Publication Date



Adolescent cigarette smoking behavior is affected by peer relationships, but how these peer relationships influence the behavior of other adolescent smokers is not well understood. Mixing among cigarette smokers in adolescent friendship networks could indicate peer influence or a homophilous association among likeminded individuals. The goal of this thesis will be to examine a set of adolescent friendship networks to determine if different cigarette smoking behaviors can be predictive of friendship nominations in the network. Examining the structure of social networks requires among other things, inspection of the presence (or absence) of relational ties. Tie formations in social networks are often conditional on the existence of other ties in the network. This conditional dependency along with purely structural network characteristics, creates a unique set of problems from a statistical modeling perspective. Fitted exponential random graph models for a group of adolescent schools will be examined, to assess how the underlying structure of these social networks is influenced by smoking behaviors.

Degree Name


Level of Degree


Department Name

Mathematics & Statistics

First Advisor

Sonksen, Michael

First Committee Member (Chair)

Michael D. Sonksen

Second Committee Member

Yan Lu

Third Committee Member

Nancy Brodsky




social network analysis, exponential random graph models

Document Type