institutions compared to females in the fourth attainment decile and
upwards.
0%
10%
20%
30%
40%
50%
60%
70%
80%
1 2 3 4 5 6 7 8 9 10
Percentage of pupils
KS4 attainment decile (mean grade)
Male school sixth form Female school sixth form
Male sixth form college Female sixth form college
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
22
Figure 11 Selectivity of institution attended by KS4 decile and gender
Across attainment deciles, females attend slightly larger institutions than do males
(Figure 12). This is likely due to greater proportions of males attending school sixth
forms across all attainment deciles – school sixth forms tend to be smaller than other
types of post-16 institution.
Figure 12 Size of institution attended by KS4 decile and gender
Figure 13 shows that only the highest attaining females travel further to their post-16
institutions than males. Below the eighth attainment decile, males travel further than
females, with the gap in distance travelled narrowing as prior attainment increases.
3.5
4
4.5
5
5.5
6
6.5
1 2 3 4 5 6 7 8 9 10
Selectivity of institution attended
(mean GCSE grade)
KS4 attainment decile (mean grade)
Male Female
0
100
200
300
400
500
600
700
1 2 3 4 5 6 7 8 9 10
Size of institution attended (pupils)
KS4 attainment decile (mean grade)
Male Female
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
23
Figure 13 Mean distance travelled to institution attended by KS4 decile and
gender
We next try to disentangle whether gender differences in choices made at age 16
result from differences in institutional availability or differences in choices made
given the same institutional availability.
Table 5 shows the effects for female pupils compared to male pupils on various post-
16 routes and outcomes. Boys are more likely to drop out of education altogether at
age 16 (9% versus 11%) and about half of this participation gap arises from choices
by gender of students with the same GCSE attainment. Girls are more likely to
attend an institution closer to home.
Girls are more likely to take at least level 2 qualifications (81% versus 70%), level 3
qualifications (68% versus 58%) and 2 or more A levels (48% versus 39%). But while
differences in rates studying for at least level 2 qualifications are largely explained by
choice, the differences in rates studying A levels are explained by gender differences
in GCSE attainment.
Girls are more likely than boys to attend HE (44% versus 36%) both because they
have higher GCSE attainment and because they have higher attendance for a given
level of GCSE attainment. The latter contributes about a third of this gender gap. The
pattern for elite Russell Group attendance is quite difference because, although girls
are slightly more likely to attend over (10% versus 9%), they are less likely than boys
to attend for a given level of GCSE attainment. Girls are more likely to study for a
SEM degree (note that this includes psychology, medicine, nutrition and nursing).
0
2
4
6
8
10
12
1 2 3 4 5 6 7 8 9 10
Mean distance travelled to institution
attended (km)
KS4 attainment decile (mean grade)
Male Female
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
24
Table 5 Post-16 routes and outcomes by gender
Effects for female pupils
compared to male pupils on…
OR/beta
s.e.
Number
of pupils
Number
of
groups
with
variation
Dropping out post-16
0.86
0.01
619,253
60,993
Staying on in a sixth form
1.03
0.01
619,253
60,993
Distance to institution (km)
-0.54
0.06
540,906
58,584
Selectivity of institution (mean
GCSE grade)
0.00
0.00
541,074
58,585
Cohort size
23.44
1.11
541,094
58,586
Studying at least level 2
1.66
0.01
619,253
60,993
Studying at least level 3
1.26
0.01
619,253
60,993
Studying at least level 3 academic
1.08
0.01
619,253
60,993
Studying 2+ facilitating A-levels
0.60
0.01
619,253
60,993
Studying 1+ SEM A-levels
0.50
0.00
619,253
60,993
Attending HE
1.12
0.01
619,253
60,993
Attending a Russell Group HEI
0.83
0.01
619,253
60,993
Studying a SEM degree
1.11
0.01
619,253
60,993
In Table 16 in Appendix IV we report how the effects of gender vary by social
background. The impact of being eligible for FSM is stronger for girls than it is for
boys in terms of staying on rates, staying in a 6
th
form setting, travelling shorter
distances to study, and studying for a level 2 or 3 qualification. However, we find no
interaction between gender and social background in terms of selectivity of
institution, cohort size, studying for A levels and HE attendance.
Understanding differences in post-16 choices by ethnic
background
Student of different ethnic backgrounds – measured here using six broad ethnic
groupings – have access to different types of institutions at age 16. Table 6 shows
that White British students have the smallest choice of institutions in a choice set,
and the greatest concentration of pupils attending the most popular institution. This is
because ethnic minority students tend to live in urban areas, where more institutions
are found as a result of greater population density, leading to greater choice and less
concentration within the most popular institution. Pupils of Indian background have
the greatest access to school sixth forms, whilst Black and Pakistani and
Bangladeshi background pupils have the greatest access to sixth form colleges.
Pupils of all ethnic groups face choice sets with a similar amount of intake diversity.
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
25
Table 6 Institutional characteristics of choice sets by ethnicity
Black
Indian
Other
ethnicity
Pakistani
/Bangladeshi
White British
White Other
Mean
S.D
Mean
S.D.
Mean
S.D.
Mean
S.D.
Mean
S.D.
Mean
S.D.
Number of institutions
12.62
4.74
11.01
4.51
9.77
4.41
10.78
4.31
7.86
3.16
9.87
4.38
% sixth form colleges
14.31
10.71
12.01
12.41
13.62
13.82
14.27
12.20
13.58
14.96
13.67
13.99
% school sixth forms
46.87
18.29
48.65
20.93
47.37
22.13
44.44
21.49
42.20
22.70
46.71
21.78
% attending most
popular institution
20.17
10.10
27.49
12.34
27.42
13.48
26.92
12.06
32.63
13.21
26.61
12.93
Intake diversity
0.93
0.16
0.93
0.16
0.90
0.19
0.88
0.19
0.88
0.19
0.91
0.19
Figure 14 below shows increasing proportions of pupils attending school sixth forms
for higher KS4 attainment deciles, with the rate of increase varying across ethnic
groups. The Indian group has the greatest percentage of pupils selecting school
sixth forms, whilst the White British group tends to be amongst the lowest across
attainment deciles. The Pakistani-Bangladeshi group is notable for the stall in school
sixth form attendance by its members between the 5
th
and 6
th
attainment deciles.
Figure 14 School sixth form attendance by KS4 decile and ethnicity
Figure 15 below shows the proportion of pupils attending sixth form colleges by KS4
attainment decile and ethnicity. The lower percentage of Pakistani and Bangladeshi
pupils choosing school sixth forms is partly explained by the higher rate of
attendance of this group at sixth form colleges.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
1 2 3 4 5 6 7 8 9 10
Percentage of pupils attending
school sixth forms
KS4 attainment decile (mean grade)
Black Indian Other Pakistani/Bangladeshi White British White Other
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
26
Figure 15 Sixth form college attendance by KS4 decile and ethnicity
Figure 16 shows that Indian pupils tend to attend slightly more selective institutions
across all KS4 deciles, whilst higher attaining Black, White British, Pakistani and
Bangladeshi pupils tend to attend less selective institutions.
Figure 16 Selectivity of institution attended by KS4 decile and ethnicity
Figure 17 shows that White British pupils tend to attend much larger institutions than
students from other ethnic backgrounds, particularly at lower KS4 attainment deciles.
This is likely due to greater proportions of White British students attending further
education colleges, which tend to be larger than sixth form colleges or school sixth
forms, than pupils of other ethnic backgrounds.
0%
5%
10%
15%
20%
25%
30%
35%
1 2 3 4 5 6 7 8 9 10
Percentage of pupils attending
sixth form colleges
KS4 attainment decile (mean grade)
Black Indian Other Pakistani/Bangladeshi White British White Other
3.5
4
4.5
5
5.5
6
6.5
1 2 3 4 5 6 7 8 9 10
Selectivity of institution attended (mean
GCSE grade)
KS4 attainment decile (mean grade)
Black Indian Other Pakistani/Bangladeshi White British White Other
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
27
Figure 17 Size of institution attended by KS4 decile and ethnicity
We next try to disentangle whether ethnic differences in choices made at age 16
result from differences in institutional availability or differences in choices made
given the same institutional availability.
Table 7 shows the chances of a post-16 outcome occurring for each ethnic minority
group compared to White British pupils, by comparing those living in the same
MSOA and with the same KS4 attainment decile. In general, ethnic minority pupils
experience better post-16 routes and outcomes than White British pupils in similar
circumstances, although the size of the difference varies according to the outcome
and the ethnic minority group under consideration.
Ethnic minority groups are more likely to drop out than are the White British
population (for whom 10% leave). This rate is the smallest for the Indian group at just
3%, both because they are higher attaining at GCSE and because they make
different choices compared to White British students living in the same
neighbourhoods with similar attainment. It is lower at just 7% for black students and
8% for Pakistani and Bangladeshi students, despite having lower GCSE attainment.
The white other and other ethnic groups are more similar to the White British group,
most likely because these categorisations mask wide variation in educational
choices. Nevertheless, all ethnic minority groups are much more likely than the
White British group to attend a school sixth form or sixth form college, even
accounting for prior attainment and neighbourhood. They are also more likely to take
level 2 qualifications, level 3 qualifications and 3 A levels.
All ethnic minority groups have much higher university participation: the rates for the
black, Indian and Pakistani/Bangladeshi group are 57%, 72% and 53%, respectively,
compared to 36% for the White British population. However, only the Indian ethnic
group is markedly more likely to attend a Russell Group university, compared to
White British pupils with similar GCSE attainment living in the same neighbourhoods.
0
100
200
300
400
500
600
700
1 2 3 4 5 6 7 8 9 10
Size of institution attended (pupils)
KS4 attainment decile (mean grade)
Black Indian Other Pakistani/Bangladeshi White British White Other
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
28
Table 7 Post-16 routes and outcomes by ethnicity
Effects for ethnic minority pupils
compared to White British pupils
on…
OR/beta
s.e.
