Conversations with J-PAL: Teaching to the Level of the Child

Louise Albertyn

Over the months of July through October 2014, the South African hub of the Centre for Education Innovations, located at the Bertha Centre for Social Innovation and Entrepreneurship will host a series of conversations on education-related topics with the Abdul Latif Jameel Poverty Action Lab for Africa.

The first of these conversations was held on the 29 July and looked at the topic of teaching to the level of the child. In the education space, as the focus has progressed from increasing schooling attendance to improving the quality of education, a body of randomized evaluations is starting to help us understand some approaches that can be taken to improve learning outcomes. In this session J-PAL discussed what studies on remedial education, classroom inputs and computer assisted learning are suggesting in terms of what works to improve the quality of education, with a focus on teaching to the level of the child.

J-PAL consists of a global network of researchers who use randomized evaluations to answer critical policy questions in the fight against poverty. J-PAL is based at the Massachusetts Institute of Technology, with an independent regional hub in Africa at the University of Cape Town. J-PAL’s mission is to reduce poverty by ensuring that policy is informed by scientific evidence. 

Researchers in the J-PAL network use randomized evaluations to determine the impact of social programs. This methodology randomly selects participants to receive the chance to participate in a program – thereby eliminating bias arising from self-selection. These studies are designed with a sufficient sample size, such that the intervention and comparison groups are statistically identical at the start of the program. Thus, we can conclude that any difference between the group who received the offer of the program and the group that did not is caused by the program – and not from other biasing factors. Randomized evaluations are considered the most rigorous method for determining the causal impact of an intervention.

Educational access in developing countries has rapidly expanded. This expansion has put stress on school systems: not only are there more students enrolled in schools, but classrooms are now more heterogeneous in learning levels.  Many of the new learners are the first generation in their families to attend school and they may struggle to attain basic skills. Much evidence shows that levels of learning are low.  The South African education system is encountering some significant challenges as can be seen from the seminar delivered by Nic Spaull in July.  Similarly, in Kenya, over 20% of students in Class 3 cannot read a word in English or Kiswahili and 37% cannot do basic subtraction (Uwezo 2012). In India, the majority of 5th graders in rural India are three or more grades behind in reading (2013 ASER survey). The potential barriers to learning in these countries include: insufficient resources (teachers, books etc); the expectations of a rigid curriculum not matching the learning levels or needs of students; teachers lacking motivation or content knowledge and schools, students and teachers lacking appropriate technologies for 21st century education.

Spending more on inputs without changing the structure of the classroom has been shown to have little impact on student learning. An evaluation of a program providing textbooks in Kenya conducted from 1995-2000 showed no impact on average test scores, although the program increased scores for the top 20%.[1] Researchers began to suspect that it was the advanced level of instruction which was holding back poor performers, not necessarily a lack of inputs. Further studies on input provision including subjects flipcharts (Kenya) and libraries (India) show little or no effect on average test scores.[2],[3] Likewise, studies on reducing pupil-teacher ratios in Kenya, India and Ghana also do not show positive impacts.[4],[5]

In light of these findings, researchers in the J-PAL network decided to test interventions which tailor educational lessons to the level of the child. They reasoned that if students have not grasped basic skills, they would be unable to participate in more advanced lessons in the classroom. Interventions they tested include the following:

  • Providing low-cost remedial tutors
  • Using computer assisted learning adjusted to the level of the child
  • Tracking classrooms by ability level

Providing low-cost remedial tutors: Several studies have shown low-cost remedial tutors to be a promising way to help students with remedial skills and improve their learning. Multiple variations on this model have been tested, with clear results:  targeted tutoring works can help improve learning, especially for weak students.

In India, researchers Banerjee et al (2007) worked with the NGO Pratham to test several variations of targeted tutoring. For example, they tested the impact of low-cost tutors providing targeted instruction during school hours by pulling struggling students out of the classroom for two hours per day. [6] Test scores for the lowest performing children increased dramatically compared to the lowest performing students in the control group. These gains were equivalent to roughly half a year of additional schooling. Banerjee et al (2010) then tested another Pratham programme in India focused on holding after-school reading sessions.[7] This program recruited local volunteer tutors and trained them for 4 days in reading pedagogy. The evaluation found that literacy increased among 3-4 graders by 7.9% in comparison to the control group who did not receive the intervention. In Ghana, the Ghanaian Education Service and Innovations for Poverty Action tested the impact of a combination of during and after-school tutoring through the Teaching Community Assistant initiative. Despite a short intervention period, the evaluation found that out-of-school assistants were most effective at increasing scores. In-school assistants focused on remedial had a small, positive impact. In-school assistants had no impact if they supported the normal curriculum and training in-service teachers to focus on remedial education had a minimal impact.

