In 2012, we published a study showing that an ambitious and expensive program involving the disbursement of free laptops to primary school students generated few academic gains after 15 months of laptop exposure. The report, examining the One Laptop per Child initiative launched by Peru’s government, generated much discussion in Peru and around the world given that many governments had been aggressively implementing similar initiatives.
While our results were aligned with subsequently published papers that showed limited effects from laptop provision, we thought that wasn’t the end of the story. There was at least the possibility that such programs could pay off over the longer term. Teachers, principals and students might need more time to learn how to use the laptops. And students might benefit from positive changes in other areas, including both their digital and socio-emotional skills, allowing them to better advance through the educational system.
With that in mind, we produced a second study presenting experimental evidence on the long-term effects of distributing laptops on children’s educational development. Despite the investment of tens of millions of dollars and the distribution of hundreds of thousands of laptops, the results were again disappointing. Students subjected to the program, our research shows, did not improve their academic performance and had even slightly lower on-time primary and secondary school completion rates than those who didn’t participate. Moreover, while students improved in their ability to use the provided laptops (known as XO laptops), they only marginally improved their ability to use PC computers and showed no improvement in internet skills.
Providing computers with little guidance on their use, it turns out, is likely to do little to improve cognitive abilities. Teachers must be sufficiently trained in digital skills, have appropriate software, and be committed to using technology in a targeted way in the classroom to allow technology’s promise to flourish. Otherwise, that promise is fleeting.
A Unique Review of the Experience in Peru
Our research, examining academic performance in 531 public, primary and rural Peruvian schools from 2009-2016 and published in November 2024, documents this issue in detail. As the first study providing long-run experimental evidence of the effect of digital technology on educational programs in developing countries, it demonstrates what happens in areas such as reading and math to second grade students provided with laptops, but with limited additional support.
One question we aimed to address was whether schools “learned” to effectively use technology over those six years. By dividing our sample into treatment and control groups—those that received laptops and those that did not—we found no evidence of such learning, and there was no significant improvement in student achievement attributable to the laptop initiative.
We also looked closely at children that received laptops as they progressed on their educational journey in the ten years from primary school on, using data from national exams at the 4th and 8th grade levels, 5th and 6th grade tests that we administered, and longitudinal data on grade progression from primary to tertiary education. Again, we found little evidence of improvement in either mathematics or reading. There was a negative effect of 2.2 percentage points in the likelihood of finishing primary school on time and there were small but negative effects of completing secondary education or applying to university on time of 3.1 and 2.3 percentage points, respectively. We didn’t encounter improvements in internet-related skills, on an index of cognitive skills, verbal fluency or coding abilities.
Guided Learning in Chile
Does this bode ill for the promise of technological solutions to learning? On the contrary, advances in computational capacity, artificial intelligence and access to computer and digital devices could be educationally revolutionary. A comparison is instructive. A highly effective program in Chile, which provided two weekly 90-minute sessions in computer labs to disadvantaged Chilean fourth graders studying math, accelerated learning for participating students by 50%. That is substantially larger than the gains produced by more traditional interventions, such as extending the school day, reducing class sizes, or providing teacher training.
A key difference between that experience and the One Laptop per Child initiative in Peru was the surrounding support system. The teachers in Chile were assisted in using the new technology by external coordinators, given appropriate software aligned with the national curriculum, and directed to systematically use the computer labs to enhance their students’ mathematical abilities. By contrast, in Peru, our surveys of teachers show that the one-week of training they received in laptop use was insufficient to markedly boost their digital skills or to encourage them to employ computers in class. Indeed, among teachers in treated schools, the use of laptops in class was minimal: less than an hour more than for teachers in the control schools, where laptop use amounted overall to less than four hours per week.
Students that participated in the One Laptop per Child program in Peru don’t appear to have used their XO laptops at home primarily for educational purposes. As additional surveys show, they were more likely to use them for entertainment — a reason, we can only speculate — may help to explain the even worse on-time completion rates in the treatment schools when it comes to finishing primary and secondary school.
Laptops, Artificial Intelligence and Pedagogical Support
Simply handing out machines doesn’t work. As an another IDB experiment in Lima between 2011-2013 demonstrates, not even offering students laptops, high-speed internet access, and a few hours of training in accessing educational websites makes a real difference to their academic growth. What matters is pedagogical support and guidance. Many studies in developing countries show that technology can have positive effects on academic achievement. These studies have evaluated after-school programs with additional instructional time and concerted efforts at guided learning. More research is needed on the integration of technology into standard school curricula. The key thing to remember, however, is that while artificial intelligence, greater computational capacity, and expanded internet access may well mark a very promising new era in education, kids will always need help in using new technological tools to expand their skills.
[Editorial Note: Several of the studies mentioned in this article were undertaken in collaboration with Santiago Cueto, of the GRADE research center.]
Leave a Reply