Number
of
pupils
Number
of
groups
with
variatio
n
Dropping out post-16
601,503
32,919
Black
0.54
0.02
Indian
0.54
0.02
Other
1.04
0.03
Pakistani/Bangladeshi
0.64
0.02
White Other
0.96
0.02
Staying on in a sixth form
601,503
32,919
Black
1.98
0.04
Indian
3.41
0.10
Other
1.49
0.03
Pakistani/Bangladeshi
2.59
0.06
White Other
1.19
0.02
Distance to institution (km)
524,677
30,532
Black
-0.51
0.15
Indian
-2.66
0.19
Other
0.45
0.17
Pakistani/Bangladeshi
-3.14
0.16
White Other
-0.11
0.13
Selectivity of institution (mean
GCSE grade)
524,837
30,540
Black
0.12
0.01
Indian
0.28
0.01
Other
0.10
0.01
Pakistani/Bangladeshi
0.16
0.01
White Other
0.03
0.01
Cohort size
524,857
30,544
Black
-16.78
3.13
Indian
-78.34
3.98
Other
-25.05
3.43
Pakistani/Bangladeshi
-55.64
3.43
White Other
-3.41
2.71
Studying at least level 2
601,503
32,919
Black
2.25
0.06
Indian
2.93
0.12
Other
1.22
0.03
Pakistani/Bangladeshi
2.01
0.05
White Other
1.26
0.03
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
29
Studying at least level 3
601,503
32,919
Black
1.80
0.04
Indian
2.37
0.09
Other
1.12
0.03
Pakistani/Bangladeshi
1.69
0.05
White Other
1.13
0.02
Studying at least level 3
academic
601,503
32,919
Black
1.78
0.04
Indian
2.83
0.09
Other
1.31
0.03
Pakistani/Bangladeshi
2.12
0.06
White Other
1.08
0.02
Studying 2+ facilitating A-levels
601,503
32,919
Black
1.32
0.03
Indian
2.16
0.06
Other
1.48
0.04
Pakistani/Bangladeshi
1.82
0.05
White Other
1.10
0.02
Studying 1+ SEM A-levels
601,503
32,919
Black
1.57
0.04
Indian
3.12
0.08
Other
1.80
0.04
Pakistani/Bangladeshi
2.38
0.06
White Other
1.08
0.02
Attending HE
601,503
32,919
Black
5.66
0.12
Indian
5.68
0.16
Other
2.45
0.06
Pakistani/Bangladeshi
4.32
0.10
White Other
1.72
0.03
Attending a Russell Group HEI
601,503
32,919
Black
1.00
0.04
Indian
1.89
0.07
Other
1.46
0.05
Pakistani/Bangladeshi
1.34
0.05
White Other
1.23
0.04
Studying a SEM degree
601,503
32,919
Black
4.67
0.10
Indian
5.38
0.15
Other
2.23
0.05
Pakistani/Bangladeshi
3.81
0.09
White Other
1.60
0.03
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
30
It is plausible that ethnic differences in post-16 choices vary further by social
background. Table 16 in Appendix IV explores whether this is the case. It shows the
effects of combinations of ethnicity and social background on post-16 routes and
outcomes and highlights the general trend of more negative effects of poverty on
White British pupils compared to non-White British pupils. For example, White British
pupils in receipt of FSM have 44% lower odds of following a level 3 academic route
than similarly attaining non-FSM White British pupils living in the same MSOA. In
comparison, the next largest gap is for the Black group, where FSM pupils have 23%
lower odds of following a level 3 academic route compared to non-FSM Black group
pupils.
Table 18 in Appendix IV explores the effects of combinations of ethnicity and gender
on post-16 routes and outcomes. Girls of all ethnic groups tend to be more likely to
stay on in education and follow all levels of qualifications than boys of the same
ethnic groups. Notably, gender gaps are often larger for pupils in the Black group
than in other groups. For example, Black females have 31% lower odds of dropping
out and have 39% higher odds of attending university than similar Black males. By
comparison, White females have only 14% lower odds of dropping out and 13%
greater odds of attending university than similar White males. The finding of no
significant difference between Black females and males in the likelihood of attending
a Russell Group university is interesting, especially given that females of all other
ethnic groups are significantly less likely to study at a Russell Group institution than
males of the equivalent groups. The lack of a gender difference in Russell Group
attending for Black pupils is likely due to the much better educational outcomes of
Black females compared to similar Black males.
Social Mobility Commission
Social and ethnic inequalities in choice available and choices made at age 16
31
Findings II – Post-16 outcomes in areas with few
school sixth forms
There is a great deal of variation in post-16 institutional set-ups across England.
However, the causation between local demographics, institutional structures and
post-16 outcomes is not clear. For example, more affluent areas are likely to have
greater provision of academic post-16 study and more rural communities will have
fewer post-16 choices available.
However, there are 20 places in England where we believe it is possible to argue
that current demographic trends have not determined the institutional mix. These are
places where there is almost no opportunity for students to attend school sixth forms
because there is very little provision. This arose where local authorities decided to
organise all post-16 provision into larger sixth form and further education colleges
from the 1960s onwards, usually as part of comprehensive reorganisation. We map
out these areas in Appendix I where it can be seen that they are distributed across
the country without any obvious shared characteristics. It is slightly arbitrary whether
we decide to include an area in this list, but we roughly adhere to the principle that
there must be five or more contiguous MSOAs where numbers attending school sixth
forms is close to zero. The list includes:
Barrow-in-Furness
Blackpool
Bury
Cambridge
Chorley South Ribble
Devon
Eastbourne
Exeter
Great Yarmouth
Hampshire
Kirklees
Lewes
Luton
Mid Cornwall
North Lincolnshire
Portsmouth
Redcar Stockton
Selby
West Cornwall
Worthing
Matching across MSOAs
These 20 areas without school sixth forms include 470 MSOAs. We match each of
these MSOAs with no school sixth forms to their three most similar MSOAs
elsewhere, via a propensity score match based on similarity of:
mean GCSE grade;
percentage of White British residents;
population density; and
percentage of residents employed in class I or II occupations.
MSOAs were sampled with replacement, meaning that the same control MSOAs
could be matched to more than one treatment MSOA.
Table 8 shows the proportions attending a school sixth form in our ‘zero’ school sixth
form areas, compared to the matched areas and compared to the country as a
whole, alongside key matching characteristics. Apart from the percentage of pupils
attending a school sixth form in an area (the very characteristic used to define
treatment and matched areas), all other matching characteristics are very similar in
treatment and matched areas. There are no significant differences between the
average characteristics of the treatment areas compared to matched areas.
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
38
Table 8: Demographic characteristics of MSOAs in areas with no school sixth
forms, compared to matched areas and England as a whole
Treatment
areas
Matched
areas
Difference between
treatment and
matched
(* indicates
statistically significant
difference)
England
Number of MSOAs
470
1,410
N/A
6,776
% attending a school sixth
form
1.6
37.0
35.396 *
36.2
Mean GCSE grade
4.7
4.7
0.003
4.7
% UK born residents
92.0
92.5
0.006
87.7
Population density
27.0
26.3
0.656
33.1
% social class 1 and 2
26.5
26.4
0.000
27.7
% employed
65.2
65.3
0.001
64.8
Note: The 1,410 matched MSOAs include some MSOAs that were drawn and matched more than
once (matching with replacement). 162 MSOAs were drawn and matched twice; 26 MSOAs were
drawn and matched three times; 3 MSOAs were drawn and matched four times.
Figure 18 shows the distribution of the percentage of pupils attending school sixth
forms across all MSOAs. The bars are shaded according to the type of MSOA –
treatment MSOAs (those with zero school sixth forms), matched MSOAs (those with
similar characteristics to the treatment MSOAs but with school sixth forms), or other
MSOAs (neither treatment nor matched). The chart clearly shows that treatment
MSOAs are concentrated in the left-most bars, which makes sense as areas with
zero school sixth forms will typically have very low percentages of pupils attending
school sixth forms. By contrast, the matched MSOAs feature throughout the
distribution, indicating that the matched MSOAs are very varied in terms of the
percentage of pupils attending school sixth forms, even if they are very similar along
other dimensions to the treatment MSOAs.
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
39
Figure 18 Distribution of the percentage of pupils attending school sixth forms
in treatment and matched MSOAs
It is important to note that the ‘treatment’ areas do not simply differ in the proportion
of places available school sixth forms because this, in turn, affects who attends
general FE colleges. Table 23 in Appendix IV shows that areas with few school sixth
forms also have far larger numbers of low and middle attaining students attending
general FE colleges. This is because sixth form colleges are generally more
selective in their intake than school sixth forms are.
MSOA-level regressions
We run MSOA-level regressions on this sample of 1,880 MSOAs – 470 with the
‘treatment’ of few school sixth forms and 1,410 others. The analysis is carried out for
all pupils and then repeated by prior attainment tercile.
The results are presented in Table 9. The first column of results of shows that living
in an area with few school sixth forms has no overall effect on levels of post-16
participation, but it does affect the types of courses student study. There are
significantly lower percentages of pupils studying at least 2 level qualifications
(73.5% versus 75.5%), at least level 3 qualifications (59.9% versus 63.0%) and at
least level 3 academic qualifications (40.0 versus 43.2%). Further, there are
significantly lower percentages of pupils studying two or more facilitating A-levels,
and also lower percentages studying one or more SEM A-levels.
The negative effects tend to follow similar patterns for each of the three GCSE
attainment groups. However, low attainers experience particularly negative effects in
terms of the percentage studying at least level 2 (-4.89 percentage points) and at
least level 3 qualifications (-5.32 percentage points). This suggests that living in an
area with school sixth forms may provide lower attaining pupils with opportunities to
study rather than take up apprenticeships or below level 2 qualifications.
The largest impacts are for those in the middle prior attainment group who are much
less likely to study for 3 A levels. This may be because sixth form colleges serve a
smaller community of students than do school sixth forms, leaving greater numbers
of middle attainers to attend general FE colleges.
0
100
200
300
400
500
600
700
800
900
0-5 5-10 10-1515-2020-2525-3030-3535-4040-4545-5050-5555-6060-6565-7070-7575-8080-8585-90
Number of MSOAs
Percentage attending school sixth forms
Matched Other Treatment
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
40
Higher prior attainment students have a five percentage point lower take-up of at
least 2 facilitating A-levels and lower rates studying for at least one science or maths
A level. One possible explanation is that sixth form colleges are able to provide a
wider range of A-levels, increasing the chances that they study less traditional
subjects. We also find this higher attaining group is slightly less likely to attend
higher education (2 percentage points) or a Russell Group university (1 percentage
points), although this last finding may not be robust to alternative model
specifications shown in Appendix VII.
Table 9 Effects of living in an area with no school sixth forms on post-16
participation by prior attainment (MSOA matching)
Effect of living in an area with no school
sixth forms on the percentage of pupils…
Effect of living
in treatment
area (no
school sixth
forms)
Effect of living
in treatment
area (no
school sixth
forms)
Effect of living
in treatment
area (no
school sixth
forms)
Effect of living
in treatment
area (no
school sixth
forms)
All pupils
Low attainers
Middle
atttainers
High attainers
beta
(s.e.)
beta
(s.e.)
beta
(s.e.)
beta
(s.e.)
Percent participating in post-16 education
0.27
0.21
0.57
0.52
-0.27
0.28
-0.14
0.19
Percent studying at least level 2
-1.91
0.28
-4.89
0.66
-1.50
0.40
-0.67
0.24
Percent studying at least level 3
-2.56
0.28
-5.32
0.52
-2.88
0.49
-0.91
0.27
Percent studying at least level 3 academic
-3.08
0.34
-1.25
0.19
-5.93
0.73
-3.00
0.51
Percent studying at least 2 facilitating A-levels
-2.75
0.26
-0.17
0.05
-3.16
0.36
-5.51
0.69
Percent studying at least 1 SEM A-level
-2.24
0.27
-0.54
0.11
-3.63
0.44
-3.06
0.66
Percent attending HE
-0.22
0.30
0.28
0.27
0.60
0.58
-2.65
0.58
Percent attending a Russell Group HEI
-0.61
0.37
0.02
0.06
0.23
0.13
-1.13
0.52
Percent studying for a SEM degree
-0.68
0.44
-0.03
0.13
0.39
0.34
-1.21
0.59
Note: N=1,880 MSOAs split into 470 treatment MSOAs and 1,410 matched MSOAs. Each model
controls for the following at the MSOA-level: percentage of rented households, the percentage of
socially rented households, the percentage of degree-qualified adults, percentage of unemployed
adults, percentage of employed adults, percentage of full-time employed adults, percentage of UK
born adults, percentage of adults belonging to social class I or II, the population density, and the
population density squared. Each model also controls for following pupil characteristics at the MSOA-
level: the percentage of FSM pupils, the percentage of female pupils, the percentage of White British
pupils, the percentage of Black pupils, the percentage of White Other pupils, the percentage of
Pakistani and Bangladeshi pupils, the percentage of Indian pupils, the percentage of pupils belonging
to Other ethnic backgrounds, the percentage of SEN pupils, the percentage of EAL pupils, the
percentage of pupils achieving 5 or more GCSEs at A*-C grade (and a squared term), the mean
number of A-A* GCSE passes (and a squared term), the mean GCSE grade (and a squared term),
and the percentage of pupils attending independent schools whilst in Year 11.
