The role of negative appraisals in explaining effects of stress and a preventive intervention on substance use

Youth from divorced families are an important group for studying pathways to substance use problems given their increased risk for heavy substance use. Moreover, research testing mechanisms underlying intervention effects on substance use among youth from divorced homes has implications for reducing the public health burden. Using data from a 6-year longitudinal follow-up sample of 240 15-19 year-old adolescents who participated in a randomized experimental trial of a preventive intervention for divorced families, two studies tested one pathway to adolescent substance use, children's negative appraisals about negative events. Study 1 tested whether the relation between childhood negative events and adolescent substance use was significantly stronger for youth with greater negative appraisals than for youth with fewer negative appraisals. Study 2 tested a mediational model of the relation between childhood negative events, negative appraisals, and adolescent substance use using both a passive longitudinal approach and an experimental approach. The passive approach used data from individuals who participated in the literature control condition of an evaluation of a preventive intervention for divorced families. It was hypothesized that negative appraisals would explain a hypothesized relation between childhood negative events and adolescent substance use. The experimental approach used data from both intervention and literature control groups of the same evaluation. It was hypothesized that the program would reduce children's negative appraisals, and that reductions in negative appraisals would account for program effects in reducing substance use six years later. Multiple regression under a path modeling framework was used to test moderation in Study 1. Structural equation modeling was used to test mediation in Study 2. There were no significant mediating or moderating effects of negative appraisals on relations between negative events and substance use. There were no program effects on negative appraisals, and negative appraisals did not explain risk by program effects on alcohol and marijuana use. There were no significant main effects of program or risk by program effects on binge drinking or cigarette smoking. Limitations of this research that may have contributed to null findings, including measurement problems, cognitive and developmental processes, model misspecification, and sample characteristics, suggest areas for future investigation.