Marginalization in education

From research to policy

Kevin Watkins Senior Fellow Brookings Institution

Why it matters

• Highly predictable failure rooted in structural disadvantage - ‘clearly remediable injustice’ linked to poverty, gender, ethnicity, disability, location and wider markers for disadvantage

• A brake on progress towards the EFA goals on access + learning

• Intrinsic importance and links from education to wider inequalities – and to wider debates

Current approaches

• OECD – data rich longitudinal studies, disaggregated learning outcomes, and school vs. pupil characteristics

• Education Gini and composite indices (e.g. HDI) – years in school

• World Bank – ‘pre-determined circumstance’ and inequality of opportunity

• Group-based analysis

• Overlapping dimensions of marginalization- UNESCO Global Monitoring Report

• Learning disparities – regional assessments/EGRA

The wealth effect: People from the poorest households who are in education poverty

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Ph

ilip

pin

es

Turk

ey

Vie

tnam

Egyp

t

Ke

nya

Co

ngo

Ind

ia

Nig

eri

a

Yem

en

Ne

pal

Pak

ista

n

Mo

rocc

o

Sen

ega

l

Ch

ad

Bu

rkin

a Fa

so Sh

are

of

the

po

pu

lati

on

wit

h le

ss t

han

4 a

nd

less

th

an 2

ye

ars

of

ed

uca

tio

n

Extreme education poverty People with less than 2 years of education

Education poverty

People with less than 4 years of education

The gender effect: Girls from the poorest households who are in education poverty

In Yemen, the poorest 20% of households have an education poverty incidence

double the national average

And, for girls from the poorest 20% of households, the proportion triples.

The education poverty threshold (age 17-22)

• In Kenya, 96% of rural Somali girls (aged 17-22) have less than 2 years of education. • The current primary net attendance rate for Somali girls is only 30%.

20%

31%

17%

8%

25%

57%

73%

84%

96%

97%

Nigeria

Kenya

Ghana

Pakistan

India

Group average

Country average

Extreme education poverty % with less than 2 years of education

(age 17-22)

, poor, Hausa, girls

, rural, Somali, girls

, northern region, rural, girls

, rural, Sindhi, girls

, poor, Uttar Pradesh, girls

Overlapping disadvantages and education poverty

Richest 20%

Poorest 20%

Poor, rural Hausa girls

Rich, rural girls

Poor, urban boys

Poor, rural girls

Nigeria

Rural Hausa

Rich, urban boys

Urban

Rural

Urban

Rural

Rich, rural boys

C. A. R.

Chad

Bangladesh

Cameroon

Honduras

Indonesia

Bolivia

Cuba

Ukraine

0

2

4

6

8

10

12

14

Ave

rage

nu

mb

er

of

year

s o

f sc

ho

olin

g

Education poverty

Extreme education poverty

3.3 years

6.4 years

3.5 years

9.7 years

0.5 years

10.3 years

2.6 years

0.3 years

Boys

Girls

6.7 years

10 years

Education marginalization – inequalities within countries

The case of Nigeria

SACMEQ – wealth-based learning disparities

33.7%

27.2%

55.7%

10.5% 13.4%

38.9%

52.9%

35.8%

62.6%

0%

10%

20%

30%

40%

50%

60%

70%

South Africa Mozambique Zambia

Proportion of students scoring at the lowest level (percentage)

National Average Highest 25 % SES Lowest 25% SES

Source: SACMEQ 2007

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Total Achievement

Male Female Highest (greater Accra)

Median (Central)

Lowest (Northern)

Per

cen

tage

of

Stu

den

ts

Ghana: Levels of Achievement in Math (P6) Not Reached Min. Competency Reached Min. Competency Reached Proficiency

Source: Ghana National Education Assessment 2012

Region

Three core research areas

• Opportunities and entitlements – Accessibility

– Financing to mitigate disadvantage/incentivize education

– Links to wider strategies for combating marginalization

• Household and pupil characteristics – Poverty and nutrition

– Parental literacy

– Attitudes

• The learning environment – Teacher motivation, competencies and attitudes

– Infrastructure

– Curriculum, language, textbooks

Factors related to learning outcomes (P3 and P6) in Ghana

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Stu

de

nt

pe

rce

nta

ge

Reached Minimum Competency Reached Proficiency

Facilities (P6)*

Textbooks (P3)

School registers (P6-Maths)

Teacher gender and training (P6)

Multi-grade classrooms (P3)

*Water, electricity, and and girls' toilets

Nigeria: EGRA Results in Sokoto State

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Neither One or more

No Yes No Yes No Yes

Pu

pai

l Me

an O

RF

Sco

re

Mean Oral Reading Fluency (ORF) Scores by Home Characteristics of Pupils

Literate parents Read to at home

Quiet place to read at home

Eats before school

Kenya’s unequal distribution of out- of-school children (47 counties)

Mandera

Turkana

Wajir

Garissa

West Pokot

Narok

Kajiado Marsabit

Samburu

Isiolo

Tana River Lamu

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%

7.0%

8.0%

9.0%

10.0%

11.0%

0.0% 1.0% 2.0% 3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% 10.0% 11.0%

Shar

e o

ut

of

sch

oo

l pri

mar

y ag

ed

ch

ildre

n

Share of primary school age children

Source: EMIS/Census 2009

Derived share of FPE spending as a proportion of school age population: 47 counties (2009)

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

Turk

ana

Waj

ir

Gar

issa

Sam

bu

ru

Man

der

a

Mar

sab

it

Tan

a R

iver

Wes

t P

oko

t

Isio

lo

Bar

ingo

Nar

ok

Kili

fi

Kw

ale

Kaj

iad

o

Lam

u

Laik

ipia

Mo

mb

asa

Bu

sia

Kak

ameg

a

Mig

ori

Tran

s N

zoia

Nan

di

Ho

ma

Bay

Kis

um

u

Kit

ui

Mer

u

Uas

in G

ish

u

Nak

uru

Siay

a

Bu

ngo

ma

Nai

rob

i

Tait

a Ta

veta

Kis

ii

Vih

iga

Elge

yo M

arak

wet

Nya

mir

a

Ker

ich

o

Thar

aka

Nit

hi

Bo

met

Mak

uen

i

Kia

mb

u

Nya

nd

aru

a

Mac

hak

os

Emb

u

Kir

inya

ga

Nye

ri

Mu

ran

g'a

*Calculated on the basis of school enrollment data and census data on school age population

Research agenda discussion

• Need for more policy-oriented research on who is being left behind – and why.

• Identifying what works • Focused research on specific groups to inform

national strategies – eg UCW on child labor • The role of state and non-state providers • Beyond the school – linkages to wider sources of

marginalization • Strengthening disaggregated data and re-coupling

quantitative with qualitative analysis