In general, it seems that living in an area with no school sixth forms has significant
effects on the types of post-16 courses studied by students across the attainment
distribution. It also lowers university participation for higher attainers.
In Appendix VII we show the results from an alternative matching technique that
matches individual students living in no school sixth form areas to those living
elsewhere in the country. The results are largely the same. This pupil matching
approach allows us to report the impact on particular pupil groups. It shows that FSM
pupils are generally less affected than are non-FSM pupils, which is logical since the
absence of school sixth forms affects higher attaining pupils the most. It also shows
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
41
that boys are more affected than are girls, particularly with respect to the proportion
studying for A levels and attending more elite universities. The numbers in minority
ethnic groups who live in these areas are generally too small to estimate ethnic
differences in impact.
Possible causes
We can only speculate as to why areas without school sixth forms see progression
into less academic post-16 routes, on average, and also have slightly less successful
higher education progression for the most able students. Historically, there have
been differences in how sixth form and FE colleges are judged, compared to school
sixth forms. Sixth form and FE colleges are judged on the proportion of their students
who successfully complete their course, which may incentivise a more cautious
approach to matching prospective students to course. By contrast, up to 2015
schools were not judged on this metric and so it is possible that schools were more
likely to give students the ‘benefit of the doubt’ who wish to progress to A levels and
other level 3 qualifications with relatively weak GCSE grades.
Another likely explanation for less academic course take-up in areas with school
sixth forms is simply that school sixth forms tend to be small and so can only offer a
relatively restricted academic curriculum. Such a restricted curriculum is expected to
encourage participation in level 3 qualifications and ‘traditional’ choices that are more
conducive to university participation.
These two explanations alone may be enough to explain differences in post-16
outcomes, including progression to higher education. But there are three other
differences between study in schools and in other institutions that warrant
consideration. Firstly, school sixth forms have smaller sized cohorts overall and
many find they must operate with quite small study classes, which may benefit the
students. Secondly, they may recruit a more homogenous intake to their classes,
making them more straightforward to teach. It is beyond the scope of this study to
analyse the ability distribution of classes within post-16 settings.
Finally, pupils who continue on to post-16 education in the same school rather than
changing to another institution may benefit from the established relationships and
familiar setting of their secondary school. Whilst changing institutions for post-16
education may well benefit some students, both academically and non-academically,
other pupils who remain in school for sixth form may gain advantages through the
consistency of educational provision and the longer “run up” to post-16 choices.
The observation that higher attaining students appear to be affected by institutional
choices available to them at 16 is confirmed by evidence on university course
selection. Figure 19 shows whether pupils make ambitious HE choices given their
GCSE results and A-level results. Our metric of course ambition is the student’s own
QCA points achieved via A-levels or equivalents divided by the average QCA points
for students attending the same university and course in this cohort. A value of
above 1 means the student will have better A level grades than their university
peers; a value below 1 means they will have lower A level grades than their peers.
We plot these average ratios for different institutional types by the Key Stage 5 (QCA
point) attainment of the student (excluding the lowest attaining since so few go to
university). The chart shows that students from independent schools tend to have
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
42
worse A level grades than their university course peers. One interpretation is that
independent schools are successful in ‘stretching’ their students to achieve an
ambitious course entry, given their grades. By contrast, sixth form college and
general FE students tend to have slightly better QCA points than their university
course peers; indeed, they might well have been able to successfully enter a more
ambitious university course with the right support. In the case of general FE colleges
where many non A-level subjects are taken, it is a signal that universities may value
their qualifications less than the value assigned under QCA conversion. Overall we
take this as evidence that school sixth forms have a better success rate at securing
the best possible university place for their students.
Figure 19: ‘Stretch’ of university attended, as measured by ratio of own post-
16 outcomes divided by the mean for university-course attended
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
5 6 7 8 9 10=high
Ratio of own QCA points to average for
uni-course attended
Student's QCA points decile
School sixth form Independent Sixth form college General FE
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
43
Appendices
Appendix I – Maps of post-16 institutional availability
Figure 20 MSOAs with no school sixth forms
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
44
Figure 21 Proportion of school sixth forms across MSOAs
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
45
Figure 22 Diversity of intakes across MSOAs
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
46
Appendix II – Year-on-year stability of MSOA choice set
indicators
Robustness checks on the year-on-year stability of our choice set measures shows
that choice set measures are generally stable with a small amount of local instability.
The relatively high year-on-year correlation between pairs of choice set indicators
can be seen in Table 10.
Table 10 Year-on-year correlation between choice set indicators
Choice set indicator
Correlation
2010-2011
Number of pupils
0.860
% attending state-funded mainstream schools Y12
0.931
% attending sixth form colleges Y12
0.935
% attending colleges of general further education Y12
0.881
Number of Y12 schools, SFCs and GFEs with at least 1
pupil
0.881
Table 11 shows the change in number of Year 12 institutions in the choice set for
each MSOA. We can see that 39% of MSOAs hardly change at all in terms of the
number of Year 12 institutions within their choice sets, while there is a difference of
no more than 6 institutions for 95% of MSOA choice sets.
Table 11 Change in number of Y12 schools, sixth form colleges and further
education colleges with at least 1 pupil 2010-2011
Number of MSOAs experiencing specific
changes in the number of post-16 institutions
7+ fewer
inst.
2 to 6
fewer
inst.
Within 1
inst.
2 to 6
more
inst.
7+ more
inst.
Total
MSOA
s
Number of
post-16
institutions
with at
least 1
pupil in
2010
<= 5
0
3
94
104
5
206
6 - 10
0
304
892
774
31
2,001
11 - 15
9
695
976
626
67
2,373
16 - 20
35
449
407
285
30
1,206
21 - 25
32
179
145
108
31
495
26 - 30
29
113
82
63
26
313
31 - 35
21
40
33
29
7
130
36 - 40
7
19
17
12
1
56
41 +
4
3
2
1
1
11
Total
137
1,805
2,648
2,002
199
6,791
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
47
In terms of the type of institutions pupils are going to, we find that there is very little
change in the number of sixth form colleges chosen by pupils. Instead, most of the
change in the composition of choice sets is due to pupils selecting different schools
and further education colleges. This can be seen in Table 12 below, which shows
how the distribution of post-16 institutions changes between 2010 and 2011 for the
4,143 MOSAs which are not in the “within 1 column” in Table 11. It is clear that there
is very little change in the number of sixth form colleges in choice sets between 2010
and 2011.
The 2011 pupil cohort is slightly smaller than the 2010 pupil cohort. This means that,
overall, we have fewer pupils going to more institutions. This leads up to assume that
there was more choice in institutions for the 2011 cohort because of the spare
capacity in post-16 institutions for this year group compared to the 2010 cohort.
Table 12 Change in composition of MSOA choice sets by institution type
Institution type
Schools
Sixth form
colleges
Further
education
colleges
Number of MSOAs with:
More institutions in choice set in 2011 than in 2010
1,168
173
1,278
Fewer institutions in choice set in 2011 than in 2010
1,010
214
1,067
One institution in 2011 of 2010 choice set
1,965
3,756
1,798
Total MSOAs
4,143
4,143
4,143
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
48
Appendix III – Descriptive statistics of choice set indicators,
institutional choices, and post-16 outcomes by pupil prior
attainment tercile
Table 13 Choice set indicators and institutional choices by prior attainment
tercile
Choice set indicators and
institutional choices
Low
attainers
Middle
attainers
High
attainers
All
attainers
Mean GCSE grade
2.7
4.9
6.5
4.7
Selectivity of post-16 institution
3.9
4.8
5.8
4.9
Size of post-16 institution (pupils)
591
551
425
516
Distance to post-16 institution (km)
9.0
7.9
6.8
7.8
Percentage attending school sixth
forms
12.7
37.5
57.8
36.0
Percentage attending sixth form
colleges
4.1
16.2
22.3
14.2
Percentage attending further
education colleges
54.1
37.1
9.2
33.5
Table 14 Percentage of pupils achieving post-16 outcomes by prior attainment
tercile
Percentage of pupils…
Low
attainers
Middle
attainers
High
attainers
All
attainers
Participating in post-16 education
78.4%
94.3%
98.0%
90.2%
Studying at least level 2
42.9%
87.2%
96.7%
75.6%
Studying at least level 3
15.1%
77.6%
96.0%
62.9%
Studying at least level 3 academic
1.9%
40.8%
88.3%
43.7%
Studying at least 2 facilitating A-levels
0.3%
9.8%
53.2%
21.1%
Studying at least 1 SEM A-level
0.9%
15.2%
54.0%
23.4%
Attending HE
6.4%
34.5%
78.2%
39.7%
Attending a Russell Group HEI
6.3%
33.2%
51.4%
30.3%
Studying for a SEM degree
2.9%
19.3%
41.3%
21.2%
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
49
Appendix IV – Post-16 route and outcomes – interacting
combinations of pupil characteristics
Table 15 Post-16 routes and outcomes – interacting social background with
prior attainment
OR/bet
a
s.e.
