The study focuses on three main questions: (1) selection for mediation, i.e., what cases are brought into mediation and why, examined on the basis of a questionnaire sent out to mediation offices and interviews with police and prosecutors; (2) the significance of mediation to the victims and offenders who participated, examined by following cases in mediation and interviewing victims and suspects; and (3) the effect of mediation on recidivism, examined by comparing offenders who participated with a matched group who did not participate. In general, prosecutors and police submitted cases to mediation on pragmatic grounds - to ease their own job - but for some there was a strong commitment to the ideology of mediation. Both offenders and victims believed that mediation as a procedure was more pleasant than what they imagined a trial would be. The study indicates that simple and easy measures usually fail to have a significant effect on recidivism and suggests the need for further research, research of a more sophisticated, preferably longitudinal, design to ascertain the effects of mediation on the probability of recidivism.