A total of 236 children (7–8-year-olds and 13–14-year-olds) attempted identification from either target-present (TP) or target-absent (TA) video parades, after either two days or two weeks, following exposure to a live target. It was found that with TP line-ups, the most frequent response was a correct identification, but in the TA condition, the most frequent response was a false identification. Under both TP and TA conditions, the 3-way interaction of delay, age and response was significant. Further analysis revealed that delay was the major contributor of variance, causing an increase in errors for both age groups in TP delayed line-ups but only for the younger age group in TA delayed line-ups. Confidence varied as a function of type of identification response made, with the highest confidence being given to a correct identification in a TP line-up. However, response confidence entered into a number of interactions involving both delay and age. Copyright © 2011 John Wiley & Sons, Ltd.