The purpose of the present program of research was to examine the influence of a previous interaction and personal knowledge of a culprit on eyewitness recall and recognition accuracy. In the main study, participants (N=184) watched a video of a crime wherein the culprit was someone who: (1) they previously interacted with and, during that interaction, the culprit shared personal information about herself (more familiar condition), (2) they previously interacted with, although they received no personal information about the culprit (less familiar condition), or (3) they had no previous exposure to (stranger condition). There were no significant differences in participants’ recall accuracy across the conditions. However, participants in the more and less familiar conditions were significantly more likely to correctly identify the culprit from a target-present lineup; there were no significant differences in the target-absent lineup. Also, more and less familiar participants were significantly more confident in their ability to correctly identify the culprit prior to the lineup task and during the lineup task than those in the stranger condition. Moreover, the confidence-accuracy relationship was strongest for those in the more familiar condition. A follow up study, completed approximately 14 months after the original study, was conducted to examine the influence of familiarity and delay on eyewitnesses who initially provided a correct lineup response. Delay was associated with a significant decline in accuracy rates and confidence for those in all conditions. Overall, results indicate that previous interaction, regardless of whether personal information is shared, may enhance eyewitness identification accuracy only when the culprit is in the identification lineup (i.e., when the suspect is guilty). Also, more familiar eyewitnesses may have a better sense of their identification accuracy, as indicated by their confidence judgements.