Lower
boun
d
Upper
boun
d
Numbe
r of
pupils
Number
of
groups
with
variation
Dropping out post-16
619253
6134
OR Low FSM to non-FSM
1.08
0.08
1.05
1.11
OR Middle FSM to non-
FSM
1.21
0.19
1.14
1.28
OR High FSM to non-FSM
1.20
0.18
1.05
1.37
Staying on in a sixth
form
619253
6134
OR Low FSM to non-FSM
0.91
-0.10
0.88
0.93
OR Middle FSM to non-
FSM
0.82
-0.20
0.80
0.84
OR High FSM to non-FSM
0.63
-0.46
0.60
0.67
Distance to institution
(km)
540906
12565
Beta Low non-FSM
9.76
0.08
9.62
9.91
Beta Middle non-FSM
8.47
0.06
8.34
8.59
Beta High non-FSM
8.44
0.07
8.31
8.57
Beta Low FSM
8.69
0.12
8.45
8.93
Beta Middle FSM
7.44
0.15
7.14
7.74
Beta High FSM
6.74
0.22
6.31
7.18
Selectivity of institution
(mean GCSE grade)
541074
12568
Beta Low non-FSM
4.07
0.00
4.06
4.08
Beta Middle non-FSM
4.86
0.00
4.86
4.87
Beta High non-FSM
5.60
0.00
5.60
5.61
Beta Low FSM
3.97
0.01
3.96
3.98
Beta Middle FSM
4.71
0.01
4.70
4.72
Beta High FSM
5.35
0.01
5.33
5.36
Cohort size
541094
12568
Beta Low non-FSM
579.43
2.21
575.1
0
583.7
6
Beta Middle non-FSM
534.10
1.90
530.3
8
537.8
2
Beta High non-FSM
456.66
2.05
452.6
3
460.6
8
Beta Low FSM
567.78
2.99
561.9
2
573.6
5
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
50
Beta Middle FSM
540.07
3.35
533.5
1
546.6
4
Beta High FSM
505.77
4.73
496.5
0
515.0
4
Studying at least level 2
619253
6134
OR Low FSM to non-FSM
0.82
-0.20
0.81
0.84
OR Middle FSM to non-
FSM
0.91
-0.09
0.88
0.95
OR High FSM to non-FSM
1.06
0.06
0.94
1.20
Studying at least level 3
619253
6134
OR Low FSM to non-FSM
0.76
-0.27
0.74
0.78
OR Middle FSM to non-
FSM
0.81
-0.21
0.79
0.84
OR High FSM to non-FSM
1.04
0.04
0.94
1.16
Studying at least level 3
academic
619253
6134
OR Low FSM to non-FSM
0.56
-0.58
0.51
0.61
OR Middle FSM to non-
FSM
0.79
-0.24
0.77
0.81
OR High FSM to non-FSM
0.78
-0.25
0.73
0.82
Studying 2+ facilitating
A-levels
619253
6134
OR Low FSM to non-FSM
0.71
-0.34
0.57
0.89
OR Middle FSM to non-
FSM
0.85
-0.17
0.80
0.89
OR High FSM to non-FSM
0.99
-0.01
0.95
1.04
Studying 1+ SEM A-
levels
OR Low FSM to non-FSM
0.81
-0.21
0.72
0.91
OR Middle FSM to non-
FSM
0.91
-0.09
0.88
0.95
OR High FSM to non-FSM
1.14
0.13
1.09
1.19
QCA points
409152
10782
Beta Low non-FSM
421.36
1.47
418.4
8
424.2
4
Beta Middle non-FSM
572.16
0.90
570.4
0
573.9
2
Beta High non-FSM
843.14
0.88
841.4
2
844.8
6
Beta Low FSM
402.15
2.76
396.7
5
407.5
5
Beta Middle FSM
551.57
2.11
547.4
4
555.7
0
Beta High FSM
792.76
2.80
787.2
7
798.2
6
Attending HE
619253
6134
OR Low FSM to non-FSM
1.00
0.00
0.96
1.04
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
51
OR Middle FSM to non-
FSM
1.07
0.07
1.04
1.11
OR High FSM to non-FSM
0.88
-0.12
0.84
0.93
Attending a Russell
Group HEI
619253
6134
OR Low FSM to non-FSM
0.58
-0.54
0.44
0.78
OR Middle FSM to non-
FSM
1.03
0.03
0.92
1.17
OR High FSM to non-FSM
0.70
-0.36
0.66
0.74
Studying a SEM degree
619253
6134
OR Low FSM to non-FSM
0.98
-0.02
0.92
1.04
OR Middle FSM to non-
FSM
1.06
0.05
1.02
1.09
OR High FSM to non-FSM
0.86
-0.15
0.82
0.91
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
52
Table 16 Post-16 routes and outcomes – interacting gender with social
background
OR/be
ta
s.e.
Lower
bound
Upper
bound
Numb
er of
pupils
Number
of
groups
with
variation
Dropping out post-16
61925
3
30937
OR FSM female to FSM
male
0.93
-0.07
0.90
0.97
OR non-FSM female to
non-FSM male
0.84
-0.18
0.82
0.85
OR female FSM to female
non-FSM
1.43
0.36
1.39
1.48
OR male FSM to male
non-FSM
1.29
0.25
1.25
1.33
Staying on in a sixth
form
61925
3
30937
OR FSM female to FSM
male
1.11
0.10
1.07
1.14
OR non-FSM female to
non-FSM male
1.02
0.02
1.01
1.03
OR female FSM to female
non-FSM
0.77
-0.26
0.75
0.79
OR male FSM to male
non-FSM
0.71
-0.34
0.69
0.73
Distance to institution
(km)
54090
6
28000
Beta non-FSM male
9.07
0.05
8.97
9.17
Beta non-FSM female
8.56
0.05
8.46
8.66
Beta FSM male
8.21
0.12
7.98
8.45
Beta FSM female
7.53
0.12
7.30
7.77
Selectivity of institution
(mean GCSE grade)
54107
4
28003
Beta non-FSM male
4.86
0.00
4.85
4.87
Beta non-FSM female
4.86
0.00
4.85
4.86
Beta FSM male
4.68
0.01
4.67
4.69
Beta FSM female
4.72
0.01
4.71
4.74
Cohort size
54109
4
28004
Beta non-FSM male
509.17
1.50
506.24
512.11
Beta non-FSM female
533.94
1.51
530.98
536.89
Beta FSM male
523.84
2.65
518.65
529.04
Beta FSM female
536.52
2.65
531.33
541.71
Studying at least level 2
61925
3
30937
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
53
OR FSM female to FSM
male
1.56
0.01
1.52
1.61
OR non-FSM female to
non-FSM male
1.70
0.01
1.67
1.72
OR female FSM to female
non-FSM
0.72
0.01
0.70
0.73
OR male FSM to male
non-FSM
0.78
0.01
0.76
0.80
Studying at least level 3
61925
3
30937
OR FSM female to FSM
male
1.16
0.01
1.13
1.20
OR non-FSM female to
non-FSM male
1.28
0.01
1.27
1.30
OR female FSM to female
non-FSM
0.69
0.01
0.68
0.71
OR male FSM to male
non-FSM
0.77
0.01
0.75
0.78
Studying at least level 3
academic
61925
3
30937
OR FSM female to FSM
male
1.13
0.02
1.09
1.17
OR non-FSM female to
non-FSM male
1.08
0.01
1.07
1.09
OR female FSM to female
non-FSM
0.69
0.01
0.67
0.71
OR male FSM to male
non-FSM
0.66
0.01
0.64
0.68
Studying 2+ facilitating
A-levels
61925
3
30937
OR FSM female to FSM
male
0.68
0.02
0.65
0.71
OR non-FSM female to
non-FSM male
0.60
0.01
0.59
0.60
OR female FSM to female
non-FSM
0.77
0.02
0.74
0.80
OR male FSM to male
non-FSM
0.68
0.02
0.65
0.70
Studying 1+ SEM A-
levels
OR FSM female to FSM
male
0.59
0.02
0.57
0.62
OR non-FSM female to
non-FSM male
0.49
0.01
0.49
0.50
OR female FSM to female
non-FSM
0.85
0.02
0.82
0.88
OR male FSM to male
non-FSM
0.71
0.02
0.69
0.73
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
54
QCA points
40915
2
18444
Beta non-FSM male
634.92
1.07
632.82
637.02
Beta non-FSM female
655.56
1.06
653.48
657.65
Beta FSM male
595.00
2.25
590.58
599.41
Beta FSM female
617.20
2.10
613.09
621.31
Attending HE
61925
3
30937
OR FSM female to FSM
male
1.11
0.02
1.07
1.15
OR non-FSM female to
non-FSM male
1.12
0.01
1.11
1.14
OR female FSM to female
non-FSM
0.80
0.01
0.78
0.82
OR male FSM to male
non-FSM
0.81
0.01
0.79
0.83
Attending a Russell
Group HEI
61925
3
30937
OR FSM female to FSM
male
0.96
0.05
0.88
1.05
OR non-FSM female to
non-FSM male
0.83
0.01
0.81
0.84
OR female FSM to female
non-FSM
0.57
0.03
0.53
0.61
OR male FSM to male
non-FSM
0.49
0.03
0.46
0.52
Studying a SEM degree
61925
3
30937
OR FSM female to FSM
male
1.11
0.02
1.07
1.16
OR non-FSM female to
non-FSM male
1.11
0.01
1.10
1.13
OR female FSM to female
non-FSM
0.78
0.01
0.76
0.80
OR male FSM to male
non-FSM
0.78
0.02
0.75
0.81
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
55
Table 17 Post-16 routes and outcomes – interacting social background with
ethnicity
OR/beta
s.e.
Lower
bound
Upper
bound
Number
of
pupils
Number
of
groups
with
variation
Dropping out post-16
601503
18632
OR Black FSM to non-
FSM
1.29
0.25
1.16
1.43
OR Indian FSM to non-
FSM
1.52
0.42
1.19
1.94
OR Other FSM to non-
FSM
0.80
-0.23
0.71
0.90
OR Pakistani/Bangladeshi
FSM to non-FSM
1.02
0.02
0.93
1.13
OR White Other FSM to
non-FSM
1.22
0.20
1.11
1.33
OR White FSM to non-
FSM
1.48
0.39
1.44
1.52
Staying on in a sixth
form
601503
18632
OR Black FSM to non-
FSM
0.88
-0.13
0.83
0.93
OR Indian FSM to non-
FSM
0.61
-0.50
0.54
0.68
OR Other FSM to non-
FSM
1.06
0.06
0.98
1.14
OR Pakistani/Bangladeshi
FSM to non-FSM
0.83
-0.19
0.78
0.87
OR White Other FSM to
non-FSM
0.83
-0.19
0.78
0.88
OR White FSM to non-
FSM
0.64
-0.45
0.63
0.65
Distance to institution
(km)
524677
16370
Beta Black non-FSM
8.75
0.17
8.42
9.07
Beta Indian non-FSM
6.32
0.20
5.94
6.71
Beta Other non-FSM
9.99
0.18
9.63
10.35
Beta
Pakistani/Bangladeshi
non-FSM
6.03
0.18
5.68
6.39
Beta White Other non-FSM
8.96
0.14
8.68
9.24
Beta White non-FSM
8.99
0.04
8.91
9.08
Beta Black FSM
7.45
0.26
6.95
7.96
Beta Indian FSM
5.66
0.57
4.53
6.78
Beta Other FSM
6.65
0.38
5.91
7.39
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
56
Beta
Pakistani/Bangladeshi
FSM
5.13
0.25
4.63
5.62
Beta White Other FSM
7.96
0.32
7.34
8.58
Beta White FSM
8.33
0.11
8.12
8.54
Selectivity of institution
(mean GCSE grade)
524837
16378
Beta Black non-FSM
4.95
0.01
4.93
4.96
Beta Indian non-FSM
5.11
0.01
5.10
5.13
Beta Other non-FSM
4.94
0.01
4.93
4.96
Beta
Pakistani/Bangladeshi
non-FSM
4.99
0.01
4.98
5.01
Beta White Other non-FSM
4.86
0.01
4.85
4.87
Beta White non-FSM
4.83
0.00
4.82
4.84
Beta Black FSM
4.83
0.01
4.81
4.85
Beta Indian FSM
4.92
0.02
4.88
4.97
Beta Other FSM
4.80
0.02
4.77
4.83
Beta
Pakistani/Bangladeshi
FSM
4.87
0.01
4.85
4.89
Beta White Other FSM
4.74
0.01
4.71
4.76
Beta White FSM
4.65
0.01
4.64
4.66
Cohort size
524857
16379
Beta Black non-FSM
510.27
3.64
503.13
517.41
Beta Indian non-FSM
446.84
4.22
438.58
455.11
Beta Other non-FSM
503.36
3.88
495.76
510.97
Beta
Pakistani/Bangladeshi
non-FSM
465.69
3.96
457.92
473.45
Beta White Other non-FSM
523.86
3.08
517.82
529.89
Beta White non-FSM
528.23
1.42
525.45
531.02
Beta Black non-FSM
520.81
5.35
510.32
531.29
Beta Indian non-FSM
489.98
11.60
467.25
512.71
Beta Other non-FSM
508.43
7.72
493.31
523.55
Beta
Pakistani/Bangladeshi
non-FSM
493.75
5.40
483.17
504.33
Beta White Other non-FSM
537.08
6.51
524.33
549.84
Beta White non-FSM
537.02
2.48
532.17
541.88
Studying at least level 2
601503
18632
OR Black FSM to non-
FSM
0.87
-0.14
0.82
0.93
OR Indian FSM to non-
FSM
0.79
-0.23
0.68
0.93
OR Other FSM to non-
1.23
0.21
1.13
1.34
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
