Transcript Cycles of Disadvantage: Poverty, Education, Inequalities and Social

Cycles of Disadvantage: Poverty, Education,
Inequalities and Social Exclusion
Andy Green
Professor of Comparative social Science and director of the Centre for
Learning and Life Chances in Knowledge Economies and Societies
(LLAKES), Institute of Education
NESSE Seminar, Brussels, May 20, 2010
The Persistence of Educational Inequality
• Major inequalities have mired education since national
systems were first created in the early nineteenth century.
• Even as participation in each phase of education - from
elementary, to secondary, then upper secondary - became
universalized, according to the Enlightenment dream,
divisions, by social class, ethnicity, religion and gender,
remained.
• When a certain phase became less segmented, so the frontier
of educational selection, as Boudon remarked, moved forward.
So that today it is at upper secondary and tertiary level where
the final sorting of students occurs and future life opportunities
are decided.
Plus Ça Change
• Since the Coleman report on ‘Equality of Educational Opportunity’ in the US
(1966), countless studies have found that social background is the main determinant
of achievement in education. Social class has been the overriding factor, although
ethnicity and gender have also been linked to under-achievement, but the picture
here is much more complex.
•
A century of reforms have ensured that all social groups get more education than
before but the relative chances of children from different social groups achieving at
the highest level has not changed dramatically.
• Studies of comprehensivization in a range of countries (Shavit and Blossfeld, 1983)
found little evidence of reductions in the social gaps in achievement, although PISA
suggests a more positive interpretation in recent years.
• The main beneficiaries of university post-1960s expansion have been the middle
classes, and particularly their daughters, rather than children of the less skilled
working class families.
Educational Stratification Occurs through Effects
at Different Levels
• Family background impacts on educational achievement 1) directly
through family socialisation and 2) indirectly through its influence on
the choice of school.
• Schools play a part in stratification through 1) ‘pure school effects’
(some schools are better resourced and provide better teaching than
others) and 2) ‘school peer effects’ (which in many countries have a
greater impact on individual student outcomes than direct background
influences) (Mostafa, 2009).
• School System effects: the design of a national education system also
affects the degree of stratification in education.
Terms
• Social inequality in education refers to the
inequalities between social groups (class, ethnicity or
gender) in educational opportunities, treatment or
outcomes.
• Unequal opportunities occur as a result of factors
external to the school which impact on a child’s
opportunity to access educational programmes and to
learn and achieve on them.
• Unequal treatment refers to differences in educational
experience encountered in EIs (and so include school
and peer effects).
Structure of the Workshop
• Evidence of inequality through the life course and
lifelong learning system.
• Theories explaining the effects of poverty and social
class on educational achievement.
• Evidence across countries on system effects on levels
of inequality.
• The impact of education on social and economic
outcomes for individuals.
• The impact of educational inequality on economic
and social outcomes at the country level.
Educational Inequality through the Life Course
• ‘Education is linked to various dimensions of social
disadvantage throughout an individual’s life span.
This includes their time in pre-primary education, in
the compulsory schooling system during their
childhood years, as young adults in post-compulsory
education and during the years of adulthood.’
• Stephen Machin (2006)
Early and Enduring Influence of Social Class on
Children’s Cognitive Test Scores.
•
Feinstein (2003) uses data from the British Cohort Study for children born
in 1970 and assessed at 22 and 42 months and at years: 5 , 10, 16, 26 and
30 (achieved sample: 11 200).
• Uses Principle Components Analysis to develop single index of
development at each stage.
• Scores at 22 months are good predictor of educational attainment at 26.
• 15% of children in the bottom quartile of cognitive scores at 22 months
achieved no general academic qualifications by age 26 compared to only
8% of those in top quartile. (Significant = P˂0.01).
• A social class differential already evident at 22 months. The gap in scores
between the top and bottom SES groups is already 13 percentage points.
Increasing impact of Social Class through Childhood
• High SES children scoring in the top quartile for cognitive
development at 22 months continue to do well aged 10,
scoring on average just under the 70th percentile.
