The study sample was selected from a previously conducted epidemiological study that involved a 10-year follow-up of 401 violent offenders. Individuals with "unexpected outcomes" were selected according to the following criteria: a baseline score greater than 12 points on the historical part of the HCR-20, which assesses risk for violent reoffending; at least 2 convictions for violent crimes and at least 5 convictions for any crime; no detention in a forensic psychiatric unit or imprisonment during the past 5 years at follow-up; and no reconvictions for any crime during the past 10 years. Only six individuals (all men) met all four criteria; four of them agreed to participate in the current study. These men were assessed through semistructured interviews to identify common themes related to desistance from crime. All of the men emphasized a specific event or factor that motivated them to make a conscious choice to stop their criminal behavior, such as an adverse experience linked with crime or during incarceration. This decision was the beginning of a long pathway to behavioral change. Restraining factors that supported the decision were family life and social-avoidance strategies. All participants reported having quit or significantly reduced drug use, including alcohol. The men constantly assessed situations in an attempt to avoid "triggers" for criminal behavior, such as avoiding meeting certain people and frequenting certain locations. One difference between the findings of this study and those of other studies of ex-criminals is the lack of establishment of links with conventional society in the current sample. For these four men desistance from crime involved avoidance of social interactions. They did not feel comfortable or safe with others, and they were unsure of their reactions to various situations and others' responses to them. 1 table and 59 references