Massachusetts Institute of Technology (Banerjee & Duflo); Columbia University (Linden)

October 2003

This 15-page report presents preliminary results obtained after the first year of a 2-year randomised evaluation of a computer assisted learning (CAL) programme implemented by the non-governmental organisation (NGO) Pratham in Vadodara, India. As background, the authors note that, when used as a supplement to regular instruction, CAL can be a promising strategy for improving the quality of education in developing countries. "Good educational software can be reproduced at nominal cost, and well-designed educational games can sustain interest and curiosity even in an otherwise dull school environment." This approach could be especially fruitful in India, where the high-tech sector is both successful and visible. "Unfortunately, despite the general excitement, there exists very little rigorous evidence of the impact of computers on educational outcomes and no reliable evidence for India or other developing countries. Furthermore, what evidence that exists is not particularly encouraging..."

Motivated by a desire to address this dearth of evidence, the authors begin by describing Pratham's CAL programme. Specifically, the government delivered 4 computers to each of 100 municipal primary schools in the city (80% of the public schools). A survey conducted by Pratham in June 2002 suggested that very few of these computers were actually used by the children, however: "most of the computers remained in their boxes, for want of anyone capable of operating them." In response, Pratham hired a team of instructors from the local community and provided them with 5 days of computer training. They then provided children in the fourth standard with 2 hours of shared computer time per week (2 children sharing one computer) to play a variety of educational computer games chosen because they emphasised some of the basic competencies in the mathematics curriculum. The instructors encouraged each child to play games that challenged the student's level of comprehension, and - only when necessary and/or prompted by the children - they helped individual children understand the tasks required of them by the game. (They provided no general instruction in mathematics.)

In brief, this evaluation found that, in its first year, the programme had a large and significant effect on test scores in math, an effect comparable for boys and girls and significant at all levels of the distribution (although somewhat larger for children at the bottom of the skill distribution range). While the first semester of the intervention showed very insignificant effects, the overall effect over the entire year showed substantial improvements over the first 4 months, with average scores on a 50-point math test rising from 14.9 to 29.0 in the "treatment group" but only from 15.5 to 25.0 in the control group. However, the programme seems to have had no spillover on language competencies; "this indicates that the more interactive, computer-based approach to learning might not have created a greater enthusiasm for learning overall."