In this paper we evaluate a Norwegian vocational training rehabilitation program by comparing employment outcomes of trainees and nonparticipants using nonexperimental data. A matching estimator is used to calculate the training effect for different subgroups of the sample. We demonstrate how...