57
FSM
OR Pakistani/Bangladeshi
FSM to non-FSM
0.97
-0.03
0.91
1.03
OR White Other FSM to
non-FSM
0.84
-0.18
0.78
0.89
OR White FSM to non-
FSM
0.69
-0.38
0.67
0.70
Studying at least level 3
601503
18632
OR Black FSM to non-
FSM
0.88
-0.12
0.84
0.93
OR Indian FSM to non-
FSM
0.84
-0.17
0.75
0.95
OR Other FSM to non-
FSM
1.05
0.05
0.97
1.13
OR Pakistani/Bangladeshi
FSM to non-FSM
0.97
-0.03
0.92
1.03
OR White Other FSM to
non-FSM
0.86
-0.15
0.81
0.92
OR White FSM to non-
FSM
0.64
-0.45
0.63
0.65
Studying at least level 3
academic
601503
18632
OR Black FSM to non-
FSM
0.77
-0.26
0.73
0.82
OR Indian FSM to non-
FSM
0.84
-0.18
0.75
0.94
OR Other FSM to non-
FSM
0.91
-0.10
0.84
0.98
OR Pakistani/Bangladeshi
FSM to non-FSM
0.82
-0.20
0.77
0.86
OR White Other FSM to
non-FSM
0.78
-0.25
0.73
0.83
OR White FSM to non-
FSM
0.56
-0.57
0.55
0.58
Studying 2+ facilitating
A-levels
601503
18632
OR Black FSM to non-
FSM
1.10
0.09
1.02
1.18
OR Indian FSM to non-
FSM
0.76
-0.27
0.67
0.86
OR Other FSM to non-
FSM
1.09
0.09
0.99
1.19
OR Pakistani/Bangladeshi
FSM to non-FSM
0.70
-0.36
0.65
0.75
OR White Other FSM to
non-FSM
0.71
-0.34
0.65
0.79
OR White FSM to non-
FSM
0.53
-0.64
0.51
0.55
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
58
Studying 1+ SEM A-
levels
OR Black FSM to non-
FSM
1.07
0.07
1.00
1.14
OR Indian FSM to non-
FSM
0.71
-0.34
0.63
0.80
OR Other FSM to non-
FSM
1.16
0.15
1.07
1.27
OR Pakistani/Bangladeshi
FSM to non-FSM
0.71
-0.34
0.67
0.76
OR White Other FSM to
non-FSM
0.83
-0.19
0.76
0.90
OR White FSM to non-
FSM
0.56
-0.58
0.54
0.58
QCA points
395711
10708
Beta Black non-FSM
657.81
2.57
652.76
662.85
Beta Indian non-FSM
693.58
2.80
688.10
699.07
Beta Other non-FSM
672.63
2.73
667.28
677.97
Beta
Pakistani/Bangladeshi
non-FSM
653.45
2.77
648.02
658.88
Beta White Other non-FSM
661.37
2.20
657.05
665.68
Beta White non-FSM
640.04
1.03
638.03
642.05
Beta Black FSM
635.49
3.97
627.70
643.28
Beta Indian FSM
680.90
8.20
664.83
696.97
Beta Other FSM
632.76
5.82
621.34
644.17
Beta
Pakistani/Bangladeshi
FSM
640.07
3.98
632.26
647.88
Beta White Other FSM
624.72
5.34
614.27
635.18
Beta White FSM
584.61
2.16
580.39
588.84
Attending HE
601503
18632
OR Black FSM to non-
FSM
1.02
0.02
0.97
1.08
OR Indian FSM to non-
FSM
0.67
-0.40
0.60
0.75
OR Other FSM to non-
FSM
1.31
0.27
1.21
1.41
OR Pakistani/Bangladeshi
FSM to non-FSM
0.87
-0.14
0.82
0.92
OR White Other FSM to
non-FSM
0.95
-0.05
0.89
1.02
OR White FSM to non-
FSM
0.56
-0.58
0.55
0.58
Attending a Russell
Group HEI
601503
18632
OR Black FSM to non-
0.70
-0.35
0.61
0.81
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
59
FSM
OR Indian FSM to non-
FSM
0.77
-0.26
0.63
0.94
OR Other FSM to non-
FSM
0.73
-0.31
0.63
0.84
OR Pakistani/Bangladeshi
FSM to non-FSM
0.84
-0.18
0.75
0.94
OR White Other FSM to
non-FSM
0.49
-0.72
0.41
0.57
OR White FSM to non-
FSM
0.35
-1.05
0.32
0.38
Studying a SEM degree
601503
18632
OR Black FSM to non-
FSM
0.97
-0.03
0.91
1.03
OR Indian FSM to non-
FSM
0.69
-0.37
0.61
0.79
OR Other FSM to non-
FSM
1.26
0.23
1.15
1.38
OR Pakistani/Bangladeshi
FSM to non-FSM
0.87
-0.14
0.81
0.92
OR White Other FSM to
non-FSM
0.92
-0.08
0.85
1.00
OR White FSM to non-
FSM
0.54
-0.63
0.52
0.55
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
60
Table 18 Post-16 routes and outcomes – interacting gender with ethnicity
OR/bet
a
s.e.
Lower
bound
Uppe
r
boun
d
Numbe
r of
pupils
Number
of
groups
with
variatio
n
Dropping out post-16
601503
31471
OR Black female to male
0.69
-0.38
0.62
0.76
OR Indian female to male
1.01
0.01
0.85
1.22
OR Other female to male
0.86
-0.15
0.78
0.95
OR Pakistani/Bangladeshi
female to male
0.99
-0.01
0.90
1.08
OR White Other female to
male
0.89
-0.12
0.82
0.96
OR White female to male
0.86
-0.16
0.84
0.87
Staying on in a sixth form
601503
31471
OR Black female to male
1.19
0.17
1.13
1.25
OR Indian female to male
1.22
0.20
1.13
1.32
OR Other female to male
0.98
-0.02
0.92
1.04
OR Pakistani/Bangladeshi
female to male
1.42
0.35
1.34
1.49
OR White Other female to
male
0.96
-0.04
0.92
1.01
OR White female to male
1.01
0.01
1.00
1.02
Distance to institution
(km)
524677
28857
Beta Black male
8.65
0.20
8.26
9.04
Beta Indian male
6.39
0.25
5.90
6.89
Beta Other male
9.30
0.23
8.85
9.75
Beta Pakistani/Bangladeshi
male
6.06
0.21
5.65
6.46
Beta White Other male
9.06
0.18
8.70
9.42
Beta White male
9.22
0.05
9.11
9.32
Beta Black female
8.19
0.19
7.81
8.57
Beta Indian female
6.10
0.26
5.59
6.61
Beta Other female
9.43
0.24
8.96
9.90
Beta Pakistani/Bangladeshi
female
5.49
0.21
5.09
5.90
Beta White Other female
8.55
0.18
8.20
8.91
Beta White female
8.62
0.05
8.52
8.73
Selectivity of institution
(mean GCSE grade)
524837
28865
Beta Black male
4.90
0.01
4.89
4.92
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
61
Beta Indian male
5.09
0.01
5.07
5.11
Beta Other male
4.91
0.01
4.89
4.93
Beta Pakistani/Bangladeshi
male
4.96
0.01
4.94
4.98
Beta White Other male
4.84
0.01
4.82
4.86
Beta White male
4.81
0.00
4.81
4.82
Beta Black female
4.95
0.01
4.93
4.97
Beta Indian female
5.10
0.01
5.08
5.12
Beta Other female
4.92
0.01
4.90
4.94
Beta Pakistani/Bangladeshi
female
4.98
0.01
4.96
4.99
Beta White Other female
4.84
0.01
4.83
4.86
Beta White female
4.81
0.00
4.80
4.82
Cohort size
524857
28869
Beta Black male
501.37
4.25
493.04
509.7
1
Beta Indian male
454.12
5.26
443.80
464.4
4
Beta Other male
495.86
4.76
486.53
505.1
8
Beta Pakistani/Bangladeshi
male
499.58
4.45
490.87
508.2
9
Beta White Other male
512.82
3.86
505.25
520.3
8
Beta White male
515.31
1.53
512.31
518.3
2
Beta Black female
522.91
4.15
514.78
531.0
5
Beta Indian female
447.93
5.45
437.25
458.6
1
Beta Other female
513.47
4.90
503.86
523.0
8
Beta Pakistani/Bangladeshi
female
447.17
4.48
438.39
455.9
5
Beta White Other female
538.85
3.81
531.37
546.3
2
Beta White female
543.68
1.54
540.66
546.7
0
Studying at least level 2
601503
31471
OR Black female to male
1.68
0.52
1.57
1.80
OR Indian female to male
1.35
0.30
1.20
1.52
OR Other female to male
1.31
0.27
1.22
1.41
OR Pakistani/Bangladeshi
female to male
1.25
0.22
1.17
1.33
OR White Other female to
male
1.49
0.40
1.41
1.57
OR White female to male
1.74
0.55
1.72
1.76
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
62
Studying at least level 3
601503
31471
OR Black female to male
1.32
0.28
1.25
1.39
OR Indian female to male
1.23
0.20
1.12
1.34
OR Other female to male
1.07
0.07
1.01
1.14
OR Pakistani/Bangladeshi
female to male
0.98
-0.02
0.93
1.03
OR White Other female to
male
1.12
0.11
1.06
1.17
OR White female to male
1.30
0.26
1.28
1.32
Studying at least level 3
academic
601503
31471
OR Black female to male
1.37
0.32
1.30
1.44
OR Indian female to male
1.27
0.24
1.18
1.36
OR Other female to male
0.97
-0.03
0.91
1.02
OR Pakistani/Bangladeshi
female to male
1.13
0.12
1.08
1.19
OR White Other female to
male
1.09
0.08
1.04
1.14
OR White female to male
1.07
0.07
1.06
1.08
Studying 2+ facilitating A-
levels
601503
31471
OR Black female to male
0.84
-0.18
0.78
0.89
OR Indian female to male
0.71
-0.35
0.66
0.75
OR Other female to male
0.65
-0.43
0.61
0.70
OR Pakistani/Bangladeshi
female to male
0.69
-0.37
0.65
0.73
OR White Other female to
male
0.67
-0.40
0.63
0.71
OR White female to male
0.57
-0.55
0.57
0.58
Studying 1+ SEM A-levels
OR Black female to male
0.66
-0.42
0.62
0.70
OR Indian female to male
0.54
-0.61
0.51
0.58
OR Other female to male
0.57
-0.56
0.54
0.61
OR Pakistani/Bangladeshi
female to male
0.56
-0.58
0.53
0.59
OR White Other female to
male
0.47
-0.75
0.45
0.50
OR White female to male
0.48
-0.73
0.48
0.49
QCA points
395711
22645
Beta Black male
636.63
3.15
630.46
642.7
9
Beta Indian male
688.85
3.54
681.91
695.7
9
Beta Other male
662.59
3.47
655.79
669.4
0
Beta Pakistani/Bangladeshi
male
642.84
3.23
636.51
649.1
8
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
63
Beta White Other male
644.15
2.89
638.48
649.8
1
Beta White male
624.25
1.13
622.04
626.4
7
Beta Black female
668.33
2.89
662.67
673.9
9
Beta Indian female
695.96
3.57
688.95
702.9
6
Beta Other female
668.87
3.39
662.23
675.5
1
Beta Pakistani/Bangladeshi
female
659.30
3.11
653.20
665.4
0
Beta White Other female
667.36
2.70
662.07
672.6
6
Beta White female
645.68
1.12
643.49
647.8
7
Attending HE
601503
31471
OR Black female to male
1.39
0.33
1.33
1.47
OR Indian female to male
1.24
0.21
1.15
1.33
OR Other female to male
1.08
0.07
1.01
1.14
OR Pakistani/Bangladeshi
female to male
0.85
-0.16
0.81
0.90
OR White Other female to
male
1.17
0.16
1.12
1.23
OR White female to male
1.13
0.12
1.12
1.14
Attending a Russell Group
HEI
601503
31471
OR Black female to male
1.06
0.06
0.95
1.18
OR Indian female to male
0.62
-0.48
0.56
0.67
OR Other female to male
0.63
-0.47
0.57
0.68
OR Pakistani/Bangladeshi
female to male
0.82
-0.19
0.75
0.91
OR White Other female to
male
0.85
-0.16
0.79
0.92
OR White female to male
0.85
-0.16
0.83
0.87
Studying a SEM degree
601503
31471
OR Black female to male
1.31
0.27
1.24
1.39
OR Indian female to male
1.21
0.19
1.11
1.31
OR Other female to male
1.05
0.05
0.98
1.13
OR Pakistani/Bangladeshi
female to male
0.88
-0.13
0.82
0.93
OR White Other female to
male
1.14
0.13
1.07
1.20
OR White female to male
1.12
0.11
1.10
1.14
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
64
Appendix V – Post-16 routes and outcomes in London
Table 19 Post-16 routes and outcomes in London – interacting gender and
ethnicity
OR/beta
s.e.