Put together, this research shows that tutors can have a positive impact, especially on the weakest students. Tutors need not be highly skilled, as their goal is to focus on literacy and numeracy basics. Thus, they can be hired at low-cost. Sufficient training can take place in 2 weeks or less as tutors are not intended to mimic the role or skills of teachers, but rather deliver very basic assistance. Tutors are most effective when they focus on remedial skills. Tutors can work in a variety of settings including after-school sessions, intensive summer camps or pulling out students during school hours.

Computer assisted learning: Research conducted by Banerjee et al (2007) has found that the provision of computers has no impact on learning outcomes beyond computer knowledge. Self-paced instruction, however, can be impactful.[8] Students in India (2002-2004) spent 2 hours a week (2 to a computer) working independently with educational software and were supported by tutors. The evaluation found that the test scores of students receiving this intervention rose in Mathematics in comparison to a control group, but the program had no impact on language[EC1] .

Tracking by ability level: In Kenya, Duflo et al (2011) evaluated the impact of adding extra teachers to Standard 1 classes.[9] Some classes were split on past student performance (tracking) and others were divided randomly. Tracking improved test scores for both higher and lower-performing students. The average child moved from the 50th to 58th percentile.

Building on these results, a South African programme, Time To Read, is in a pilot phase of development. The programme would target learners who have already fallen behind their grade level and are transitioning from mother-tongue to English as the medium of instruction (Grade Fours). The aim is to improve English literacy and comprehension. Learners are tested with extended EGRA and split into three reading levels. The two strategies to improve literacy will be extended learning time in the holidays and a repair system in the form of a remedial curriculum and tutor.  The program has a further youth employment focus with tutors being drawn from the pool of unemployed matriculants.

The first conversation in this series provided compelling evidence for an education strategy that teaches to the level of the child.

Join us for the next Conversation with J-PAL on 27th August to explore the impact of school-based deworming on learning outcomes.

[1] Glewwe, Paul, Michael Kremer, and Sylvie Moulin. 2009. "Many Children Left Behind? Textbooks and Test Scores in Kenya." American Economic Journal: Applied Economics 1(1): 112-35.

[2] Glewwe, Paul, Michael Kremer, Sylvie Moulin and Eric Zitzewitz. 2000. "Retrospective vs. Prospective Analyses of School Inputs: The Case of Flip Charts in Kenya.” NBER Working Paper 8018.

[3] Borkum, Evan, Fang He and Leigh L. Linden. 2013. “The Effects of School Libraries on Language Skills: Evidence from a Randomized Controlled Trial in India.” NBER Working Paper 18183.

[4] Duflo, Esther, Pascaline Dupas, and Michael Kremer. 2011. "Peer Effects, Teacher Incentives, and the Impact of Tracking: Evidence from a Randomized Evaluation in Kenya." American Economic Review 101(5): 1739-74.

[5] Banerjee, Abhijit, Shawn Cole, Esther Duflo, and Leigh Lindon. 2007. “Remedying Education: Evidence From two Randomized Experiments in India.” The Quarterly Journal of Economics 122(3): 1235-1264.

[6] Ibid.

[7] Banerjee, Abhijit, Rukmini Banerji, Esther Duflo, Rachel Glennerster, and Stuti Khemani. 2010. "Pitfalls of Participatory Programs: Evidence from a Randomized Evaluation in Education in India." American Economic Journal: Economic Policy 2(1): 1-30.

[8] Banerjee, Abhijit, Shawn Cole, Esther Duflo, and Leigh Lindon. 2007. “Remedying Education: Evidence From two Randomized Experiments in India.” The Quarterly Journal of Economics 122(3): 1235-1264.

[9] Duflo, Esther, Pascaline Dupas, and Michael Kremer. 2011. "Peer Effects, Teacher Incentives, and the Impact of Tracking: Evidence from a Randomized Evaluation in Kenya." American Economic Review 101(5): 1739-74.


See more Teacher Training and Evaluation blogs


If a child can not learn in a way teacher teaches, he or she must teach in a way the child can learn. Teachers should focus to educate the children with mind of the above statement to ensure that children get opportunity to learn beyond traditional teaching and learning setting.

Add new comment

14 + 3 =
Solve this simple math problem and enter the result. E.g. for 1+3, enter 4.