The assertive case management (ACT) study by Bond and colleagues illustrates the problems of evaluating new mental health service modalities applied to multi-dimensional problems. Both characteristics of large urban areas and increasing consumer self-awareness affect implementation of random assignment and follow-up studies. In the study reviewed here, possible lack of fit between study subjects and the control condition, a drop-in center, may have contributed to a high attrition rate. As most of the controls never received the treatment, neither ACT nor the drop-in center were adequately tested. And without explication of how the control condition relates to other peer-oriented interventions, study findings cannot be generalized to self-help. The authors' conclusions concerning self-help are therefore not supported by their findings.