The research reflected earlier studies indicating that the outcome of scent identification is significantly affected by factors such as the way in which the odors are presented and whether the dog is assessed as willing to work. Each of the 6 dog/handler teams was used in 10 experiments that varied in whether the suspect was the same as the perpetrator and in the type of corpora delicti. Results revealed that a performance check regarding the dog's ability or willingness to work produces a significant positive effect on the results. Findings indicated that scent identification following the improved performance check used in the study are reliable enough to be a useful forensic tool. Nevertheless, scent identification should be performed only by dogs that are part of a comprehensive quality control effort in which the performance level of each dog is monitored so that court's can be informed of the dog's reliability if necessary. Tables and 11 references (Author abstract modified)