A total of 449 students were tested for their event memories and autobiographical memories of hearing about the Columbia shuttle disaster of 1st February 2003. Four different groups were tested 2, 18, 27, or 51 days after the event. All participants were then re-tested after 5 months (second session) and again after 1 year (third session) from the first interview. Dissociations between consistency and confidence and between event and autobiographical memories were found. Consistency and confidence in event memories, but not in autobiographical memories, were affected by time. In contrast, repeated testing selectively enhanced autobiographical memories, in accordance with the narrative and rehearsal hypothesis of Neisser and Harsh (1992). For event memories, veridicality was inversely correlated to consistency, which in turn was inversely correlated to confidence, and mainly based on omissions. As regards veridicality, the analyses showed an increase of false memories at long time delays. Results are discussed with reference to recent studies contrasting autobiographical and event memories.