Lower
bound
Upper
bound
Number
of
pupils
Number
of
groups
with
variation
Dropping out post-16
76285
6813
OR Black female to male
0.64
-0.44
0.54
0.76
OR Indian female to male
0.82
-0.20
0.66
1.02
OR Other female to male
0.83
-0.18
0.75
0.93
OR Pakistani/Bangladeshi
female to male
0.77
-0.26
0.69
0.86
OR White Other female to
male
0.85
-0.16
0.78
0.93
OR White female to male
0.84
-0.17
0.82
0.86
Staying on in a sixth form
76285
6813
OR Black female to male
1.18
0.17
1.09
1.29
OR Indian female to male
1.39
0.33
1.27
1.53
OR Other female to male
1.10
0.09
1.02
1.18
OR Pakistani/Bangladeshi
female to male
1.71
0.54
1.61
1.81
OR White Other female to
male
1.01
0.01
0.95
1.07
OR White female to male
1.02
0.02
1.01
1.03
Distance to institution
(km)
68991
6434
Beta Black male
6.96
0.23
6.51
7.42
Beta Indian male
4.69
0.39
3.92
5.46
Beta Other male
6.62
0.35
5.93
7.31
Beta Pakistani/Bangladeshi
male
5.07
0.36
4.37
5.77
Beta White Other male
6.90
0.29
6.33
7.47
Beta White male
7.73
0.17
7.39
8.06
Beta Black female
6.58
0.22
6.14
7.02
Beta Indian female
4.75
0.40
3.96
5.55
Beta Other female
5.71
0.37
4.99
6.43
Beta Pakistani/Bangladeshi
female
4.72
0.36
4.02
5.41
Beta White Other female
6.17
0.29
5.60
6.73
Beta White female
7.90
0.17
7.56
8.24
Selectivity of institution
(mean GCSE grade)
68991
6434
Beta Black male
4.88
0.01
4.85
4.91
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
65
Beta Indian male
5.04
0.02
5.01
5.08
Beta Other male
4.92
0.02
4.88
4.95
Beta Pakistani/Bangladeshi
male
4.86
0.02
4.82
4.89
Beta White Other male
4.82
0.02
4.79
4.86
Beta White male
4.87
0.01
4.84
4.89
Beta Black female
4.93
0.01
4.90
4.95
Beta Indian female
5.09
0.02
5.05
5.13
Beta Other female
4.92
0.02
4.89
4.96
Beta Pakistani/Bangladeshi
female
4.92
0.02
4.89
4.96
Beta White Other female
4.85
0.02
4.82
4.88
Beta White female
4.88
0.01
4.85
4.90
Cohort size
68994
6435
Beta Black male
408.56
4.83
399.10
418.03
Beta Indian male
331.71
7.83
316.36
347.06
Beta Other male
381.97
7.03
368.18
395.75
Beta Pakistani/Bangladeshi
male
412.96
7.20
398.86
427.07
Beta White Other male
406.30
5.87
394.80
417.80
Beta White male
377.78
3.73
370.47
385.09
Beta Black female
426.42
4.67
417.27
435.57
Beta Indian female
319.24
8.09
303.40
335.09
Beta Other female
385.24
7.33
370.87
399.61
Beta Pakistani/Bangladeshi
female
363.32
7.16
349.27
377.36
Beta White Other female
416.55
5.78
405.22
427.88
Beta White female
395.92
3.78
388.51
403.33
Studying at least level 2
76285
6813
OR Black female to male
1.70
0.53
1.53
1.89
OR Indian female to male
1.79
0.58
1.55
2.06
OR Other female to male
1.33
0.29
1.23
1.45
OR Pakistani/Bangladeshi
female to male
1.74
0.56
1.62
1.87
OR White Other female to
male
1.50
0.41
1.41
1.60
OR White female to male
1.67
0.51
1.65
1.69
Studying at least level 3
76285
6813
OR Black female to male
1.21
0.19
1.11
1.32
OR Indian female to male
1.30
0.26
1.17
1.44
OR Other female to male
1.11
0.10
1.03
1.19
OR Pakistani/Bangladeshi
female to male
1.22
0.20
1.15
1.30
OR White Other female to
male
1.12
0.12
1.06
1.19
OR White female to male
1.24
0.21
1.22
1.25
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
66
Studying at least level 3
academic
76285
6813
OR Black female to male
1.40
0.34
1.29
1.53
OR Indian female to male
1.34
0.29
1.23
1.46
OR Other female to male
0.97
-0.03
0.91
1.05
OR Pakistani/Bangladeshi
female to male
1.35
0.30
1.27
1.43
OR White Other female to
male
1.02
0.02
0.96
1.08
OR White female to male
1.06
0.06
1.05
1.07
Studying 2+ facilitating A-
levels
76285
6813
OR Black female to male
0.88
-0.12
0.79
0.99
OR Indian female to male
0.77
-0.26
0.71
0.84
OR Other female to male
0.73
-0.31
0.67
0.80
OR Pakistani/Bangladeshi
female to male
0.81
-0.21
0.75
0.87
OR White Other female to
male
0.77
-0.27
0.71
0.82
OR White female to male
0.68
-0.39
0.67
0.69
Studying 1+ SEM A-levels
OR Black female to male
0.67
-0.40
0.61
0.74
OR Indian female to male
0.57
-0.57
0.52
0.62
OR Other female to male
0.62
-0.47
0.57
0.67
OR Pakistani/Bangladeshi
female to male
0.57
-0.56
0.53
0.61
OR White Other female to
male
0.46
-0.77
0.43
0.50
OR White female to male
0.51
-0.68
0.50
0.52
QCA points
54415
5226
Beta Black male
615.02
4.20
606.78
623.25
Beta Indian male
666.03
6.08
654.10
677.95
Beta Other male
645.27
5.79
633.93
656.62
Beta Pakistani/Bangladeshi
male
620.63
5.87
609.12
632.14
Beta White Other male
631.04
5.05
621.14
640.93
Beta White male
613.56
3.38
606.94
620.18
Beta Black female
647.79
3.89
640.16
655.41
Beta Indian female
667.25
6.16
655.18
679.32
Beta Other female
657.49
5.78
646.16
668.83
Beta Pakistani/Bangladeshi
female
637.73
5.70
626.56
648.90
Beta White Other female
660.64
4.74
651.34
669.94
Beta White female
633.82
3.35
627.25
640.38
Attending HE
76285
6813
OR Black female to male
1.40
0.34
1.29
1.52
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
67
OR Indian female to male
1.35
0.30
1.23
1.47
OR Other female to male
1.19
0.18
1.11
1.28
OR Pakistani/Bangladeshi
female to male
1.06
0.05
0.99
1.12
OR White Other female to
male
1.27
0.24
1.20
1.35
OR White female to male
1.09
0.09
1.08
1.11
Attending a Russell Group
HEI
76285
6813
OR Black female to male
1.06
0.06
0.88
1.28
OR Indian female to male
0.68
-0.39
0.61
0.76
OR Other female to male
0.54
-0.62
0.48
0.60
OR Pakistani/Bangladeshi
female to male
0.72
-0.33
0.64
0.81
OR White Other female to
male
0.84
-0.18
0.76
0.92
OR White female to male
0.80
-0.22
0.79
0.82
Studying a SEM degree
76285
6813
OR Black female to male
1.31
0.27
1.19
1.45
OR Indian female to male
1.29
0.26
1.17
1.43
OR Other female to male
1.15
0.14
1.05
1.26
OR Pakistani/Bangladeshi
female to male
1.02
0.02
0.95
1.09
OR White Other female to
male
1.23
0.21
1.15
1.33
OR White female to male
1.06
0.05
1.04
1.07
Table 20 Post-16 routes and outcomes in London – interacting social
background and ethnicity
OR/bet
a
s.e.
