ABSTRACT
This chapter describes methods to test contrasts among mediated effects in multiple mediator models. First, multiple mediator models are described as a way to summarize drug prevention programs. Second, technical material on testing of specific mediated effects and the standard errors of specific mediated effects is presented. A methodology to test contrasts among mediated and direct effects in multiple mediator models is then presented and illustrated in the evaluation of a multicomponent program to prevent anabolic steroid use among high school football players. Limitations of multiple mediator models and methods to test contrasts in these models are discussed.
