Most past research on detecting deception has relied on the assumption that liars often fabricate a story to account for their whereabouts, whereas truth tellers simply recall an autobiographical memory. However, little research has examined whether liars, when free to choose the topic of their own reports, will actually choose to fabricate information rather than use a different strategy for constructing their lies. The authors describe two studies that evaluated liars' strategies for selecting the content of their lies when given the freedom to choose whatever content they desired. In studies 1 (N= 35) and 2 (N= 22) participants (a) described a truthful story in order to identify a salient event, then (b) lied about the event, and finally (c) described their strategies for choosing the content of the reported lies. Liars overwhelmingly chose to report a previously experienced event for the time period they were to be deceptive about (67 percent and 86 percent in studies 1 and 2, respectively). The majority of discrete details reported were experienced, occurred relatively frequently, occurred relatively recently, and were typical or routine. These findings have significant implications for the development of cognitive-based interventions for detecting deception. In particular, some methods of deception that rely on content analysis may be ineffective if liars choose to report previous experiences rather than outright fabrications. Abstract published by arrangement with John Wiley & Sons.