ICT in Education in Developing Countries: What the Evidence Actually Says

Education technology — tablets, laptops, e-learning platforms, digital content, computer labs — has been one of the largest investment categories in ICT4D since the early 2000s. Programs ranging from One Laptop Per Child (OLPC) to national tablet distribution initiatives to continental connectivity programs have put devices and connectivity into schools across Africa, Latin America, and Asia at a scale of hundreds of millions of dollars.

The evidence on whether these investments improve educational outcomes is, to put it gently, not encouraging. The disconnect between the investment rationale and the evaluation evidence is one of the more striking puzzles in ICT4D.

The Scale of Investment

Since the early 2000s, education technology programs in low- and middle-income countries have included:

One Laptop Per Child (OLPC): Launched in 2005 with the ambition of producing a $100 laptop for every child in the developing world. The XO laptop was deployed at scale in Peru, Ethiopia, Uruguay, Rwanda, and numerous other countries between 2007 and 2015. Total investment exceeded $800 million.

National tablet and device programs: Peru, Brazil, India, South Africa, and many other countries distributed devices — tablets, laptops, smartphones — to school students at national scale, with investment running into the hundreds of millions of dollars.

School connectivity programs: The World Bank’s $5 billion investment commitment in school connectivity (announced as part of the Giga initiative with UNICEF), African Union programs, and bilateral donor investments have funded broadband connectivity to schools across sub-Saharan Africa, South Asia, and Southeast Asia.

Digital content and platform programs: E-learning platforms, digital curriculum content, and learning management systems have been deployed in thousands of schools — often with donor support from organizations like the Gates Foundation, Hewlett Foundation, and bilateral development agencies.

The Evaluation Evidence

OLPC: The Benchmark Study

The most rigorous evaluation of a large-scale ICT-in-education program in a developing country is the Inter-American Development Bank’s evaluation of OLPC in Peru, published in 2012. This randomized evaluation — the closest thing to a gold standard in impact evaluation — examined the program’s effects on learning outcomes after 15 months of implementation.

The findings: negligible effects on reading and mathematics test scores. The program increased computer use significantly. It did not improve educational outcomes as measured by the standard tests.

Subsequent evaluations of OLPC in Ethiopia and other contexts found similar patterns. The devices were used, often in ways parents and students valued — for games, entertainment, and non-school uses — but the hoped-for learning gains did not materialize.

Tablet Programs

Evaluations of tablet distribution programs — in India, South Africa, Peru, and other countries — have also found limited evidence of learning outcome improvements. A meta-analysis of randomized controlled trials of education technology in low- and middle-income countries (published in Educational Research Review) found that on average, technology interventions had small positive effects on learning outcomes, but with very high heterogeneity — some programs showed meaningful gains, many showed nothing.

What Positive Results Look Like

Cases where technology produced measurable learning gains share several characteristics:

Software quality: Programs with adaptive software that personalized instruction to student level — rather than simply digitizing textbook content — showed better results
Teacher integration: Programs where teachers were trained to integrate devices into instruction (not just to supervise students using them) consistently outperformed programs where devices were given to students with minimal teacher involvement
Curriculum alignment: Programs where digital content was aligned with the national curriculum and examination system showed better learning transfer
Targeted content: Programs targeting specific skills (literacy, numeracy) with well-designed software outperformed comprehensive “digital learning” programs

Why Technology Fails to Improve Learning

ICT4D education researchers have proposed several explanations for the persistent gap between technology investment and learning outcomes:

The “tool” fallacy: Technology is a tool, not a pedagogy. Giving a student a device does not change how they learn unless the pedagogical approach changes. In most large-scale deployment programs, pedagogical transformation was not adequately invested in.

Teacher training and buy-in: In almost every context, teachers determine whether technology is meaningfully integrated into learning. Programs that distributed devices without substantial ongoing teacher professional development produced little pedagogical change. Teachers who lack confidence in using technology, who see it as threatening or as additional burden, or who were not consulted in program design rarely use it effectively.