• By contrast, the low SES group scoring in the top quartile at
22 months experience a relative decline, scoring on average at
the 40th percentile aged 10.
• The low SES group in lowest quartile at 22 months scored on
average at the 27th percentile at aged 10, compared to average
scores at the 58th percentile for the high SES children in lowest
quartile at 22 months.
• High SES children scoring in lowest quartile at 22 months
have overtaken low SES top quartile group at 22 months by
aged 10.
Relative Cognitive Progress Of High SES Children (Blue) and
Low SES Children (Red) between 22 Months and 10 years in the
UK
Effects of Poverty on Early Learning
• Research by Goodman and Gregg (2010) using longitudinal data for
the UK shows that the gap in attainment between children from the
poorest fifth and the richest fifth is already large by the age of 5,
grows more rapidly during the primary school years and increases at a
lower pace during the secondary years. They explain the gap in terms
of attitudes.
• By age 11 only around three quarters of children from the poorest fifth
of families reach the UK Gov’t’s expected level at KS2 compared to
97% of the richest fifth.
•
By the end of lower secondary education only 21% of poorest fifth of
children get 5 A-C grades including Maths and English, compared
with 75% of those from the richest fifth.
Performance of Immigrant Students in PISA
•
•
•
•
We do not have data on the average performance of different ethnic groups
across a range of countries. PISA tells us only about 1st and 2nd generation
immigrant children.
Across OECD countries there are a higher proportion of top performers
amongst native students than amongst those with an immigrant
background, but this partly reflects different in socio-economic
backgrounds.
In half of the countries, if you control for student SES, the differences are
no longer significant.
Relative performance of native and immigrant students varies from country
to country depending on various factors, including immigration policies.
Countries where immigrants do as well or better than natives include:
Australia, Canada, Greece, Ireland, New Zealand, Norway, and Portugal.
Estimated Effects of Family Background of Students’ Test scores Across
Countries
Youths Participating in Tertiary Education by
Educational Attainments of their Parents, 1994-95
Participation Rate in Non-Formal Job-Related Continuing Education and Training for
the Labour Force 25-64 Years, by Level of Educational Attainment, 2003
Explanation of the Links between Poverty and
Educational Attainment.
• Micro level explanations focus on the characteristics of
children and their families, pointing to differences in child
health, parenting styles, family material and cultural resources,
and attitudes towards learning.
• Meso level explanations focus on the neighbourhood and the
school, focussing on the negative effects of neighbourhoods
with high unemployment and low social capital on aspirations
and positive attitudes towards the school. Schools with high
concentrations of disadvantaged children may exercise
negative peer effects on achievement.
• Macro-level explanations focus on the way in which school
systems reproduce social inequality and systematically
disadvantage children from poor families (through poorly
resourced schools , low aspirations, stereotyping of students
Levels of Explanation
Raffo, Dyson, Gunter, Hall, Jones and Kallambouka (2007)
distinguish between ‘functionalist’ and ‘socially critical’ theories,
each of which operate at three levels:
• Micro
• Meso
• Macro
The Importance of Values, Attitudes and Attitudes.
Research on educational disadvantage has focussed on the importance child and
parental aspirations since Douglas (1964) identified the importance of the mother’s
educational aspirations for her children. Goodman and Gregg (2010) in their recent
report on poverty and achievement in the UK follow in this tradition.
•
81% of the richest fifth of mothers hoped their 9 year old children would go to
university compared to 37% of the poorest mothers.
•
The authors claim this factor explains 6% of the attainment gap by 11
•
Children of poorer families tend to see themselves as less scholastically able, are
less concerned with school results and exhibit higher levels of hyperactivity.
•
The differences in child attitudes and behaviours, according to the authors, account
for 12% of the attainment gap at 11 years.
•
25% of the test score gap is associated with factors present across the generations
(like not being read to as a child).
Goodman and Gregg
Over-Arching Theories
There are various over-arching theories about how social
inequalities are reproduced through Education. The most
influential are currently those of:
• Pierre Bourdieu
• Raymond Boudon
Pierre Bourdieu and Social Reproduction
• Bourdieu (1986) distinguished between three types of capital which
are unequally distributed across families of different social classes: 1)
economic capital; 2) social capital and 3) cultural capital.
