This study proposes to use network analysis theory to tailor a nationally recognized effective substance abuse prevention program, Towards no Drug Use (TND). Network analysis theory will be used to identify group leaders and group members. TND will then be adapted to be used by these groups, completing the exercises and lessons in TND within these naturally occurring groups to determine if the effectiveness of TND can be improved by harnessing the power of peer influence. Recent evidence from an earlier trial has shown that this network approach to health promotion can be effective. The shortcoming to this prior work, however, has been that it has been implemented with a novel prevention program for which prior effectiveness had not been demonstrated. Thus this study provides the opportunity to test the network adaptation on an existing evidence-based program. This study will randomly assign 46 continuing high school classrooms (N=800) to receive TND or TNDNetworked and compare outcomes on substance abuse behavior at one and two year intervals. The two TND curricula will be similar with the exception that all group activities in TND-Networked will be conducted with groups based on students' selection of peer leaders. It is expected that TND-Networked will be more effective than TND by resulting in lower incidence and prevalence of smoking, marijuana, and hard drug use. This study will also use the network data to test hypotheses concerning peer influence. Specifically, we will test whether peer influence, measured with social network analysis, mediates the influence of TND. This study also proposes to conduct basic research on the role of social network influences on substance abuse behavior among these high-risk youth. We propose to study the ethnic composition of friendships choices, and investigate whether this facilitates or impedes substance abuse. It is linked with Project 1 in that it uses the baseline data from Project 1 and will investigate whether associative memories are affected differentially in the two conditions, and whether they vary by network choices.
Showing the most recent 10 out of 65 publications
