1) Adult average schooling has
tripled between 1950-2010
2) Rapid expansion of
schooling - lots of schools
being built
3) Even in non-democratic,
repressive, corrupt, slow-growing
economies, schooling is increasing
4) in 2010, the average Haitian or
Bangladeshi had more year of
schlooing than the average
Frenchman or Italian in 1960.
PROBLEM!
Learning profiles too flat
e.g. Andhra Pradesh, India
Examined a variety of interventions:
Performance pay
Increasing school grants
Schools randomly selected for treatment
Developed test (ASER) to assess
student's rote learning and deeper
conceptual understanding
ASER test
Literacy
Numeracy
Telling time
Handling $$
Less than 2/3 of children
aged 15-16 mastered all 4
components
Test student in multiple grades
Produce cross section
learning profiles
RESULTS
FLAT LEARNING PROFILES!
e.g. single digit addition
Grade 2: 35%
Grade 5: 61%
PISA standardised test
Average Score 2009
OECD: 496
Denmark: 503
Tamil Nadu: 350
Also more students
placed in th bottom
category in Tamil Nadu
Sampling was mostly school based
Don't know the
learning outcome of
children not in school
Progress may actually be exaggerated
WHY???
Barriers to universal education
1) current page of
educational progress is too
slow and unlikely to change
because system is centralised
Lack of evidence
based planning
Only Education
Management Information
Systems (EMIS) visible goals
More inputs: books, teachers, buildings, etc.
Perpetuates illusion of progress
Tendency to replicate rich
countries' and educational system
Existing systems
cannot easily
change and flow of
innovation unlikely
Centralised "spider" system
- top down approach
Good at logistics and can
crank out primary
schools quickly
But has delivered:
More schooling
Resources for
marginalised children
that are not yet schooled
in remote areas
But education needs
judgement and control
Teaching a child
with top down
approach may be
forcing a round peg
into a square hole
Mostly run by government
Teachers
are mostly
civil servants
Nationalisation wave in
1870s -1920s beacuse
Increased demand for
skilled labour especially
during industrial revolution
rise of centralising
ideologies and nationalisms
Mimicry of successful
education systems
especially in France
and Japan
Why not follow them?
Most OECD achieved current levels
before 1970 - used to know how.
They now know how to improve but
not build from scratch.
Hasn't changed because
Camouflage: e.g. hiring building
contractors, putting colour in
textbooks to gain legitimacy as
if they are creating learning
System perpetuates
itself - it is more
convenient to be
blind/ignorant to
parents' concerns
can more spending improve learning outcomes?
NO!
Countries have
exactly same
learning outcomes
with different level
of spending
US PPP $105,000 per child
Poland PPP $39,000 per child
Countries have
different learning
outcomes with
same level of
spending
Finland PPP $71,000
But Finland outperforms
Spain by 50 points on
average PISA results
Spain PPP $74,000
International
assessments:
East Asian
countries -
highest scores
Most OECD
clustered around
OECD mean of 500
in PISA
Eastern European
countries near
OECD mean
Most developing
countries score
around 400 or below
Solution
"Starfish"
bottom up
approach
Qualities
Examples
Community controlled schools:
Community hired teachers paid
1/5 amount as civil service
teacher but produce equivalent
or better learning outcomes.
E.g. Andhra Pradesh
Private Providers of
education like in Kenya
(Bold et al. 2011)
Schools under small
governmental
jurisdiction
Putting cameras in classrooms and have teachers take
date-time stamped photos >> increase student attendance
and scores in India (Duflo, Hanna and Ryan 2010)
Open
have many different types of
schools providing education
instead of only having
schools under spider control
Locally operated
Given
autonomy
to operate
Performance pressure on learning
through accountability
professionally networked
communities of professional practice
Flexibly financed
finance follows students and
performance rather than mainly
being directed at the teachers,
independent of performance