An experiment investigated the relationship between the ability to encode verbal stimuli during an attentional lapse. The task employed a variation on the sustained attention to response task (SART, Robertson, Manly, Andrade, Baddeley, & Yiend, 1997) which involved the detection of an infrequent target against a background of words. As a manipulation, participants were either instructed to encode the stimuli or were merely exposed to the stimuli. Retrieval was measured using process dissociation. Irrespective of the instructions given to the participants during the task, participants were more likely to retrieve information on the basis of recollection after an error was made than before, whilst the likelihood of retrieving information on the basis of familiarity remained invariant over the same period. The implications of this result for methods of investigating subjective experience are discussed.