18 October 2005

Primary school completion in India, 1950-2000

The share of the Indian population who attended school has increased steadily over the past 50 years. Data on educational attainment in India shows that about 80% of all persons born around 1990 have attended at least primary school. Of the generation born between 1950 and 1970, only about 55% ever attended school.

Not every child who enters primary school completes that level of education. High dropout rates are a particular concern in Sub-Saharan Africa, the region with the lowest school life expectancy values worldwide. In India, on the other hand, most children who enter the first grade stay in primary school until they graduate after five years of education. The graph below compares the percentage of Indians who attended primary school (independent of the number of years) with the percentage who completed primary education between 1950 and 2000. The data, from a 2000 Multiple Indicator Cluster Survey (MICS), is disaggegrated by sex and area of residence.

Primary school attendance and completion by year of birth, India 1950-2000
Chart with share of population who attended and completed primary education by year of birth, India 1950-2000
Data source: India 2000 MICS.

The population is grouped by year of birth and for each cohort three values are shown:
  • the percentage who attended primary school or higher (blue area)
  • the percentage who completed primary school (green area)
  • the percentage of primary entrants who completed primary school (ratio completed primary/attended primary, red line)
In urban areas, virtually all residents who attend primary school also complete that level of education. In rural India, completion rates are slightly lower but even here children typically stay in school until they graduate. 85% to 90% of primary school students born between 1950 and 1970 went on to graduate. Since 1970, this share has increased to more than 93% for India as a whole and in urban areas this value is as high as 97%. (The gap between primary school attendance and completion rates since 1990 is explained by the fact that children born since then were not old enough to have completed primary school at the time of the survey.)

The education system in India succeeds at keeping children in school. Children currently out of school are therefore likely to complete at least the primary level of education once they take the crucial step of enrolling in the first grade. A disaggregation of primary school attendance rates in India makes clear that many of the children out of school are girls from poor rural households. Educators and policy makers have to focus their efforts on this group of children to reach the Millennium Development Goal of universal primary education.

Related articles
Friedrich Huebler, 18 October 2005 (edited 12 October 2008), Creative Commons License
Permanent URL: http://huebler.blogspot.com/2005/10/primary-school-completion-in-india.html


ristoharma said...

Based on existing studies, the concensus is that the Indian educaiton system does not succeed in keeping children in school to the end of the first five years primary cycle, let alone the full eight. The poor quality of education in government schools which most of the population attend is the key reason, cost being another key reason but not the key reason as is often assumed as shown by a study (K. Muralidharan and M. Kremer (2006) Private and Public Schools in Rural India, Harvard University, mimeo.) cited by Geeta Kingdon (Oxford Review of Economic Policy, Vol 23, No 2, 2007, p. 187, The Progress of School Education in India), that shows the existence of a private school, contrloing for other factors, is more likely in villages with a disfunctional/low quality government school. This this statement, "The education system in India succeeds at keeping children in school", is probably very misleading.

- Risto Harma, education researcher and India country programme specialist, formerly with Free Schools India, Jyotiba Phule Nagar district, Uttar Pradessh, India

Friedrich Huebler said...

Thank you for your comments. The MICS data tells us that 9 out of 10 children who enter primary school end up graduating. These numbers can be confirmed with data from the 2005-06 DHS (or NFHS, as it is called in India). According to the DHS, the survival rate to grade 5 is 95 percent. On the other hand, the most recent data from the UNESCO Institute for Statistics (UIS) shows a much lower survival rate to grade 5 of only 73 percent in 2004. The discrepancy between the survey data from the MICS and DHS and the administrative data from UIS calls for further research.

Risto F. Harma said...

With reference to the paragraph above where you write about attendance data and gender being a significant issue with regard to it determining whether a child is in or out of school, the other key issues determining who is out of school and/or in poor quality government schools are caste and religion. A very good short account of these issues can be found in Degrees Without Freedom by Craig Jeffrey, Patricia Jeffery and Roger Jeffery, 2008 (yes Craig's last name is spelled the way it is). They focus on district in Uttar Pradesh bordering Jyotiba Phule Nagar district to the north. The basic story is that in that the lower caste and non-Hindus are being kept out of good schools by higher Hindu castes. A comprehensive recent study of this is to be found in Jyotsna Jha and Dhir Jhingram (2005), Elementary Education for the Poorest and Other Deprived Groups: the Real Challenge of Universalisation, (New Delhi: Manohar). Surprisingly, this is perhaps the only recent comprehensive studies on these issues which are pivotal to understanding the background behind the statistics. There are a some rules of thumb that apply to Indian and developing country education statistics in general: in the case the Indian, given the well known issues of gender, caste and religion in developing a very effective system of discrimination, the more and more the data move toward 80 or 90 percent attendence, completion etc the harder it is to get the remaining children into school precisely because the last who remain to be enroled are from severely deprived communities or groups; te final element is then very poor quality teaching and teacher behaviour). Second, in the case of India, nearly all of the focus in education data discussions has been on enrolement and attendance, that is, given the suspiciiously high enrolment data and the well documented reasons for this (widespread fraud for personal financial gain and other distorted incentives (numerous studies site these), researchers wanted to know how can actual attendence to accurately measured; the issues of completion rates plays a distant second to this issue since there was high doubt about the enrolment data being an accurate measure of attendance (or actual in school child numbers); therefore high levels of dropout before class 5 were not in dispute, ie these levels were clearly much more than the 5 percent dropout suggested by the MICS data.