• Cultural capital refers to the forms of knowledge, skills, education,
and advantages that a person has, which give them a higher status in
society. Habitus refers to acquired sensibilities, disposition and taste.
• Parents provide their children with cultural capital and a specific
habitus which correspond to what is demanded in school and allows
them to succeed in the current educational system (Bourdieu and
Passeron, 1973).
• The theory was developed through research on the French education
system but is offered as a relatively universal theory. It ignores
varying ways in which different systems reproduce inequality.
Raymond Boudon and ‘Positional
Theory’
Boudon (1974), a radical rational choice theorist, argues that family
backgrounds also influence educational achievement through the way in
which they affect choices made by parents and students. Depending on
where they are positioned socially students are likely to make different
calculations about the relative costs and benefits of selecting the most
prestigious academic routes in education which will affect their
educational achievements.
Educational systems which offer more ‘branching points’ (opportunities
for academic selection) are likely to allow social background to have
more influence the achievement of individual students through the
choices they themselves make.
Student Choices in France
Duru-Bellat and Mingat’s (1999) analysis of streaming at the
end of lower secondary education in 1990 in France well
illustrates how this works.
Looking at transitions to different types of lycée they show
that:
• 25 per cent of the social class differences in transitions
to high or low streams could be attributed to ability, but
• a further 25 per cent derived from the choices made by parents and
their children which are unrelated to achievement.
School System Effects
How comparing across education systems can help to
explain what system characteristics impact on the level
of inequalities in educational outcomes in different
countries.
Average Variation in Science by Country Group in PISA
2006
130
124.55
113.55
110
106.85
99.8
94.8
94.575
Nordic
S. Europe
90
70
50
US, UK
Anglo
Germanic
E. Asia
Percentage of Within Country Variance in Science Scores Explained
by PISA ESCS Index by Average for Country Group
20
17.98
15.9
15
13.87
13.4
9.6
10
7.75
5
0
Germanic
US, UK
S. Europe
Anglo
Nordic
E. Asia
Score Point Difference Associated with One Unit on ESCS - Social
Gradient
50
48.5
46.2
44.83
40
35.5
34.6
30
31.75
20
10
0
US, UK
Anglo
Germanic
E. Asia
Nordic
S.Europe
Between School and Within School Variation
150
100
50
0
US, UK
Anglo
Germanic
E. Asia
Betweeen Schools Within Schools
Nordic
S. Europe
Typical Range of Average Socio-Economic Status of
Schools by Average for Country Group
0.8
0.6
0.4
0.2
0
US, UK
Anglo
Germanic E. Asia
Nordic
S. Europe
Nordic Countries – Low Variation
• Nordic countries, on average, have the smallest variation in individual
scores of any country group (along with southern Europe).
• Small differences between schools in social intake
• Small differences between schools in average performance
• School peer effects are low
• Even in Nordic countries where individual background impacts
substantially on performance, since the impact of the school is low, total
variation between students remains quite low.
• Norway, for instance, has a steeper social gradient than Spain, meaning
that differences in social intake would predict larger differences in school
performance than in Spain. However, in fact, schools have very similar
social intakes in Norway so the overall differences between schools in
mean scores remain very low and lower than in Spain, which has very high
differences in social intakes between schools.
Countries with Selective Systems:
Germany, Austria and Belgium
• Highest between-school variation in average performance
• Large differences in social intake of school
• School peer effects are strong
• Overall variation in student performance is high, as is the
impact of social background on individual performance,
measured both by the percentage of variance explained
and the social gradient.
Countries with Selective Systems:
Germany, Austria and Belgium
• Highest between-school variation in average performance
• Large differences in social intake of school
• School peer effects are strong
• Overall variation in student performance is high, as is the
impact of social background on individual performance,
measured both by the percentage of variance explained
and the social gradient.
UK and USA – Highest variation and
social gradients
•
Total variance between individual scores is very high in both countries because the
education systems combine very high within-school variation with average levels
of between-school variation (whereas most comprehensive systems have high
within-school variation and low between-school variation).
