Can we select the right peers in Indian Education? Evidence from Kolkata: Research by Asad Islam

Can the peers a student lives with truly change their academic destiny? Inspired by the rigorous research of Asad Islam and his colleagues, I want to share powerful lessons from a path-breaking study at a Kolkata college. Through the lens of random dormitory assignments, this research uncovers how much—and in what ways—peers influence higher education outcomes.

Key Findings: The Power of Peer Ability

This study used a unique natural experiment: students of varying backgrounds were randomly assigned to dormitory rooms, eliminating the usual self-selection that clouds most peer effect estimates. The findings?

• Peer ability matters—a lot: Roommate ability had about one-third the impact on student academic achievement as the student’s own prior ability.

• Biggest gains for the most vulnerable: Students from non-urban regions and non-English-medium backgrounds made the largest academic strides when paired with more able peers.

• Non-linear effects: The impact isn’t the same across the board—high achievers benefit the most from being grouped together, but the right mix can also lift up the least advantaged.

Why Does This Matter for Policy?

What happens when colleges try to “engineer” peer groups? The researchers simulated several assignment strategies:

• Assortative matching (grouping by ability): Maximizes average grades but can widen gaps.

• High-low matching (mixing top and bottom ability): Raises the minimum grades and reduces disparity, especially helping weaker students.

• Middle-based/random matching: Falls between the two extremes.

The choice here depends on policy priorities: maximizing excellence, reducing inequality, or seeking a balance.

Mechanisms: Demonstration Over Competition

The study reveals that peer effects mainly operate via demonstration—students mimic and adopt the motivation and habits of their peers, particularly in the first college year. Group study and tuition with peers help academic performance, especially for those who need it most. Contrary to some expectations, fierce competition is not the main driver—mutual support and modeling matter more.

Diversity and Social Mobility

India’s educational landscape is incredibly diverse. This research shows that mixing students from different backgrounds—urban, rural, English, non-English, and social classes—can yield educational gains and support social mobility. The biggest benefits occur when strong students from more privileged backgrounds are grouped with talented students from less privileged ones.

Global Implications

Asad Islam’s scientific rigor and real-world focus offer both inspiration and a roadmap for education systems around the world. Randomized peer-group studies can guide housing, grouping, and support programs to both raise performance and close achievement gaps.