Room: Phillips 203
Time: Mon 15:00 PM-16:30 PM

New Directions in Estimating Peer Effects in Health

Chair: H. Shelton Brown

Session Description

This session brings together new empirical methods using richer models in order to provide more credible estimates. The authors focus on the large empirical challenges in estimating peer effects by confronting the endogeneity of peer groups, interdependent decision making, and shared environmental factors. In the first paper by Trogdon, peer formation is explicitly modeled using agent-based simulation methods. The authors extend the seminal work of Burke and Heiland (2007) and pay careful attention to the mechanisms that may lead to peer group formation based in part on weight as well as the potential social influences of peers on weight outcomes. The second paper by Eisenberg et al, the authors use new data from the University of Michigan to extend knowledge about potential peer effects in mental health outcomes by leveraging random roommate assignment. In the third paper by Fletcher, the authors focus on friendship networks of adolescents in determining smoking and drinking behaviors. The paper presents a new empirical methodology to leverage arguably quasi-random variation in friendship networks using a cross-cohort design and the unique features of the Add Health data.

Key Terms None

Session Organizer: Jason Fletcher (Yale University)

---

Presentations

1. An Agent-Based Model of Social Influence on Obesity

   Presenter: Justin Trogdon (RTI International)

   Discussant: Debjani Kanjilal

2. Peer Effects in Mental Health and Health Behaviors: Evidence from College Roommates

   Presenter: Daniel Eisenberg (University of Michigan)

   Discussant: Stephanie von Hinke Kessler Scholder

3. Friendship Networks and Health Behaviors

   Presenter: Jason Fletcher (Yale University)

   Discussant: Jeff DeSimone (University of Texas at Arlington)