About
This study in Pakistan, led by Harvard University in partnership with the Centre for Economic Research in Pakistan (CERP), examines how tech-enabled guidance for parents and teachers influences children’s learning, using randomised controlled methods to measure impact.
The randomised controlled trial (RCT), conducted in partnership with the Government of Khyber Pakhtunkhwa, measured learning outcomes in Maths, English, and Urdu using assessments originally developed for the Learning and Educational Achievement in Pakistan Schools (LEAPS) longitudinal study to measure foundational skills among primary school students from kindergarten through Grade 5.
Key Findings
On average, students in the TIP program treatment groups outperformed their peers in the control group in maths by 0.14 standard deviations.
We find that workbooks alone, with no technology component, are effective in helping increase personalisation and, therefore, learning—especially for students at the bottom of the learning distribution. The information and guidance treatment arms—both of which leverage technology—are marginally more effective than workbooks but provide more even improvements for students regardless of baseline learning level.
Data analysis suggests that most learners benefit equally from the interventions. While the intervention was effective on average across the baseline test score distribution, paper-based materials particularly benefited those who were most behind.
Bridging the Gap
The Challenge
Many schools and teachers in marginalised communities face structural barriers to engaging fully with digital personalised learning (DPL). Evidence on effective approaches to scale low-cost, tech-enabled learning support in both home and school settings is limited.
Why It Matters
Understanding how parents, teachers, and schools can be supported to use DPL tools is essential for improving learning outcomes. The study seeks to identify which delivery models and enabling conditions best support effective, scalable implementation in Pakistan.
How this Work Aims to Address It
The study tests three approaches through a large-scale randomised control trial covering over 2,700 schools, both public and private. These approaches focus on home-led learning, school-led learning, and customised informational inputs on student progress. A design-based implementation research approach is used to iteratively refine the application and assess how different actors contribute to learning gains.
Objective
This randomised controlled trial (RCT), working closely with the government of Khyber Pakhtunkhwa, Pakistan, examines the key pathways to home- and school-based facilitation of low-cost DPL at scale. This study addresses gaps in evidence around the implementation of personalised learning, facilitated by actors at both home and school, and how the support of those actors contributes to learning gains, and demonstrates potential for scalability.
The Research Questions
- Is there an information failure that inhibits learning progress? Do actors know enough about child learning levels and possible learning activities they could utilise to enhance their child’s learning? Does alleviating this failure boost learning, and by how much and at what cost?
- Are actors able to deploy information on the child’s learning gaps and learning activities, to better guide the child’s learning? If not, is there a “guidance” failure? Does alleviating this failure boost learning, and by how much and at what cost?
- Are school-based and home-based actors differentially affected by the above failures? Is the impact of addressing them, and costs of adopting them, differential across these types of actors?
- Are school-based and home-based actors complements, substitutes or have separable and independent effects in helping children learn?
Study Design and Methodology
This is a randomised controlled trial (RCT) with significant Design-Based Implementation Research (DBIR) components, allowing the team to iterate on and refine the application design throughout the study. The experiment has three arms:
- Home-led DPL
- School-led DPL
- Customised informational inputs on learning progress (test scores and relevant content)
The sample covers 2,700 schools, split evenly between public and private, all within the Khyber Pakhtunkhwa province.
The Importance of the Results
The study demonstrates the importance of supporting overburdened teachers, lifting constraints on parental involvement, and encouraging engagement.
Teachers in Pakistan struggle with an ambitious curriculum and large, mixed-ability classrooms, with some students three to four grades behind. The TIP program effectively reduced teachers’ administrative and mental workloads (e.g., homework), giving them more time to teach. The research also shows that teachers respond differently to structure versus autonomy—a crucial factor for program design.
The TIP program successfully helped ease parental burdens at home, allowing parents to focus their time on other children or tasks. The program was also successful in enabling those with low literacy levels to engage with the program materials.
Implications for Policy and Practice
This study has generated significant interest in taking up the TIP program in various ways among both government and private-sector partners. This includes a Phase 2 of the study with the Ministry of Federal Education and Professional Training (MoFEPT) in Islamabad, as well as interest from the Khyber Pakhtunkhwa Elementary and Secondary Education Department (KP E&SED), the Punjab Monitoring and Implementation Unit (PMIU), The Citizens Foundation (TCF), Taleemabad in Balochistan Province, and the Pakistan Foundational Learning Hub (PFLHub). Please see the final report for details.
Study Team
- Christina Brown, Co-Investigator (University of Chicago)
- Tahir Andrabi, Co-Principal Investigator (Pomona College)
- Asim Khwaja, Principal Investigator (Harvard University)
- Isabel Macdonald, Co-Investigator (Lab for Inclusive FinTech, Haas School of Business, UC Berkeley)
- Julia Ladics Collins, Senior Research Manager
Previous contributors to this study:
- Akanksha Bapna
- Katy Jordan
Key Partners