Infrastructure and maintenance: Devices become unusable without maintenance, charging infrastructure, internet connectivity, and software updates. Many device programs failed to plan for these ongoing costs, resulting in devices that became non-functional within 12–18 months of distribution.

Content quality: Most deployed digital content was simply digitized textbook content — the same material in digital form rather than material designed to leverage the interactive capabilities of digital devices. Low-quality digital content produces the same results as the paper content it replaces.

Assessment alignment: Where examinations are paper-based (as in most low-income countries), student performance on paper examinations is a poor measure of skills developed through digital interaction. Technology programs that improve digital skills may not show effects on traditional examination scores even if real learning gains are occurring.

What Works: Conditional Evidence

Research has identified education technology approaches with stronger evidence:

Computer-assisted learning (CAL) with adaptive software: Studies in India (Pratham’s CAL program), sub-Saharan Africa, and Latin America have found positive effects on literacy and numeracy from well-designed adaptive software programs that adjust content to student level. The software quality matters more than the device.

Instructional radio and audio: One of the oldest education technology programs — instructional radio, developed in several African countries in the 1970s — produced some of the strongest education outcomes in the ICT4D literature. Audio-based learning has shown effects in multiple contexts and is more robust to infrastructure constraints than digital technology.

Teacher support tools: Technology used to support teachers — providing lesson planning resources, enabling remote supervision and coaching, giving access to professional communities — shows more consistent effects than technology given directly to students.

Targeted remedial programs: Technology programs specifically targeting students who are behind grade level in literacy or numeracy, with software designed for remedial learning, show stronger effects than comprehensive programs for all students.

Implications for ICT4D Practice

The education technology evidence suggests several lessons for ICT4D practitioners:

Hardware is the easy part. Distributing devices is logistically challenging, but it is far simpler than the pedagogical transformation required to produce learning gains. Programs should invest at least as much in teachers and pedagogy as in devices.
Pilot with evaluation before scale. Many large-scale education technology programs were scaled before rigorous evaluation of effectiveness. The OLPC evaluation’s results came after hundreds of millions had been spent.
Ask what problem technology is solving. In contexts where the primary constraint on learning is teacher quality, large class sizes, or frequent school absences due to household poverty, technology addresses none of these constraints directly.
Maintenance is not optional. Plans for device repair, charging infrastructure, software updates, and connectivity maintenance must be part of the original program design, not afterthoughts.

Frequently Asked Questions

Has any ICT-in-education program worked at scale in a developing country? Yes — Uruguay’s Plan Ceibal, which distributed one laptop per student nationwide starting in 2007, is one of the few programs that has shown some positive effects at scale. Factors cited include sustained political commitment, robust technical support, investment in teacher professional development, and integration with the national curriculum.

Why do governments continue to invest in education technology despite disappointing evidence? Several reasons: the evidence is not widely known among policymakers; the visible gesture of distributing devices is politically popular; technology companies actively market to governments; and there is genuine hope that the right implementation will produce the results the technology promises. The gap between research evidence and policy is a pervasive feature of the development landscape, not specific to education technology.

What is the Giga initiative? Giga is a UNICEF-ITU initiative to connect every school in the world to the internet. It maps school connectivity and works with governments and the private sector to close connectivity gaps. It is infrastructure-focused — it addresses the connectivity prerequisite, not the pedagogical question of whether connectivity produces learning gains.

Do mobile phones in schools improve learning? Most evidence suggests not significantly — and there is evidence that unrestricted phone use in class is associated with reduced learning. The distinction between student-owned unrestricted mobile phones and teacher-guided use of specific educational applications matters considerably.

How does ICT4D research inform education technology policy? The research-policy connection in education technology is notably weak. Academic findings rarely reach policymakers in usable form, procurement cycles are driven by technology company relationships rather than evaluation evidence, and political incentives reward visible technology programs regardless of their effectiveness. Strengthening research-policy translation is a priority for the ICT4D research community.

Development Informatics