Can internet access at home stimulate learning among children? It is an especially urgent question in developing countries where digital access and educational achievement lag considerably behind those in developed ones.
Ofer Malamud, Santiago Cueto, Diether Beuermann and I decided to address the issue in an experiment where we provided both laptops and internet access to low-income children in Lima. Our findings were not encouraging, but are far from the end of the story. A significant body of research points to the fact that guided learning, in which children use well-designed software at school to augment their course material, can be among the most effective ways to help children develop skills.
The problems with One Laptop Per Child
A key landmark in the debate over the ability of computers to speed intellectual growth among children is the 2008 One Laptop Per Child program in Peru. More than $200 million was spent distributing 900,000 computers. But as documented in an IDB study that we produced, the results were disappointing. If computer literacy rose, children’s abilities in reading and math showed no improvement at all.
We decided we would provide different conditions under which children engaged with computers outside of school. Some observers had argued that a problem with the One Laptop Per Child program in Peru was that it didn’t encompass internet access and was carried out in mostly rural areas of the country where neither teachers nor parents had the experience to help the children engage with digital technology. Our experiment, by contrast, unfolded between 2011-2013, in Lima. It offered both laptops and high-speed internet access. And it provided an additional benefit: children were given eight training sessions on how to access educational websites and search for information on sites like Wikipedia.
Unguided computer use doesn’t foster learning
The results, nonetheless, disappointed again. Conducted with a sample of about 1,800 low-income children in the third through five grades, our experiment confirmed a basic lesson from the One Laptop Per Child experience: Children don’t usually benefit from unguided computer use. They are spontaneous creatures, intrigued by play above all things. That is to say when unsupervised they tend to spend most of their computer time not working out academic problems, but watching YouTube and playing video games. They may improve their digital skills. But they make no real gains in math and reading, general cognitive skills, or grades. Even the novelty wears off. At the beginning of the study 40% of the laptops were used every day. After three months, this was reduced by half.
When the Internet makes a difference
Guided programs are another story. In joint work with Elena Arias, we looked at six studies conducted in India and China in which computers were provided to schools for after hours work in areas like math and language. These programs tended to share certain features. Computers were shared among students; software was especially tailored to supplement course learning, and teachers were there to help with any problems. The results were impressive. As we demonstrate in our SkillsBank, an online resource with evaluations of studies done around the world, guided learning on average improved educational attainment by 40% for third or fourth graders over what they normally would have achieved in the course of a year. That is almost three times greater than reducing class size and four times greater than expanding the length of the school day. It is also much cheaper. The only intervention more effective was tutoring, which yielded an 80% boost in learning, but is often prohibitively expensive.
Latin Americans are far more connected to the internet than they were at the turn of the millennium, but they still trail considerably behind people in North America, Western Europe and Developed East Asia in internet usage. That has tempted policymakers in the region to think wistfully that if only computers and broad-band access were more readily available, low-income children might begin to close their educational gap with their peers in the developed world. Our experiment and the One Laptop Per Child experience show that assumption to be false. Handing out equipment in-and-of-itself does little. Guided computer learning with well-designed software and committed teachers, however, can be transformative, stoking learning and taking children to a new level.