The basic point being made here is that extreme caution is required for use of Indian and developing country education data, especially where figures in the 90 percent range are being quoted.

One study, while an older one, makes the point only too clearly, citing enrolment data being 20 to 40 percent inflated (too high) - unfortunately I do not have the full reference available, it was Kurien 1983.

I'm currently working on a paper, the outline of which was recently presented at a conference at the European University, Cyprus, on this issue of education statistics with India and Nigeria as case studies. The case of Nigeria is very illuminating: data from the 2000-01 ILO-FOS (International Labour Organisation and Federal Office of Satistics, Nigeria)shows enrolment above 90 percent, however it also shows that 25 percent of children 5-14 years of age miss 4 days of school per week, 3 percent miss 3 days and 10 percent miss 2 days. So it is clear that we can not rely on there being over 90 percent of enrolment meaning 90 percent attendance. A further comparison is to notice that the Indian state of Kerala is widely agreed to have enrolment and attendance above 90 percent, and this is match with high levels (in the desireble direction) for a range of other human development indicators (HDIs). Nigeria on the other hand does not have such a good record for HDIs, hence lending mre weight to the conclusion that over 90 percent enrolment is a false statistic as a measure of actual education system success (lack of). This is the type of analysis that has to accompany reporting of all such references to developing country education statistics, hence more view as expressed in my first comment on your web page.
Further, I have been informed of some issues with regard to the MICs methdology, but I will need to investigate these further before commenting.

- Risto F. Harma

Risto F. Harma said...

Sorry for the incomplete reference
"2000-01 ILO-FOS (International Labour Organisation and Federal Office of Satistics, Nigeria", what I forgot to say was that this is the ILO-FOS 2000-01 Child Labour Survey, details of which I think are available on the ILO website.

- Risto F. Harma

Friedrich Huebler said...

I appreciate your thoughtful comments, thank you for taking the time to share them. On the issue of caste and religion, you may be interested in my article on "Caste, ethnicity, and school attendance in Nepal". An article on "Primary school attendance in Nigeria" has some analysis of disparities by gender, area of residence, and household wealth. I would be interested to hear more about your views of the MICS methodology. Also, please feel free to post a link to your paper on India and Nigeria on this blog.

Risto Harma said...

A question regarding the completion rate for primary education 6-10 years (classes 1 to 5): how are the DHS and the MICS collecting this data?

Also, with regard to the approximate 95% completion rate for India (primary school 6-10 years, classes 1 to 5) for the year 2000 cited above, the Census of India 2001 gives an attendance rate of 68.8%. It's these kinds of discrepancies that need to be addressed. I was reminded of a paper recently (currently attempting to locate it) that emphasised the fact that India Census data and the National Sample Survey (NSS) data show a consistent difference over time, with the NSS data on school attendance being higher than the Census. I believe the reason was thought to be the fact that the NSS is a smaller sample whereas the Census is the the population, not a sample. Someone needs to explain how it is possible to have 95% completion and 68.8% attendance. (For readers unfamiliar with the NSS it is a government department of the Indian government, the NSS Organisation, that carries out surveys fairly regularly, for example the one I refer to hear is their employment survey which includes current education attendance data for children and the survey is out every five years, most recent being 2004-05, referred to as the 61st Round. from which poverty data is also calculated, the Census of India.) the MICS 2000 survey you cite on the page on participation in Indian education puts net attendance at 72%, which isn't that far off the Census 2001 figure, but it stills shows the potential upward bias. The National Family Health Survey (NFHS-II) for 1998-99 has considerably more upward bias as I recall, there was a suggestion by one researcher that the survey question was badly worded that it was not clear enough to parents whether it meant children currently actually in school, attending regularly, or simply enrolled, which does not equal attendance as we know. (I presently would need to relocate the reference for that.)

