To examine the effectiveness of brief behavioral interventions adapting the principles and techniques of Motivational Interviewing (MI) to four behavioral domains: substance abuse, smoking, HIV risk and diet/exercise. We conducted a systematic review of 29 randomized trials of MI interventions. Data on methodological quality were extracted and tabulated. Between-group behavior change effect sizes and confidence intervals were calculated for each study. Due to varying intervention time lengths, targeted problem behaviors, settings and interventionists' backgrounds and skill levels, outcomes were not combined meta-analytically. Sixty per cent of the 29 studies yielded at least one significant behavior change effect size. No significant association between length of follow-up time and magnitude of effect sizes was found across studies. There was substantial evidence that MI is an effective substance abuse intervention method when used by clinicians who are non-specialists in substance abuse treatment, particularly when enhancing entry to and engagement in more intensive substance abuse treatment treatment-as-usual. Data were inadequate to judge the effect of MI in the other domains. Client attribute-treatment interactions were understudied and the sparse and inconsistent findings revealed little about the mechanism by which MI works or for whom it works best. To determine more effectively how well MI works in domains other than substance abuse and for whom it works best in all domains, researchers should study MI with risk behaviors other than substance abuse, while examining both interactions and the theoretical components of MI.