The recently released Annual Status of Education Report (Aser) 2013, which covers every rural district in India, contains some good but mostly bad news. Enrolment in India’s primary schools is still at an impressive 96%, and the number of schools compliant with the Right to Education (RTE) Act norms, such as providing drinking water and usable toilets, continues to rise. But much of this is undermined by a fundamental challenge faced by public and private schools alike. Learning levels are unacceptably low and have been stagnant or getting worse over the past few years, especially in government schools. In basic reading and numeracy skills, the majority of our schoolchildren are falling several years behind where they should be based on their ages and the classes they attend. Although we have schools that are open and children that attend, many of these schools are failing in their basic mission of teaching.
In some ways, this isn’t new — Aser over the past four years has highlighted this very fact. But in our conversations with education departments in several states, we find many of them are only beginning to shift their focus to learning outcomes. While a few states are experimenting with new and innovative policies aimed at improving learning, there is very little consensus on interventions that are scalable and proven effective at improving learning in primary schools. Do we really know so little? Do we need to be at the mercy of our instincts, ideology or inertia, or can we do better and use rigorous evidence from the field to accelerate the process of learning for millions of children?
Researchers affiliated with MIT’s Abdul Latif Jameel Poverty Action Lab (J-PAL) have launched over a hundred rigorous randomised impact evaluations in the field to test what works, and what does not, in improving education. Many of these studies were conducted in India with various implementing partners, including state governments such as Haryana and NGOs such as Pratham. One of the clear lessons from this body of research is that teaching at the right level — classroom instruction based on the actual learning level of the child, rather than what the government-prescribed syllabus for that class may be — is consistently successful in improving student achievement.
Evidence for the teaching at the right level approach has been building up for over a decade. An early illustration of this approach was Pratham’s Balsakhi programme, implemented in Mumbai and Vadodara in municipal primary schools in 2001-03. Pratham mobilised local youth, often women, to work with children identified by the school as having fallen behind. These ‘balsakhis’ held two-hour pull-out classes in the school, using a teaching-learning package designed by Pratham, to teach core skills in literacy and numeracy starting from the current level of the children. An evaluation by J-PAL affiliated researchers found that the programme raised test scores in language and mathematics, with the largest gains going to lower-achieving students.
Subsequently, J-PAL affiliated researchers have rigorously evaluated several variations of this approach in different contexts. In Uttar Pradesh, an evaluation showed that this method can be implemented successfully by community volunteers. In Bihar, Pratham worked with government teachers to use the same methods and materials during school hours and in summer camps. The model worked during the camps but not during school hours, suggesting that government teachers can deliver learning gains with this approach when they are not under pressure to complete the defined curriculum and are instead given time to focus on specific learning goals. Variations on the teaching at the right level approach have also been adopted in Kenya and Ghana and have been found effective by randomised evaluations.
Perhaps most exciting of all, new results from a pilot programme implemented by the government of Haryana during the 2012-13 academic year show that significant learning gains can be achieved by applying the teaching at the right level approach through regular government teachers during school hours. Pratham trained cluster-level education officers in this teaching methodology, who in turn trained their teachers. For one hour each school day, teachers regrouped the students by their current reading levels, rather than by their age, and taught them with level-appropriate materials. The cluster-level education officers and Pratham staff also provided training, mentoring and monitoring support to teachers throughout the year. A randomised evaluation conducted by J-PAL found significantly higher scores in basic oral and written Hindi after one school year.
Haryana’s experience shows that state governments in India can improve learning outcomes by taking this proven model to scale, rather than continuing with some of the current, failing approaches. Governments can partner with organisations like Pratham, which can provide innovative teaching materials, training, and technical assistance, that contribute to the success of these programmes. In fact, some state governments such as Maharashtra, Madhya Pradesh and Bihar have begun taking some of these ideas to scale.
For the first time, the 12th Five Year Plan strongly emphasises ‘improving learning outcomes at all levels’ and talks of the need to design specialised programmes to achieve this goal and to evaluate such interventions. This is a great start, but too much is at stake here for governments and policy-makers to ignore what we already know works and start from scratch. Teaching at the right level works and has been validated through rigorous scientific research in multiple settings — a rare feat for any education intervention. Hopefully, when the next Aser is released a year from now, we will report that more education departments are addressing the problem by designing programmes based on hard evidence.
Iqbal Dhaliwal is deputy director of J-PAL at MIT, Shawn Powers is the education program manager for J-PAL, and Jasmine Shah is the head of policy outreach for J-PAL South Asia in Delhi
The views expressed by the authors are personal