Low accessibility of threatening information is a crucial element in the theory of repression. Despite many attempts, no empirical evidence has been found for such an inverse memory bias in normal subjects. An alternative approach is to study repressive processes as an individual dierence. Adopting such a trait approach, several studies have shown impaired memory for threatening memories of real-life episodes in so-called repressors. Several other studies have yielded evidence of avoidant processing of mildly threatening information in the early stages of information processing in repressors. Thus far, little evidence for a similar avoidant bias in memory has been oered for repressors. To demonstrate the presence of memory deยฎcits for mildly threatening information, we tested recall and cued recognition after a 30 min delay, for (i) threatening, (ii) pleasant, and (iii) neutral words in repressors and non-repressors. No proof was found for poorer recall and recognition of threatening words in repressors, or for lower numbers of threatening semantic intrusions in these individuals, while the power of the study was adequate. While repression-related avoidant processing may be present for even mildly threatening stimuli at earlier stages of information processing, the results of the present study indicate that repressed memory may only show up for relatively intense emotional events.