The document discusses a study on private schools in India that examined learning outcomes and progress across different types of schools. Some key findings from the study include:
1) There is significant variation in test scores and learning progress between school types, with private unaided schools showing higher average scores and learning gains.
2) Students from wealthier families and with more educated parents are more likely to attend private schools, indicating that school choice is linked to family background.
3) When controlling for student background factors, private schools generally "add more value" and have higher learning effectiveness than other school types.
4) Within the private school sector, higher fee schools tend to be more consistently effective than lower fee schools,
2. A Thorny Debate?
1. Growing enrolmentin private schools across India
• Driven by both rising demand (incomes, preferences) and supply
2. Some evidence of better learning progressin private schools
• Potentially widening gaps
• Potential knock-on effects in government system e.g. weakening accountability
3. Some evidence of greater efficiencyin low-cost private schools
• Even if outcomes are the same, low fee private schools often have much lower
unit costs than government schools (especially teacher salaries)
4. But evidence that choice and competition benefit all studentsis much less
clear
• Overall (in govt and private) learning levels are low – little evidence of ‘a rising
tide lifting all ships’; potential widening inequality
Fortunately, we don’t attempt to answer all these questions
3. 1. What are the characteristics of children
attending different school types?
2. How do learning and learning progress compare
across different types of school? How does this
change when we include controls for student
background?
3. Within private schools, what is the relationship
between fees paid and learning gains?
4. Considering all of these things, what are the
implications for equity within the Indian
education system?
Our Questions
4. Young Lives School Survey - India
India Secondary School Survey (2016-17)
• Grade 9 students
• Four different types of school
• Progress in Maths and English in Grade 9
• Tests at the beginning & end of the school year
• School, teacher and student background data
5. Young Lives School Survey: Sample
Sample designto explore school
choice available in each of the 20
Young Lives sites
Sample stratified by 4 school types:
•State government schools
•Tribal/SocialWelfare schools
•Private Aidedschools
•Private Unaidedschools
Number of schools sampled in each
site proportional to the total
number of schools in that site:
Total number of schools in a
site Proportion sampled
> 80 schools 10% sampled
21-80 20% sampled
8-20 schools 50% sampled
<8 schools 100% sampled
(exception: less prevalent school types
are oversampled)
7. A lot of variation in test scores
There is considerable variation in mean maths test scores
at Wave 1 and Wave 2.400500600700
300 400 500 600 700
Mean school maths score (wave 1)
8. Including variation within school types
Variation within and between school management types,
especially for govt schools – some very high scoring and
others very low
400500600700
300 400 500 600 700
Mean school maths score (wave 1)
Private Aided Private Unaided
State Govt Tribal Social Welfare
9. Along with significant variation in progress
By the end of Grade 9, private unaided schools have a
higher mean score and make more progress on average
School type
Mean maths
score at start
of Grade 9
Mean maths
score at end
of Grade 9
Mean points
of progress
made
PrivateAided 470 483 13
Private
Unaided
538 577 39
State
Government
454 474 20
Tribal / Social
Welfare
412 435 23
Total 500 531 31
10. • Incoming test score differences the result
of prior school and home ‘inputs’
• Progress is the result of ‘continuing’
influences of home and school inputs over
the year
• We can control for some of the home
inputs over the year
• BUT certain factors like motivations are
unobserved and may explain both school
choice and progress
What Explains Differences?
11. There are big disparities in student intake
There are large disparities in the characteristics of students attending
different types of schools – wealth, gender, caste, parental education
Suggests school ‘choice’ is linked to wealth and family background (as
described by Srivastava, 2007)
0
20406080
100
Percent
Private Aided Private Unaided State Govt Tribal Social Welfare
School Type & Wealth Terciles
bottom tercile middle tercile top tercile
12. Private schools ‘add more value’
‘Value-added’ is a
statistical measure of
‘school effectiveness’.
It measures how much
schools or teachers ‘add’
to student learning over
and above the progress
they might be expected
to make.
Value-added can also be
contextualised by adding
in student background
factors.
-20-10
0
1020
MathsValue-Added
Private Aided Private Unaided State Govt Tribal Social Welfare
Mean School VA (unconditional)
Mean School VA (conditional)
Young Lives data suggests that private
schools add more value than any of the
other school types – even when we
control for student background
14. A lot of variation between private schools
Distribution of fees charged to Grade 9 students0
.00002.00004.00006.00008
Density
0 10000 20000 30000 40000 50000
Annual fees (rupees)
kernel = epanechnikov, bandwidth = 2.5e+03
Mean: Rs.13,476 (£150);
Median: Rs.12,000 (£133).
We classified
the private
schools in our
sample into 3
categories using
ratio of
fees/average
state per capita
income
(method used by
Kingdon, 2017)
15. Higher fees = more learning?
0
1020304050
MathsValue-Added
High fee Moderate fee Low fee
Mean School VA (unconditional)
Mean School VA (conditional)
Data suggests that school effectiveness varies by fee level –
even when we control for differences in student background
(more advantaged children in high fee schools)
16. Low fee offers better ‘value for money’?
Do low fee schools offer better value for money?
An additional 1000 rupees gets you more points of progress in a low fee
school than a high fee school on average – but a lot more variation. High
fee schools have a lower median but this is more guaranteed.
-20
0
204060
High fee Moderate fee Low fee
17. Low fee schools more of a ‘quality lottery’
Some low fee schools add just as much value as high fee schools
But high fee schools are more consistently above average – paying
more for lower risk? Opting out of the ‘quality lottery’?
-100
-50
0
50
100
0 20 40 60
Private schools by value-added rank
High fee Moderate fee Low fee
19. Discussion and implications
• In the Indian education system, there is too much inequality
and too much poor performance – not all children have the
same opportunity to meet minimum learning standards.
• Data suggests that more advantaged children are ‘sorted’
into ‘better’ schools, even when the background of children
is controlled for. This is true within the private school sector
as well as between govt and private.
• Huge variation in performance between schools, but limited
information available about school performance. How are
‘school choices’ made?
• Spending more appears to be a means of opting out of the
‘quality lottery’ – higher fee schools are consistently more
effective. Consequences for equity?