The studies were located through a computer search using the PsycInfo database from 1984 to March 1997 and the keywords eyewitness, age crossed with identification, child crossed with lineup, witness, photo and identification and witness and identification. Studies were restricted to those that included children and an adult comparison group. Information was coded for each study on age, presence versus absence of the target person in the photographs shown to participants, the method of lineup presentation, and the mode of target exposure. Preschoolers were less likely than adults to make correct identifications. However, children over age 5 years did not differ significantly from adults with regard to the rate of correct identifications when the target person was present. In addition, children of all ages were less likely than adults to correctly reject a lineup in which the target person was absent. Moreover, even adolescents did not reach an adult rate of correct rejection. Furthermore, compared to simultaneous lineup presentation, sequential lineups increased the child-adult gap for correct rejections. Finally, providing child witnesses with identification practice or training did not increase their correct rejection rates. Findings suggested that children reach the same rate of accurate hits as adults earlier than the 12 years of age suggested in earlier research, but that the rejection rates for target-absent lineups are lower for both children and adolescents than for adults. Other research suggests possible reasons for the differences. Further research is recommended. Tables, footnotes, and 58 references (Author abstract modified)