Another interesting point is that for Uttar Pradesh (UP) for example, there is a survey, simply referred to as the UNICEF survey for 1999-2000, (probably the MICs 2000 you cite here on your website but this source below does not specify)referred to as being carried out for eight Indian states, cited in the book based on the survey called The Economics of Elementary Education in India (2006, (ed) Santosh Mehrotra, Uttar Pradesh chapter by Ravi Srivastava, and for UP, Srivastava cites Indian Department of Education data for drop out from classes 1 to 8 as 52.5%. Now the Census 2001 population data cohort, for some reason given as years 6 to 11, as opposed to 10, gives the total number of children (whether in school or out) at 24,823,641 (million); now obviously attendance data is typically higher the for the lower grades, but the above figure of 52% is suggestive of possibly high drop out between classes 1 t 5. This is interesting too because government data on childrenin school are usually severely biased upwards. And then you have Bihar state, Jharkhand, Orissa, Madhya Pradesh, Chhattisgarh and Rajasthan, and I'm just not having a great deal of confidence in 95% completion for classes 1 to 5. When I asked one of my education contacts on the ground in UP about 95% completion, for UP he said that would be "complete rubbish".

This can be looked at from another perspective: from the 1991 Census, there were 56,464,208 out of school of children. So we need to understand how we can get from this figure to only about 13 million in around 2005 (this is from a survey cited on the Sarva Shiksha Abhiyan (SSA)website, which is the government of India's flagship universalisation programme for education for classes 1-8). The out of school number for Census 2001, classes 1 to 5, is 42,221,422 (million) out of a total population cohort of 135,212,659 (million). So, starkly, we have to get from 42 million out of school in 2001 to the currently cited situation of around 13 million out of in 2005.

2001 Census data is from: data set name: "Attending Educational Institution", http://www.censusindia.gov.in/Census_Data_2001/Census_Data_Online/Social_And_Cultural/Attending_educational_institutions.aspx

-- Risto Harma

Risto Harma said...

To clarify the above data in the last comment for 21 October:

56 million (1991 Census, India) is for ages 6-10 years (classes 1 to 5).

13 million (2005 survey cited on the Sarva Shiksha Abhiyan (SSA) website, see also Education for All (EFA) Global Monitoring Report 2007, chapter two, box 2.1) is for ages 6-13; the figure for 6-10 years of age (classes 1 to 5) is 7.8 million. (For the EFA 2007 report and exact page reference, readers can go to http://www.unesco.org/education/GMR/2007/chapter2.pdf ).

Friedrich Huebler said...

Risto, thank you for your detailed comments. What you refer to as the "completion rate" is in fact the survival rate to grade 5. You can find information about this indicator in the article Survival rate to the last grade of primary school. The survival rate to grade 5 is the share of children entering grade 1 that eventually reach grade 5. The survival rate to the last grade of primary school is the share of children entering grade 1 that eventually reach the last grade. The survival rate can have very high values in spite of low attendance rates. Assume for example that only 20% of all children in a country attend primary school. If all of these children reach the last grade, the survival rate is 100%. The attendance rate of 68.8% from the 2001 census in India is therefore not incompatible with the survival rate of 95% calculated from DHS data. I hope this explanation was useful.

Risto Harma, Research Manager, The Loomba Trust said...

Let me clarify where I’m coming from:

In the above I do of course of realise the issues with regard to have, say, 90% of 20% of children you enter school, completing, or surviving, which ever we want to concentrate on.

However, for India, in say 40 years ago, you will have an elite mainly going to primary school and all levels of education, so you would expect something like 99% completion survival and attendance of that 20% who go to/enter school.

But as we move forward in time from 40 years ago, and education is opened up to more and more of the population, you would expect a deterioration of these high 90’s results above, for the simple reason, that more people from less affluent backgrounds are entering the education system with less chance of success because they do not have the range of educational support that wealthier families have. And this is to say nothing about the gender split (discrimination against girls due to discrimination against women) that increases as you move down the social hierarchy away from the elite.

The only locations in the country (India) where we can talk with confidence about 90% this, and 90% that, is Kerala (in the far southwest) and Mizoram (in the far northeast), and these are small states. Notice also that they have a small gender gap in male and female literacy, (I’m citing 2001 Census of India data , from p. 105, India Development Report 2002, Oxford University Press, edited by Kirit S. Parikh and R. Radhakrishna).

So whatever variable we are talking about in an Indian (or most of South Asia) or Sub-Saharan African context, I find it highly suspect when they record 90% and above.

- Risto Harma
Research Manager,
The Loomba Trust