The aim of this study was to investigate the respective and interactive effects of memory aids, derived from the levels-of-processing framework ( Craik and Lockhart, 1972; Lockhart and Craik, 1990), depending on task characteristics in young healthy adults. To this end, semantic word lists, active encoding, and retrieval supports (i.e., cued-recall and recognition tests) were used as memory aids. Two experiments were designed to manipulate presentation duration and organization of the to-be-remembered items, respectively. In this context, we expected that difficult task conditions, such as short presentation duration and unorganized lists, would reduce memory performance without cognitive aids, but that the benefit from interactions between the memory aids would be greater in these difficult conditions. The results are consistent with our assumption, indicating increased interactions, but only when difficulty is explicitly reported by the subjects. These results are discussed in the Bäckman and Dixon's model of compensation (e.g., 1992) and using the notion of awareness. Moreover, we propose that our results could be useful in the design of memory-aid software, and in the domain of eyewitness testimony.