In a study of 5285 8th graders from the Gang Resistance and Education Training (G.R.E.A.T.) research, this study applied Travis Hirschi’s social bonding theory to examine the curriculum’s efficacy in increasing conventional bonding (friends with positive peers, succeeding at education etc.) and decreasing non-conventional bonding (drug use, truancy, law violations etc.). The results suggest that across the full models, multiple group models (i.e., receive the G.R.E.A.T. curriculum or not) and models with indirect effects, attachment to parent, education and positive peers is the most consistent construct for increasing youth bonding. In the multiple group model, commitment (i.e., feelings about joining gangs, being involved with gang behavior etc.) is reduced significantly for those youth who received the G.R.E.A.T. curriculum versus those who did not. In the full model, belief about gangs in school, and pressure to join gangs, as well as involvement with delinquent peers and drug using peers are significant constructs for increasing bonding in the full models (with and without indirect effects), but are not significant in the multiple group model.
Regarding the latent construct G.R.E.A.T. (i.e., gang knowledge and knowledge about gang influence - selling drugs for power, interfering with goals and neighborhood peace) these variables are consistently significant across models regardless if youth received the G.R.E.A.T. curriculum or not. Regarding the manifest variables, youth use drugs because of peer pressure, and youth use drugs because of low self-esteem, youth view these two factors as consistently salient across all models.
Results for bias corrected, resampled confidence intervals for indirect effects on latent constructs suggest that large samples and large resampling (i.e., over 5000) are required for stability of loading estimates.