Relating information to the self (self-referent encoding) has been shown to produce better recall than purely semantic encoding. This finding has been interpreted as demonstrating that self-reference produces a more elaborate memory trace than semantic encoding, and it has been cited frequently as evidence that the self is one of the most highly elaborated structures in memory. The experiments reported in this article challenge this interpretation of the self-reference effect by demonstrating that self-referent and semantic encodings produce virtually identical free recall levels if they are first equated for the amount of organization they encourage. On the basis of our findings we conclude the following: (a) Organization, not elaboration, is responsible for the superior recall performance obtained when information is encoded self-referentially, and (b) organization is not a necessary component of selfreferent encoding and can be orthogonally varied within self-referent and semantic encoding tasks. Finally, we discuss how a single-factor theory based on organization can account for many of the selfreferent recall findings reported in the literature. One of the most influential approaches to the study of human memory has been the Depth of Processing (DOP) framework proposed by Craik and Lockhart (1972). These investigators suggested that retention of a memory trace is determined by the nature of encoding operations carried out on stimulus material. Deep, meaningful analyses such as those prompted by semantic encoding tasks allow formation of a more durable trace than do shallow, structural analyses of the sound or appearance of stimuli. Until 1977, semantic encoding was commonly considered the optimal way of achieving good retention (e.g.