•
The differences between schools in terms of social intake are high for
‘comprehensive systems’ (although average for OECD countries as a whole)
•
The impacts of individual background and the school peer group effects are above
average in both countries
•
These combine to produce the very high social gradients.
•
Comprehensive schools in the US and the UK would appear to be both
substantially more socially selective than those in the Nordic countries and more
internally differentiated.
Japan and Korea
• Combine high between-school variation in performance, with quite low
within-school variation.
• Differences between schools in social intakes are slightly below average.
• Social background has relatively little impact on an individuals’
performance but there are strong school peer group effects.
• Schools do not vary particularly highly in their social intakes, but small
differences in the average social composition of the peer group seems to
have a large effect on individual performance. One explanation of this may
be the strong normative effect of pedagogies in Japanese and Korean
schools which tend to standardise aspirations across student cohorts and
thus enhance peer group effects.
Explaining Differences between
Countries in levels of Inequality
Different levels of inequality could be explained by
differences in national socio-economic contexts or by
differences in education systems . Three main
explanations by context:
• Effects of income inequality
• Political ideologies
• Effects of welfare systems
Impact of Income Inequality
Clearly where countries have greater variations in income and wealth
there is a likelihood that the rich will buy advantages for their children
through private schools, private tutoring etc.
Given the embeddedness of cultural capital in economic relations, it is
also likely that Bourdieu’s ‘cultural capital’ will be more unevenly
distributed amongst families thus probably making a more unequal
distribution of learning outcomes.
In fact there is no clear correlation between income inequality (Gini
measure) and inequalities of student achievement at 15 yrs (although
there is a clear correlation between adult skills distribution and income
distribution) (OECD)
Incom e Ine quality
45
40
35
Gini
30
25
20
15
10
5
0
Nordic
Core
Europe
Southern
Europe
Re gion
AngloSaxon
Welfare System Effects
Esping -Andersen argues that the more equal outcomes
of Nordic countries has nothing to do with their
schooling systems.
He attributes it entirely to the effects of universalist
welfare systems that include ample social support for
parents, and particularly lone parents, and where all
children are entitled to full-time pre-school education
from infancy.
Welfare 2
• Universal and largely free pre school education ensures early
socialisation of the children away from the home and thus
mitigates the effects on learning of unequal social capital in
the home.
• Generous social support and family – friendly employment
policies encourage the vast majority of mother to work.
• Lone mothers are generally not poor these countries which
means that the learning disadvantages of children of lone
parents resulting from low family income in countries like
England are not apparent in Nordic Countries
Political Ideologies
It can be argued that creating egalitarian educational systems requires
considerable political will and a strong consensus favouring equality. This
is most likely to be achieved in countries like the Nordic states which are
social cohesive and have strong social democratic ideologies.
This could provide an alternative explanation of the relative egalitarianism
of the East Asian (quite cohesive) and Nordic states. It would also suggest
reasons why less cohesive states like the UK and USA find it harder to
achieve consensus around such controversial issues.
However, the explanation do not work so well for the southern European
states which are neither strongly social democratic nor particularly
cohesive.
The Country Clustering
The most equal countries are in three groups:
• Nordics
• North East Asian
• Southern European
The most unequal countries are in two group:
• English-Speaking (and particularly USA and UK)
• ‘Germanic’ : Germany, Austria, Switzerland, Belgium,
Luxembourg.
Decentralisation
One common factor shared by all the unequal systems is that have a
degree of decentralisation to regional levels. They are either effectively
federal systems (USA, UK, Germany) or give strong powers to individual
regions or cantons. This may lead to the unequal allocation of resources
and thus to educational inequality.
The more equal countries do not generally have federal systems (although
Spain in southern Europe comes close to this.)
OECD statistical analysis shows that there is a strong correlation between
degree of regional decentralisation and inequality of outcomes
Decentralisation
One common factor shared by all the unequal systems is that have a
degree of decentralisation to regional levels. They are either effectively
federal systems (USA, UK, Germany) or give strong powers to individual
regions or cantons. This may lead to the unequal allocation of resources
and thus to educational inequality.
