Segregating deviant youths into groups with similar deviant peers is a common practice in the fields of education, mental health, and juvenile justice. Although this practice may prevent deviant youth from corrupting nondeviant youth, it may aggravate and escalate the problems of deviant youth, since they are continually interacting with and being influenced by the attitudes and behaviors of other deviant youth. This analysis focuses on issues associated with this policy. The paper first shows that the number of deviant juveniles--relative to the size of the relevant population or to the number of assignment options--could determine whether the harm-minimizing assignment calls for diffusion, segregation, or some of both. The second issue examined was the way in which individual behavior caused collective harm. In cases where bad behavior has a direct, detrimental effect on others who share the assignment, including youths who are not deviant, then there is a stronger case for segregation compared to harm that occurs outside of the location where interaction with other youth occurs. A third issue discussed pertains to the capacity for behavioral control. The ability to control behaviors is a factor in the choice between segregation or integration of deviant youth. The structuring of group interactions to maximize the control of harms caused by negative behaviors may make the integration of deviant and nondeviant youths more attractive. A review of the relevant literature does not provide sufficient empirical findings to offer clear guidance for harm-minimizing assignment strategies. 40 references