JOURNAL OF VERBAL LEARNING AND VERBAL BEHAVIOR 9, 684-688 (1970) Immediate Free Recall and Three-Week Delayed Recognitio#

Twenty-five Ss were shown 10 lists of words, each list followed by a free-recall test. Overt rehearsal of the word lists was required uring study and was tape-recorded. Three weeks later Ss returned for a recognition test on these words. Initial recall and three-week delayed recognition were increasing functions of amount of rehearsal ccorded the item. The prob-ability of delayed recognition decreased asa function of the initial serial study position, and appeared to increase as a function of its output position on initial recall. Further analysis of this latter esult suggested that delayed recognition first rises and then falls as a function of output position in initial recall. In a recent study by Craik (1970) Ss were presented with 10 lists of words with a single immediate free-recall test following each list. At the end of this series of trials a final test was given in which Ss were asked to recall as many words as possible from all of the preced-ing lists. Two interesting results emerged from the study. First there was a marked difference between the shapes of the serial position curves for immediate and final recall. The recency effect, which had been quite strong in immediate recall, was absent from final recall; in fact, final recall was found to decrease over serial input positions. The second result con-cerned the relationship between the prob-ability of recall on the final test and the posi-t ion of that item in S's initial output protocol: the later an item appeared in S's initial recall, the higher was its probabil ity of recall on the final test. In the present study Ss were given a series of word lists with each list followed by a free-recall test on that list. During the study of these lists Ss were required to rehearse list items aloud; this overt rehearsal was re-corded and analyzed following the procedur