Lower
boun
d
Upper
boun
d
Numbe
r of
pupils
Number
of groups
with
variation
Dropping out post-16
601503
18632
OR Black FSM to non-
FSM
1.29
0.25
1.16
1.43
OR Indian FSM to non-
FSM
1.52
0.42
1.19
1.94
OR Other FSM to non-
FSM
0.80
-0.23
0.71
0.90
OR Pakistani/Bangladeshi
FSM to non-FSM
1.02
0.02
0.93
1.13
OR White Other FSM to
1.22
0.20
1.11
1.33
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
68
non-FSM
OR White FSM to non-
FSM
1.48
0.39
1.44
1.52
Staying on in a sixth
form
601503
18632
OR Black FSM to non-
FSM
0.88
-0.13
0.83
0.93
OR Indian FSM to non-
FSM
0.61
-0.50
0.54
0.68
OR Other FSM to non-
FSM
1.06
0.06
0.98
1.14
OR Pakistani/Bangladeshi
FSM to non-FSM
0.83
-0.19
0.78
0.87
OR White Other FSM to
non-FSM
0.83
-0.19
0.78
0.88
OR White FSM to non-
FSM
0.64
-0.45
0.63
0.65
Distance to institution
(km)
524677
16370
Beta Black non-FSM
8.75
0.17
8.42
9.07
Beta Indian non-FSM
6.32
0.20
5.94
6.71
Beta Other non-FSM
9.99
0.18
9.63
10.35
Beta
Pakistani/Bangladeshi
non-FSM
6.03
0.18
5.68
6.39
Beta White Other non-
FSM
8.96
0.14
8.68
9.24
Beta White non-FSM
8.99
0.04
8.91
9.08
Beta Black FSM
7.45
0.26
6.95
7.96
Beta Indian FSM
5.66
0.57
4.53
6.78
Beta Other FSM
6.65
0.38
5.91
7.39
Beta
Pakistani/Bangladeshi
FSM
5.13
0.25
4.63
5.62
Beta White Other FSM
7.96
0.32
7.34
8.58
Beta White FSM
8.33
0.11
8.12
8.54
Selectivity of institution
(mean GCSE grade)
524837
16378
Beta Black non-FSM
4.95
0.01
4.93
4.96
Beta Indian non-FSM
5.11
0.01
5.10
5.13
Beta Other non-FSM
4.94
0.01
4.93
4.96
Beta
Pakistani/Bangladeshi
non-FSM
4.99
0.01
4.98
5.01
Beta White Other non-
FSM
4.86
0.01
4.85
4.87
Beta White non-FSM
4.83
0.00
4.82
4.84
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
69
Beta Black FSM
4.83
0.01
4.81
4.85
Beta Indian FSM
4.92
0.02
4.88
4.97
Beta Other FSM
4.80
0.02
4.77
4.83
Beta
Pakistani/Bangladeshi
FSM
4.87
0.01
4.85
4.89
Beta White Other FSM
4.74
0.01
4.71
4.76
Beta White FSM
4.65
0.01
4.64
4.66
Cohort size
524857
16379
Beta Black non-FSM
510.27
3.64
503.1
3
517.4
1
Beta Indian non-FSM
446.84
4.22
438.5
8
455.1
1
Beta Other non-FSM
503.36
3.88
495.7
6
510.9
7
Beta
Pakistani/Bangladeshi
non-FSM
465.69
3.96
457.9
2
473.4
5
Beta White Other non-
FSM
523.86
3.08
517.8
2
529.8
9
Beta White non-FSM
528.23
1.42
525.4
5
531.0
2
Beta Black FSM
520.81
5.35
510.3
2
531.2
9
Beta Indian FSM
489.98
11.6
0
467.2
5
512.7
1
Beta Other FSM
508.43
7.72
493.3
1
523.5
5
Beta
Pakistani/Bangladeshi
FSM
493.75
5.40
483.1
7
504.3
3
Beta White Other FSM
537.08
6.51
524.3
3
549.8
4
Beta White FSM
537.02
2.48
532.1
7
541.8
8
Studying at least level 2
601503
18632
OR Black FSM to non-
FSM
0.87
-0.14
0.82
0.93
OR Indian FSM to non-
FSM
0.79
-0.23
0.68
0.93
OR Other FSM to non-
FSM
1.23
0.21
1.13
1.34
OR Pakistani/Bangladeshi
FSM to non-FSM
0.97
-0.03
0.91
1.03
OR White Other FSM to
non-FSM
0.84
-0.18
0.78
0.89
OR White FSM to non-
FSM
0.69
-0.38
0.67
0.70
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
70
Studying at least level 3
601503
18632
OR Black FSM to non-
FSM
0.88
-0.12
0.84
0.93
OR Indian FSM to non-
FSM
0.84
-0.17
0.75
0.95
OR Other FSM to non-
FSM
1.05
0.05
0.97
1.13
OR Pakistani/Bangladeshi
FSM to non-FSM
0.97
-0.03
0.92
1.03
OR White Other FSM to
non-FSM
0.86
-0.15
0.81
0.92
OR White FSM to non-
FSM
0.64
-0.45
0.63
0.65
Studying at least level 3
academic
601503
18632
OR Black FSM to non-
FSM
0.77
-0.26
0.73
0.82
OR Indian FSM to non-
FSM
0.84
-0.18
0.75
0.94
OR Other FSM to non-
FSM
0.91
-0.10
0.84
0.98
OR Pakistani/Bangladeshi
FSM to non-FSM
0.82
-0.20
0.77
0.86
OR White Other FSM to
non-FSM
0.78
-0.25
0.73
0.83
OR White FSM to non-
FSM
0.56
-0.57
0.55
0.58
Studying 2+ facilitating
A-levels
601503
18632
OR Black FSM to non-
FSM
1.10
0.09
1.02
1.18
OR Indian FSM to non-
FSM
0.76
-0.27
0.67
0.86
OR Other FSM to non-
FSM
1.09
0.09
0.99
1.19
OR Pakistani/Bangladeshi
FSM to non-FSM
0.70
-0.36
0.65
0.75
OR White Other FSM to
non-FSM
0.71
-0.34
0.65
0.79
OR White FSM to non-
FSM
0.53
-0.64
0.51
0.55
Studying 1+ SEM A-
levels
OR Black FSM to non-
FSM
1.07
0.07
1.00
1.14
OR Indian FSM to non-
FSM
0.71
-0.34
0.63
0.80
OR Other FSM to non-
1.16
0.15
1.07
1.27
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
71
FSM
OR Pakistani/Bangladeshi
FSM to non-FSM
0.71
-0.34
0.67
0.76
OR White Other FSM to
non-FSM
0.83
-0.19
0.76
0.90
OR White FSM to non-
FSM
0.56
-0.58
0.54
0.58
QCA points
395711
10708
Beta Black non-FSM
657.81
2.57
652.7
6
662.8
5
Beta Indian non-FSM
693.58
2.80
688.1
0
699.0
7
Beta Other non-FSM
672.63
2.73
667.2
8
677.9
7
Beta
Pakistani/Bangladeshi
non-FSM
653.45
2.77
648.0
2
658.8
8
Beta White Other non-
FSM
661.37
2.20
657.0
5
665.6
8
Beta White non-FSM
640.04
1.03
638.0
3
642.0
5
Beta Black FSM
635.49
3.97
627.7
0
643.2
8
Beta Indian FSM
680.90
8.20
664.8
3
696.9
7
Beta Other FSM
632.76
5.82
621.3
4
644.1
7
Beta
Pakistani/Bangladeshi
FSM
640.07
3.98
632.2
6
647.8
8
Beta White Other FSM
624.72
5.34
614.2
7
635.1
8
Beta White FSM
584.61
2.16
580.3
9
588.8
4
Attending HE
601503
18632
OR Black FSM to non-
FSM
1.02
0.02
0.97
1.08
OR Indian FSM to non-
FSM
0.67
-0.40
0.60
0.75
OR Other FSM to non-
FSM
1.31
0.27
1.21
1.41
OR Pakistani/Bangladeshi
FSM to non-FSM
0.87
-0.14
0.82
0.92
OR White Other FSM to
non-FSM
0.95
-0.05
0.89
1.02
OR White FSM to non-
FSM
0.56
-0.58
0.55
0.58
Attending a Russell
601503
18632
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
72
Group HEI
OR Black FSM to non-
FSM
0.70
-0.35
0.61
0.81
OR Indian FSM to non-
FSM
0.77
-0.26
0.63
0.94
OR Other FSM to non-
FSM
0.73
-0.31
0.63
0.84
OR Pakistani/Bangladeshi
FSM to non-FSM
0.84
-0.18
0.75
0.94
OR White Other FSM to
non-FSM
0.49
-0.72
0.41
0.57
OR White FSM to non-
FSM
0.35
-1.05
0.32
0.38
Studying a SEM degree
601503
18632
OR Black FSM to non-
FSM
0.97
-0.03
0.91
1.03
OR Indian FSM to non-
FSM
0.69
-0.37
0.61
0.79
OR Other FSM to non-
FSM
1.26
0.23
1.15
1.38
OR Pakistani/Bangladeshi
FSM to non-FSM
0.87
-0.14
0.81
0.92
OR White Other FSM to
non-FSM
0.92
-0.08
0.85
1.00
OR White FSM to non-
FSM
0.54
-0.63
0.52
0.55
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
73
Appendix VI – Factor analysis of MSOA characteristics
Table 21 Factor analysis of MSOA characteristics - rotated factor loadings and
unique variance
MSOA characteristic
Factor 1
Factor 2
Factor 3
Uniqueness
MSOA population density
-0.13
0.91
-0.04
0.16
MSOA population density squared
-0.01
0.84
-0.02
0.30
Percentage of rented households
-0.50
0.68
-0.28
0.21
Percentage of social rented households
-0.59
0.43
-0.32
0.36
Percentage of adults with A-level 4 qualification
0.82
0.27
0.33
0.14
Percentage of unemployed adults
-0.65
0.44
-0.43
0.20
Percentage of employed adults
0.32
-0.20
0.96
-0.06
Percentage of full-time employed adults
0.28
-0.09
0.88
0.15
Percentage of UK-born adults
-0.05
-0.76
0.13
0.41
Percentage of adults belonging to social class I and
II
0.86
0.08
0.29
0.17
Percentage of pupils attending grammar schools
0.29
-0.08
0.02
0.91
Mean GCSE grade
0.93
-0.13
0.21
0.07
Mean GCSE grade squared
0.94
-0.13
0.19
0.07
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
74
Appendix VII – Pupil-level matching and regressions of impact
of living in an area with no school sixth forms
We perform a second matching approach to ensure the individual pupils living in
these areas without school sixth forms are matched as precisely as possible to
individual pupils with the same characteristics living elsewhere in England. Our
matching approach uses the individual pupil and local area variables listed in Table
22. In order to simplify the matching process, a factor analysis was carried out on the
MSOA-level variables. Thirteen MSOA characteristics were analysed and combined
into three factors using factor analysis. These thirteen local area characteristics and
their factor loadings can be seen in Table 21 in Appendix VI.
The three resulting factors, alongside the individual pupil variables, were used in the
matching process. An analysis of the quality of the matching process (a balancing
test) can be seen in Table 22. We can be more confident that the pupils in the
treatment area sample are similar to those they are matched to in the control area
sample when the bias percentage is less than five per cent and there are no
significant differences in the mean values of the variables between the treatment and
matched control samples. As can be seen in Table 22, the level of bias is slightly
high with regards to matching on the White British characteristic. There are also
various pupil and MSOA-level characteristics that vary significantly between the
treatment and control samples. However, although significant, the size of the
differences tends to be very small and so we are not overly concerned about these
significant differences. The low level of bias seen after matching is also reassuring.
