(with Rembrand Koning)
When do peers affect performance? In this article, we propose that individuals or teams with many preexisting connections interact with fewer of their nearby peers and are thus less affected by them. We test this hypothesis using data from a field experiment conducted at a startup bootcamp in New Delhi, India. We find that teams whose members have existing relationships with others in the bootcamp interact less with neighboring teams, and thus their performance is unaffected by their peers. In contrast, peers affect the performance of teams without preexisting connections. Our findings highlight how preexisting connections, which are often a source of knowledge and influence, can limit new interactions and thus the power of social incentives and peer effects to improve performance.