ABSTRACT
This chapter outlines the advantages and demonstrates the use of the Solomon Four Group Design in the context of a multiple groups structural equation model (SEM) with latent means. The data are aggregated at the level of the classroom. Hypotheses that test threats to internal validity such as pretest sensitization, history, and maturation are illustrated and analogs in appropriately constrained models are presented. This analytic approach is then demonstrated with data from an ongoing D.A.R.E. evaluation. Pretest sensitization was found for two of the D.A.R.E. core constructs; acceptance of risky behavior and institutional bonds. Classrooms of students who participated in D.A.R.E. reported higher self-esteem, greater resistance to peer pressure, stronger institutional bonds, and were less accepting of risky behaviors.
