Examining proximity exposure in a social network as a mechanism driving peer influence of adolescent smoking.

Background: Adolescent peers' influence on tobacco smoking is a dynamic process affected by close friends and other network peers. Although research has examined the influence of immediate friends on smoking behavior (i.e., by cohesion exposure), the influence of all peers according to closeness (i.e., proximity exposure) remains unknown. This study introduces proximity exposure as a potential driver of peer influence.

Methods: Using the Teenage Friends and Lifestyle Study dataset, we examined 160 adolescents followed for 3 years and assessed their friendship ties and health behavior. Proximity exposure was calculated as the proportion of an individual's network peers who smoked, considering their distance from the individual. Path analysis was conducted with cross-lagged models testing the effect of proximity exposure on smoking frequency over time.

Results: Among nonsmokers without cohesion exposure (n = 80), proximity exposure at year 1 was significantly associated with smoking initiation by year 3. Path analysis (n = 160) indicated that smoking at year 1 predicted cohesion exposure by year 3. When proximity exposure was included, the effect of smoking on cohesion exposure was lost. Early smoking predicted future proximity exposure. However, the predictive value of early proximity toward future smoking was stronger.

Conclusion: These results suggest that proximity exposure can predict smoking even among nonsmokers without direct ties to friends who smoke. In support of a peer selection hypothesis, early smoking predicted friendship formation with smokers through cohesion. Conversely, in support of a peer influence hypothesis, proximity exposure predicted smoking. Researchers may consider developing interventions that decrease proximity exposure among adolescents.