Ehlers and Clark [(2000). A cognitive model of posttraumatic stress disorder. Behaviour Research and Therapy, 38, 319-345] propose that a predominance of data-driven processing during the trauma predicts subsequent PTSD. We wondered whether, apart from data-driven encoding, sustained data-driven processing after the trauma is also crucial for the development of PTSD. Both hypotheses were tested in two analogue experiments. Experiment I demonstrated that relative to conceptually-driven processing (n = 20), data-driven processing after the film (n = 14), resulted in more intrusions. Experiment 2 demonstrated that relative to the neutral condition (n = 24) and the data-driven encoding condition (n = 24), conceptual encoding (n = 25) reduced suppression of intrusions and a trend emerged for memory fragmentation. The difference between the two encoding styles was due to the beneficial effect of induced conceptual encoding and not to the detrimental effect of data-driven encoding. The data support the viability of the distinction between data-driven/conceptually-driven processing for the understanding of the development of PTSD. (C) 2008 Elsevier Ltd. All rights reserved.