Malingered amnesia, feigning loss of memory, is a recurring phenomenon in criminal investigations of violent crimes. Little research has been conducted to investigate its characteristics, or the consequences of withholding of information on later memory for the event. Participants were tested as perpetrators after having read a story about ‘themselves’ committing a violent crime. At the first test occasion, one group tested on what they actually recalled. A second group was not tested at all. The two remaining groups were asked to simulate memory impairment by making either omission errors or commission errors. All groups responded genuinely at the second test occasion. Test results showed that different kinds of simulation had different impact on later recall. Results pertaining to free recall showed that the performance of the simulators increased from the first test occasion to the second but did not reach the same level as for those individuals who responded genuinely at both test occasions. When simulating, individuals omit some bits of information and retrieve others, possibly leading to retrieval-induced forgetting. Simulation by omission caused more detrimental effects than simulation by commission on free recall, while the reverse was true for responses to open-ended questions. Simulators showed no impairments on the multiple-choice questions. Since it cannot be stated clearly whether these results is owing to non-rehearsal or a combination of non-rehearsal and retrieval-induced forgetting, future research may focus on distinguishing between these two possible explanations by incorporating these variables into the research design. 5 tables, 35 references, appendix