objective: to describe the components and development of the early prevention of violence program in the city of medellín, colombia, and to evaluate the results of its first phase, three years after implementation. methods: a before (2001) and after (2004) study of four variables-direct aggression, indirect aggression, prosocial behavior, and scholastic achievement-was conducted among a convenience sample of 339 program participants and their families. results: several program benefits were noted. decreases in both direct and indirect aggression were observed, though the latter was reduced only in girls and in those over 12 years old. prosocial behavior increased among children of all ages and both genders, including those who exhibited low levels of prosocial behavior in 2001. in addition, improved school performance was seen in the group as a whole. results may have been affected by some changes to the prevention program？s implementation and by the dangerous nature of the neighborhood, which limited the home visits. conclusions: the program seems to be an effective intervention for highly aggressive children, among whom a decline in direct aggression was observed. it also offers preventive benefits, as evidenced by the rise of prosocial behaviors in less aggressive children.