Production increases both true and false recognition

The production effect is the finding that reading information aloud enhances memory relative to reading information silently. In five experiments, we examined the influence of production on true and false memory in the DRM paradigm. In Experiments 1a, 1b, 3a, and 3b, reading aloud was compared to reading silently. In Experiment 2, reading aloud was compared to reading silently while hearing the words spoken by another voice. In all experiments, reading aloud consistently resulted in better recognition of studied words, but it also consistently resulted in more false alarms to unstudied lures that were semantically related to the studied words. We advance an argument based on current theoretical accounts of false memory wherein reading aloud selectively enhances relational or gist processing—the encoding of shared features across items—rather than item or verbatim processing—the encoding of specific details of individual items. This selective enhancement could be for the shared semantic network (gist), for the shared context of reading aloud (misattributed source memory), or for both. Thus, the benefit of production is best captured by the combination of adding new features (contextual information) together with enriching existing features (semantic information).