The experiments compared the identification of children aged 3-15 (N = 307) and undergraduates (N = 384). Correct identification rates tended not to vary across either age of subject or identification procedure. However, children showed a significant tendency to guess as indicated by their lower rate of correct rejection when the target was absent. The tendency for children to make false positive choices was particularly evident with showups (presentation of a single individual, the suspect). The tendency of children to guess seriously compromises the reliability of eyewitness identification by children. Neither training nor alternative identification procedures have been successful in reducing the problem. Even at the end of elementary school, children's identification performance was inferior to adults. This research also demonstrated the showup is a dangerous procedure. Not only did showups appear to reduce correct identification rates relative to lineups, but the risk of false identification was as great or greater. Notes, tables, references