The more equal countries do not generally have federal systems (although
Spain in southern Europe comes close to this.)
OECD statistical analysis shows that there is a strong correlation between
degree of regional decentralisation and inequality of outcomes
Differences between Comprehensive
Systems
If selective secondary school organisation is associated with
outcomes inequality, and comprehensive organisation with
more equal outcomes, how then do we explain the
differences between comprehensive systems in terms no
inequality?
One strategy is to identify different characteristics within the
structures of comprehensive schools which may be conducive
to equality and to see if we can identify different forms of
comprehensive school systems.
Nordic and East Asian Countries
These all had (in the relevant reference period) a number of
characteristics which may be seen to favour low levels of
inequality.:
•
•
•
•
Mixed ability teaching throughout
Low levels of school diversity
Low social segregation in intakes
Low levels of school choosing (most children going to their
local school) – encouraged in Nordic countries by the unique
system of all-through comprehensive schools where few
children change school in transition to lower secondary.
Features of Quasi Market
Comprehensive Systems
• High level of school diversity within the state
system (specialist schools, Academies, Faith
schools etc in England.
• Strong promotion of school choice
• Culture of league tables and inter-school
competition
• High level of in setting and streaming in
schools
Branching Points
What is common to the ineqalitarian systems, and more absent in the
egalitarian systems, is – to use the generic American term – tracking.
The Germanic countries track through formal selection to different types of
school.
The English-speaking systems track through ability grouping in schools and
through more subtle forms of selection to different types of school (which are
inevitably hierarchically ranked.)
Tracking creates more ‘branches points’ and it is the proliferation of these
which Raymond Boudon claims makes some education systems more
inegalitarian than others.
Individual Economic Benefits
Earnings Differential by Educational Attainment in 2002
Education Earnings Premium
• Earnings increase with each level of education.
• The earnings premium for tertiary education is substantial in
most countries and exceeds 50% in 17 out of 28 OECD
countries.
• Males with Type A tertiary education earn on average 80%
more than those with only upper secondary or post-secondary
mom-tertiary in Hungary, Czech Republic, Poland, Portugal,
Israel and the USA.
• The education earnings gap increases with age in most
countries.
Employment Rates by Educational
Attainment
Individual Social Benefits
Social Benefits
Research in various countries has show that higher levels of
education are associated with a variety of social benefits to the
individual including:
• Better health (Taubman and Rosen, 1982; Feinstein et al,
2003)
• Reduced propensity to crime (Behrman and Stacey, 1996)
• Reduced risk of teenage pregnancy and drug abuse (Bynner
and Parson, 1998)
• Increased social capital (trust, civic association and political
engagement (Putnam, 2000)
Macro-Social Effects
Macro Social Benefits
Cross-country studies show that higher average levels of
education are associated with:
• Lower levels of violent crime (McMahon, 1999)
• Improved human rights, political stability and
democratization( McMahon, 1999)
• Greater income equality (Mcmahon, 1999)
• Reduced family size (Sen, 1999)
• Increased social capital (Putnam, 2000)
More equal distributions of skills and qualifications are also
associated with wider social benefits (Green, Janmaat and
Test score ratio and income inequality
1.60
1.50
USA
POR
Income inequality
1.40
CAN
1.30
PO
SZ
IRL
B
NW
AU
1.20
FIN
SW
DEN
UK
NL
D
1.10
1.00
20.00
25.00
30.00
35.00
Test score ratio
40.00
45.00
Social Cohesion Index
Figure 2. Relationship between Social Cohesion and Income Inequality
4
Nw
3
2
De
1
Ne
Swt
Be
0
Ir
Ca
Swe Fi
Au
Pol
UK
-1
US
Por
Ge
-2
-3
20
25
30
35
100-Gini Coefficient (100=total inequality)
40
45
Educational Inequality and Social
Cohesion
Social Mobility
The Centre for Learning and Life
Chances in Knowledge Economies
and Societies (LLAKES) is an research
Centre funded by the UK Economic
and Social Research Council. Details
can be found at:
Llakes.org