Are private schools really performing better than goverment schools?

The debate on whether private schools provide better quality primary education as compared to government schools is
heating up in India. This is completely understandable in the current scenario. On the one hand, for almost ten years, through
Sarva Shiksha Abhiyan, the government has intensified the move towards universalizing elementary education and more
recently the Right to Education Bill has been passed in the Parliament. This push has led to impressive increases in provision
and enrolment. On the other hand, ASER as well as other data show a clear rising trend in private school enrolment in rural
India.
At the all India level, private school enrollment increased from 16.3% in 2005 to about 22.6% in 2008 – an increase of about
40%. In the last year, between 2008 and 2009, private school enrollment has marginally dropped to 21.8% in rural India.
There is considerable variation across states. On the one end of the spectrum are states like UP, Haryana, Punjab and Kerala
where private school enrollment is as high as 40% and on the other end of the spectrum are states like Bihar and West Bengal
with enrollment in private school closer to 5%. However, what is clear is that whether enrollment in private schools is high or
low, it has been increasing over time.1
What has led to this shift towards private schools in rural areas? The standard answer and the common perception is that
private schools provide a better quality education. This trend was started by parents living in urban areas – the elite having
opted out of the government school system and the middle and lower income classes trying their level best to send their
children to private schools – and now their rural counterparts are coming to the same conclusion. After all teacher attendance
is much better in private schools and these schools often give instruction in English, mastery of which leads to better job
prospects in the future.2
Indeed, the ASER results indicate that this might be the case. In the ASER 2008 report, I wrote a preliminary piece on the
differences between learning outcomes in government and private schools and how these differences narrow when household
and other characteristics are controlled for.3 ASER 2009 has additional controls available, mainly tuition and father’s education,
and this note takes advantage of that. In addition, ASER 2009 tests children for English as well.4 This gives us another
learning outcome to check for differences between government and private schools. More importantly, it gives us a learning
outcome – ability to read and comprehend basic English – which is often cited as the reason for sending children to private
schools.
In 2009, in classes 1-5, the percentage of children who could read at least a class 1 level text was 43.6 in government schools.
The corresponding figure in private schools was 52.2 – a whopping 8.6 percentage point advantage. However, this is an
uncontrolled difference in learning outcomes – one that is obtained in a simple cross-tabulation of learning outcomes against
type of school. It does not take into account that many different things affect a child’s learning level. For instance, it is well
established that the mother’s education has a positive impact on the probability that a child goes to school as well as her
learning ability. Supplemental help offered at home, in the form of paid tuition or by family members will also improve learning
outcomes of children, regardless of whether they go to government or private school. If any of these factors is positively
correlated with the probability of going to a private school, their impact will show up as enhanced learning outcomes in private
schools.
For instance, richer households can afford to pay for additional tuition for their children. It is also well established that a larger
proportion of children from more affluent homes attend private schools.5 In this case, part of the observed learning differential
between government and private schools would be due to the extra help that private school children were getting at home
and not because of the better quality of education being imparted in private schools. Therefore, it becomes important to try
and estimate the learning differential once other things that impact learning are taken into account.
Apart from type of school, ASER 2009 has information on many other factors that can impact learning. A simple model is built
to try to disentangle the effect of other factors from that of private schools on learning outcomes of children. Two learning
outcomes for children in primary school (class 1 – 5) are considered:6 7
• Ability to read a class 1 level text or more in their local language
• Ability to read simple words or more in English
The model controls for child characteristics like age, gender, number of siblings, education of both parents; household
characteristics like type of house (“katcha”, “semi-pucca”, and “pucca”), whether the house had a television, phone,
electricity, some kind of vehicle; and characteristics of the village the child lives in like whether the village had a bank, post
office, government primary, middle or secondary school, private school, STD booth, etc.
All the variables are significant in the model and have the expected impact. Learning increases with age, but then levels off.
(This is to be expected as the learning measure is a very basic and “floor” level indicator for reading.) A larger number of
siblings, presumably, reduces time spent on learning and reduces learning outcomes. Education of both parents is positively
correlated with their children’s learning level. Further, the impact of parents’ education rises monotonically with their education
level. Tuition has a large impact on learning – almost as large as the impact of mother’s education. Finally, all household
characteristics signifying greater affluence are positively correlated with learning outcomes.
Once we control for characteristics other than the type of school the child goes to, the learning differential between government
and private schools falls drastically from 8.6 percentage points to 2.9 percentage points – from 20% to a measly 5%. This
means that 2/3rd of the learning differential between government and private schools can be attributed to factors other than
the type of school. So at least in the case of reading in the local language private schools perform no better (or worse) than
government schools.