In this study, a mixed group of 60 substance-dependent patients were randomly assigned to 12 weeks of treatment in either a high-structure, behaviorally oriented (HSB) or a low-structure, facilitative (LSF) individual counseling style. We tested the hypothesis that patients with a more severe pretreatment drug problem will realize greater treatment benefit in HSB counseling, while those with a less severe problem will benefit more in the LSF approach. Six counselors provided the treatments in a counterbalanced design that controlled for possible differences in counselor effectiveness. Treatment benefit comparisons with respect to the counselors' posttreatment ratings, the number of counseling sessions attended, reduction in problem severity, and substance use during treatment were consistently in the hypothesized direction. These findings provide at least partial support for the notion that treatment benefit for substance abuse patients can be improved through appropriate patient-treatment matching on the basis of addiction severity.