Reconstructive processes in the free recall of categories.

Three experiments, involving three different groups of Ss, were conducted. For each experiment, one group of Ss never saw the stimulus materials but produced "recall outputs" on the basis of someone else's description of the input list. A second group produced a description of, and then recalled, the list, while a third group recalled the list under standard conditions. Each experiment involved different materials: for Experiment I, exhaustive categories were used; for Experiment 11, high-frequency category items were used; and for Experiment III, low-frequency items were used. Results showed all lists were reconstructible with the degree of reconstructibility following the order of experiments. An examination of descriptions produced indicated significant differences between those leading to lengthy outputs and those leading to less extensive outputs; basically, these differed in terms of the degree to which specific information was or was not present in the description. Results were discussed in terms of Bartlett's theory of remembering.