For parolees, reintegration is a highly stressful event that provides a leveraged opportunity for counseling. From a group of first felony offenders about to be released on parole, 18 matched pairs were identified. One member of each pair was randomly selected and offered intensive crisis counseling; the second member of each pair became a control subject, receiving no direct service. However, the interviews, which were designed to compare adjustment upon return home and 7 months later, provided all returnees with a sympathetic listener for about 4 hours. Crisis counseling began 3 weeks before release from prison and continued for 10 weeks after release. Official arrest records and the two evaluation interviews provided data for outcome analysis. Although small, consistent negative findings were observed. In general, data analysis revealed no significant differences between groups. Secondary analysis looked more closely at counseled cases, comparing outcomes against specific components of the counseling process and against key demographic variables. Results suggest that positive outcomes were associated with counseling of good quality while recidivism was associated with counseling of poorer quality. It appears likely that most of the counseling was of too poor a quality to be an important factor in the client's adjustment. Therefore, there was no direct evidence that either the counseled group or the control group was performing more satisfactorily on any of the variables monitored. Reasons for this low level of service include counselor inexperience, overestimation of the ex-con's crisis, experience and inadequate program funding. It is possible that, for some parolees, the presence of a counselor may have increased the sense of immunity from police surveillance and thus discouraged responsible behavior. However, this experiment provides specific experience that suggest modifications in methods and approach that could increase the effectiveness of subsequent work with ex-con populations. Seven appendixes display forms, scales and questionnaires used in the study. Four data tables and 57 references are provided. (Author abstract modified.)