The Generation Effect: Delineation of a Phenomenon

Five experiments are reported comparing memory for words that were generated by the subjects themselves with the same words when they were simply presented to be read. In all cases, performance in the generate condition was superior to that in the read condition. This held for measures of cued and uncued recognition, free and cued recall, and confidence ratings. The phenomenon persisted across variations in encoding rules, timed or selfpaced presentation, presence or absence of test information, and between- or within-subjects designs. The effect was specific to the response items under recognition testing but not under cued recall. A number of potential explanatory principles are considered, and their difficulties enumerated. It is concluded that the generation effect is real and that it poses an interesting interpretative problem. This is an empirically oriented article whose purpose is to report a set of simple experiments that establish the existence of a robust and interesting phenomenon of memory. This phenomenon, called the generation effect, is robust in that it manifests itself across a variety of testing procedures, encoding rules, and other situational changes. It is interesting in that it does not seem to be easily or satisfactoril y accommodated by any of the currently familiar explanatory notions. We expect that once the phenomenon is described in its initial form, it will be the subject of wider experimental analysis and will eventually become better understood. In contrast to the usual objective reasons for embarking upon a line of research, the present work was neither initiated by any extant theoretical issue nor inspired by any previously published findings. It was carried out with the sole purpose of arriving at a