Rani Kumari, 11, from Bhopal in Madhya Pradesh has been going to school for the past six years. A Standard 6 student, she is in the appropriate grade for her age and attends school regularly. By all accounts Rani Kumari should be able to read poetry. Yet, ask her to read a few sentences and she is unable to do so. What then have all these years of schooling done for her?

In order to assess learning levels and access to quality education of primary school-aged children (children between the ages 6 -14) in rural India, a district-wise national survey was conducted recently. This survey, named Annual Status of Education Report (ASER) and released in January 2006, was conducted in 509 of India's 575 districts. Of the districts surveyed, the data from 487 districts is usable. Conducted across 9,521 selected villages, ASER tested the children of about 203,000 households and surveyed approximately 9,500 schools. The villages and children were randomly selected; the selection method used was developed in consultation with the National Sample Survey of India.

A citizens' initiative, this survey was coordinated by the national NGO Pratham, but was conducted on a voluntary basis by more than 770 NGOs, citizen and student groups, and universities. ASER is the most extensive survey of its kind.

The data generated by the ASER survey paints a telling picture. Enrolment levels are high across the country, yet learning levels are surprisingly poor. Although the gender gap in enrolment has decreased, fewer girls than boys are enrolled in schools. Moreover, a larger number of boys than girls are enrolled at private schools. Learning levels at private schools are significantly higher than those at government schools. The data shows that while the government has moved towards its goal of increasing enrolment, it needs to now focus on learning and assessment.

Across rural India, 93.4 per cent of primary school-aged children are enrolled in school. Of these, 75.1 per cent are enrolled in government schools and 16.4 per cent in private schools. While 22.2 per cent more boys than girls are enrolled in government schools, 50 per cent more boys than girls attend private schools.

These numbers are significantly better than those given in the well-known Public Report on Basic Education (PROBE) Report 1999. According to the PROBE Report, only 70 per cent of children were enrolled in school. Perhaps, this difference can be accounted for by the fact that government programs to increase enrolment, such as the midday-meal scheme, have had their intended effect. It may also be because the number of private schools across the country has risen, and an increased number of children now attend private schools.

Ostensibly, it seems like we should be excited that only 6.6 per cent children are out of school. Yet, this percentage represents 11 million children. If each child is recognized as an individual person deserving of education, an inexcusable number of children are not in school. Of these 11 million children, 52 per cent in the 6-10, and 55 per cent in the 10-14, age groups are girls. Moreover, there is a wide variation between states with respect to enrolment levels. Goa, Kerala, Karnataka and Uttaranchal have hardly any children out of school. Meanwhile, Bihar, Rajasthan, Orissa, Andhra Pradesh and Uttar Pradesh account for 71.2 per cent of out-of-school children. The government's drive to increase enrolment is perhaps not as successful as seems at first glance.

Besides, enrolment is only one of the indicators of education; actual learning levels being the other.

Of all enrolled children, 51.9 per cent cannot read an easy short story, and 65.5 per cent cannot perform a simple division problem. (Cannot read: 30 per cent of children in private schools, 40 per cent of children in government schools; cannot solve written numerical problems: 45 per cent of children enrolled in private schools, 67 per cent of those in government schools.) Uttar Pradesh, Madhya Pradesh, and surprisingly Tamil Nadu and Karnataka, feature as the states with the worst levels of both reading and math. Gujarat and Orissa do badly with respect to reading and math respectively. Bihar, Haryana and Chhattisgarh - all states that are commonly assumed to have low development indicators - have unexpectedly high levels in both.

A closer glance at state-wise data makes it clear that there is no one factor that guarantees learning. In Bihar, for example, enrolment is low at 86 per cent, yet learning levels of Standard 5 children are high, with close to 90 per cent being able to read and close to 85 per cent being able to do math. Tamil Nadu and Karnataka show high enrolment at 97 per cent and 98 per cent respectively, but both have abysmal learning levels. Of Standard 5 children in these states, only about 75 per cent and 65 per cent can read and do math.

Enrolment clearly does not correspond to learning, and thus to education. Here is another example: although the pupil-teacher ratio in West Bengal, at 48:1, is much higher than elsewhere in the country (the national average is 38:1), learning levels are high at 92 per cent for reading and 90 per cent for math. It is clear then that learning does not directly follow from any one cause.

Why, though, are learning levels across the country so low? This question will elicit different responses in different places, and the ASER survey does not engage with it. However, one possible reason might be that the government's emphasis on child enrolment and retention has meant that that not much attention is given to learning assessment. For example, Sarva Shiksha Abhiyan, the recent government initiative to promote elementary education, only assesses the enrolment of children and the infrastructure of schools. Another possible reason for low learning levels could be that, in some states, teachers are penalized if children in their classes are seen to be not learning well. Teachers here will often promote children to the next standard irrespective of their actual learning levels.

Education, then, cannot be measured simply by enrolment. In promoting universal education, the focus of the government must include learning assessment. Of course, this does not mean that the government should stop work on increasing enrolment - those 11 million children out of school cannot be ignored - yet, it is not enough only to guarantee that all children go to school. If we are to prevent more stories like that of Rani Kumari, the government must ensure that all children in school are actually learning, and learning well.