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
75
Table 22 Quality of propensity score matching of pupils in treatment and
matched control areas
Variable
Treatment
mean
Control
mean
Significant
difference
between
treatment mean
and matched
mean
% Bias
Well balanced
(%bias < 5)
FSM
0.106
0.106
No
0.1
Yes
Female
0.488
0.486
No
0.5
Yes
Indian
0.006
0.009
Yes
-3.7
Yes
Other ethnicity
0.020
0.022
No
-1
Yes
Pakistani/Bangladeshi
0.024
0.026
Yes
-1.4
Yes
White Other
0.033
0.038
Yes
-2.6
Yes
White British
0.887
0.870
Yes
5.1
Marginal
Missing ethnicity
0.020
0.023
Yes
-1.9
Yes
School Action
0.128
0.128
No
0.1
Yes
School Action Plus
0.088
0.092
Yes
-1.7
Yes
Statement
0.036
0.036
No
-0.2
Yes
EAL
0.049
0.055
Yes
-2.7
Yes
Five GCSEs at A*-C
0.530
0.533
No
-0.6
Yes
Mean GCSE grade points
4.687
4.708
No
-1.2
Yes
Mean GCSE grade points
squared
24.953
25.133
No
-1.2
Yes
Number of GCSEs at A-A*
1.669
1.715
Yes
-1.6
Yes
MSOA characteristics
factor 1
-0.089
-0.076
No
-1.3
Yes
MSOA characteristics
factor 2
-0.206
-0.203
No
-0.6
Yes
MSOA characteristics
factor 3
0.178
0.163
Yes
1.3
Yes
North West
0.162
0.159
No
0.7
Yes
Yorkshire and the Humber
0.095
0.095
No
-0.2
Yes
East Midlands
No matches so excluded from the matched analysis
West Midlands
No matches so excluded from the matched analysis
East of England
0.105
0.104
No
0.2
Yes
London
No matches so excluded from the matched analysis
South East
0.425
0.430
No
-1
Yes
South West
0.150
0.153
No
-0.8
Yes
When we look at the distribution of pupils across school types in treatment and
matched control areas (Table 23), we can see that greater proportions of pupils
across all three attainment bands attend school sixth forms in matched areas than
treatment areas (as would be expected given that treatment areas are those without
school sixth forms). Attendance gradients are similar in treatment and matched
areas in that greater proportions of pupils attend some sort of institution as
attainment increases. Given the lack of school sixth form options in treatment areas,
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
76
there are greater proportions of pupils attending sixth form colleges and FE colleges
in treatment areas than matched areas.
Table 23 Distribution of low, middle and high attaining pupils across school
types in treatment and matched control areas
School sixth form
Sixth form college
FE college
Treatment
Matched
Treatment
Matched
Treatment
Matched
Low attainers
0.5%
12.7%
12.1%
3.4%
60.6%
54.3%
Middle attainers
1.6%
38.6%
43.8%
14.2%
45.1%
38.0%
High attainers
2.8%
60.3%
69.6%
20.0%
16.8%
9.7%
Pupil-level regressions
The tables below show the results of the pupil-level matched analysis. Each set of
results estimates the effect of living in areas with no school sixth forms (treatment
areas) on the likelihood of pupils achieving each of our post-16 outcomes compared
to matched pupils who live in areas with school sixth forms. We first look at the
results for the full matched sample before breaking the sample down according to
various pupil characteristics. Table 24 shows the sample sizes for each analysis.
Whilst the full pupil sample is sufficiently large, some pupil sub-samples are quite a
lot smaller (such as the Black and Indian pupil samples) and so we are more
cautious about the findings here, particularly when we take into account the non-
random geographical concentration of ethnic groups across England.
Table 24 Sample sizes for pupil-level matched regressions
Pupil sub-group
Number of pupils
Number of MSOAs
All pupils
85,834
4,469
Low attainers
28,846
3,880
Middle attainers
28,996
4,053
High attainers
27,992
3,978
Non-FSM pupils
76,739
4,456
FSM pupils
9,095
2,406
Male pupils
44,043
4,345
Female pupils
41,791
4,311
Black pupils
971
532
Indian pupils
636
386
Other ethnicity pupils
1,819
1,032
Pakistani/Bangladeshi pupils
2,135
593
White Other
3,019
1,432
White British
75,405
4,423
Missing ethnicity pupils
1,849
1,016
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
77
The results for all attainers in Table 25 show us that pupils living in areas with no
school sixth forms are significantly less likely to be studying for at least level 2 (11%
lower odds), at least level 3 (13% lower odds) and at least level 3 academic (15%
lower odds) qualifications. They are also less likely to be studying for at least two
facilitating A-levels and at least one SEM A-level. The pattern of results is
reasonably similar for low, middle and high attainers.
Table 25 Effects of living in an area with no school sixth forms on post-16
participation by prior attainment (pupil matching)
Effect of living in an area with no
school sixth forms on the likelihood
of a pupil…
All attainers
Low
attainers
Middle
attainers
High
attainers
OR
(s.e.)
OR
(s.e.)
OR
(s.e.)
OR
(s.e.)
Participating in post-16 education
1.04
0.03
1.09
0.04
0.93
0.05
1.00
0.10
Studying at least level 2
0.89
0.03
0.85
0.03
0.85
0.03
0.90
0.07
Studying at least level 3
0.87
0.03
0.65
0.03
0.81
0.03
0.85
0.06
Studying at least level 3 academic
0.85
0.03
0.47
0.05
0.76
0.03
0.77
0.04
Studying at least 2 facilitating A-levels
0.82
0.03
0.42
0.15
0.66
0.03
0.83
0.03
Studying at least 1 SEM A-level
0.84
0.03
0.31
0.06
0.67
0.03
0.91
0.03
Attending HE
0.95
0.03
0.91
0.06
1.03
0.04
0.87
0.03
Attending a Russell Group HEI
0.93
0.05
1.00
0.36
1.21
0.14
0.93
0.04
Studying a SEM degree
0.97
0.03
0.98
0.12
1.07
0.05
0.96
0.03
Note: Models were estimated on the following numbers of pupils: all attainers (85,834), low attainers
(28,846), middle attainers (28,996), high attainers (27,992).
Turning to the effects of living in an area with no school sixth forms on demographic
sub-samples of pupils, Table 26 shows there are few differences in the magnitude of
the estimates for FSM and non-FSM pupils. Some of the estimates on the FSM
sample are not statistically significant because the sample is so much smaller.
The findings by gender show bigger gaps for males than females. For example,
males in treatment areas have 18 per cent lower odds of studying for a level 3
academic qualification than matched males in control areas, whilst females in
treatment areas have only 12 per cent lower odds of following this route than are
matched females. Although the pattern of findings is generally similar for males and
females, one area of divergence occurs when looking at the likelihood of attending a
Russell Group institution. Here, there are no significant differences between females
in treatment and matched control areas, but males in treatment areas have lower
odds than males in control areas of attending a Russell Group HEI. This suggests
that limiting access to school sixth forms is not just to the disadvantage of males
during the post-16 phase, but also in terms of access to higher status universities.
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
78
Table 26 Effects of living in an area with no school sixth forms on post-16
participation by social background and gender (pupil matching)
Effect of living in an area with no
school sixth forms on the likelihood
of a pupil…
Non-FSM
FSM
Males
Females
OR
(s.e.)
OR
(s.e.)
OR
(s.e.)
OR
(s.e.)
Participating in post-16 education
1.05
0.04
0.99
0.06
1.03
0.04
1.05
0.29
Studying at least level 2
0.89
0.03
0.87
0.05
0.87
0.03
0.92
0.09
Studying at least level 3
0.87
0.03
0.82
0.05
0.84
0.03
0.90
0.04
Studying at least level 3 academic
0.84
0.03
0.98
0.07
0.82
0.03
0.88
0.02
Studying at least 2 facilitating A-levels
0.82
0.03
0.88
0.10
0.82
0.03
0.83
0.01
Studying at least 1 SEM A-level
0.84
0.03
0.79
0.08
0.83
0.03
0.85
0.01
Attending HE
0.94
0.03
1.03
0.08
0.93
0.04
0.97
0.01
Attending a Russell Group HEI
0.93
0.05
0.84
0.16
0.86
0.05
1.00
0.00
Studying a SEM degree
0.98
0.03
0.88
0.10
0.98
0.04
0.97
0.00
Note: Models were estimated on the following numbers of pupils: non-FSM (76,739), FSM (9,095),
males (44,043), females (41,791).
Table 27 repeats the analysis for four ethnic group sub-samples. As would be
expected, the results for the White British group closely follow those for the sample
of all attainers. The sub-sample of Black pupils is small (only 971), but our results
indicate that this group is especially disadvantaged in terms of participation in level 3
study. Given that we do not see similar findings for the Indian sub-sample and the
Pakistani/Bangladeshi sub-sample, the lack of access to school sixth forms seems
particularly disadvantageous for Black pupils.
We only see positive effects of living in an area with no school sixth forms on the
likelihood of attending higher education for Indian pupils and Pakistani/Bangladeshi
pupils (35 per cent more likely). Interestingly, these two ethnic groupings do not
experience significantly lower likelihoods of following level 2, level 3 and level 3
academic routes in treatment areas, unlike all the other sub-samples investigated
here. The lack of significant differences could be because of small sample sizes,
however.
Social Mobility and Child Poverty Commission
Social and ethnic inequalities in choice available and choices made at age 16
79
Table 27 Effects of living in an area with no school sixth forms on post-16
participation by ethnicity (pupil matching)
Effect of living in an area with no
school sixth forms on the
likelihood of a pupil…
White
British
Black
Indian
Pakistani/
Bangladeshi
OR
(s.e.)
OR
(s.e.)
OR
(s.e.)
OR
(s.e.)
Participating in post-16 education
1.04
0.04
1.55
0.51
3.24
2.29
0.82
0.15
Studying at least level 2
0.89
0.03
0.83
0.14
1.10
0.38
0.94
0.11
Studying at least level 3
0.87
0.03
0.69
0.09
1.24
0.31
1.05
0.11
Studying at least level 3 academic
0.84
0.03
0.69
0.10
1.23
0.24
1.09
0.12
Studying at least 2 facilitating A-
levels
0.82
0.03
0.56
0.10
1.10
0.20
1.01
0.13
Studying at least 1 SEM A-level
0.84
0.03
0.56
0.09
1.07
0.20
1.14
0.13
Attending HE
0.95
0.03
0.99
0.14
1.60
0.30
1.35
0.13
Attending a Russell Group HEI
0.93
0.05
0.64
0.22
1.40
0.31
0.72
0.16
Studying a SEM degree
0.98
0.03
0.76
0.13
1.08
0.19
1.15
0.14
Note: Models were estimated on the following numbers of pupils: White British (75,405), Black (971),
Indian (636), Pakistani/Bangladeshi (2,135).
Interpretation of the magnitude of findings
In this section we interpret the magnitude of the findings by estimating how many
pupils miss out on each of our post-16 routes and outcomes because of the school
sixth form provision available to them. We base our estimates on the number of
pupils living in the treatment (no school sixth forms) used in our earlier analysis. We
further split our estimates by pupil social background. Table 28 shows the results of
our interpretation of the magnitude of the findings.
As can be seen in the analysis of all attainers, pupils tend to miss out when they live
in treatment areas i.e. areas with no school sixth forms. For example, we estimate
that 1,713 pupils miss out on studying for level 3 academic qualification because of a
lack of school sixth form provision in their local areas, whilst 502 pupils miss out on
university attendance.
Table 28 Estimated number of pupils missing out in areas with no school sixth
forms
All pupils
Non-FSM
FSM
Studying at least level 2
927
761
163
Studying at least level 3
1,422
1,220
200
Studying at least level 3 academic
1,713
1,696
Studying at least 2 facilitating A-levels
1,279
1,247
Studying at least 1 SEM A-level
1,263
1,181
80
Note: The “all attainers” estimates are based on a sample of 42,949 pupils in treatment (no school
sixth form) areas. The non-FSM estimates are based on a sample of 38,387 pupils in treatment (no
school sixth form) areas. The FSM estimates are based on a sample of 4,562 pupils in treatment (no
school sixth form) areas.
1
© Crown copyright 2016
Any enquiries regarding this publication should be sent to us at the Social Mobility
Commission, Sanctuary Buildings, 20 Great Smith Street, London, SW1P 3BT.
Email: contact@smcpcommission.gsi.gov.uk
This document is also available from our website at:
www.gov.uk/government/organisations/social-mobility-commission
