Motivational interviewing with late-adolescent/college underage drinkers: An investigation of therapeutic alliance

Many college-bound late adolescents increase their substance use during their first years of college. Recent studies with late adolescents have found that receipt of a brief motivational interviewing (MI) intervention resulted in decreases in their substance use and related problems. However, despite its efficacy, the mechanisms underlying MI remain unknown. This study examined the role of therapeutic alliance and empathy in MI with a late-adolescent substance-using sample. It was predicted that client ratings of alliance would predict behavior change. In addition, it was posited that agreement between therapist and client ratings of alliance would correspond with greater behavior change. Finally, this study explored participants' attitudes towards and desires for therapy, and their reasons for substance use. For the initial assessment, exploratory analyses with the full sample (N= 136) revealed that if they went to therapy, most participants would like assistance with stress reduction and problem solving, along with improving relationships and health. While few concerns about substance use emerged, participants stated that they generally drink to have a good time. For the second component, 55 participants were randomized into an MI (n = 40) and no-intervention control group (n =15), based upon an endorsement of high levels of binge drinking and/or alcohol-related problems. Correlations, linear regression, and MANOVA were employed to evaluate the hypotheses. In contrast with predictions, no significant trends emerged surrounding therapeutic alliance and empathy. Notably, both the MI and the control sample evidenced reductions in substance use and related problems, but without a clear pattern emerging for either group. Post-hoc analyses investigated the possible influence of various factors upon the findings. However, no trends emerged through these analyses. In conclusion, several theories were proposed to explain the effects found. Specifically, the selection criteria for the sample, lack of this sample's substance use severity, the length follow-up period, semester effects, age of the sample, reporting error, size of the control group, and number of MI sessions were proposed as possible confounding factors. In addition, the importance of future explorations of the efficacy of MI and other empirically-supported interventions across developmental periods and risk-taking behaviors was underscored. (PsycINFO Database Record (c) 2010 APA, all rights reserved)