Fergus I. M. Craik and Michael J. Watkins conducted two experiments which oppose many widely accepted models of memory, stating that an items length of stay in short term storage (STS) has an effect on the item being transferred into long term storage (LTS).
Previous researchers postulate the more an item is rehearsed in STM there is a better chance of the item being transferred into LTM, for recall later, this can explain the negative recency effect in free recall; items presented at the end of a list are not rehearsed as often, being poorly retrieved later.Craik and Watkins experiments showed that neither the amount of time an item stayed in STM nor the number of overt rehearsals was related to subsequent recall (Craik and Watkins). They have concluded that the maintenance and elaborative aspects of the rehearsal can by separated. Maintenance does not lead to an improvement in memory performance (Craik and Watkins).
Experiment # 1 Participants in this experiment were instructed to listen to a series of word lists, reporting only the last word beginning with a specific (critical) letter after each list was completed.The participants were given the critical letter prior to the presentation of the list, therefore able to ignore all other words not containing the critical letter. Once another word with the critical letter was presented the participant could drop the first word and rehearse the next.
Continuing until the list ended. Once the list ended the participant wrote down the last critical word. Three rates of presentation were used; the time a critical word was held in STS varied both as a function of presentation rate, and the number of noncritical words monitored between presentation and replacement (Craik and Watkins).Following the presentation of all the lists, participants were unexpectedly asked to recall as many words as possible from all the lists presented previously. For each rate of presentation, the delayed recall of both the replaced and reported critical words was examined as a function of the number of noncritical presentations during which words were held (Craik and Watkins). If time in STS predicts LTS retrieval, then final recall performance should increase directly with the number of items monitored during the retention of the critical words, and inversely with presentation rate (Craik and Watkins).Method Fifty four introductory psychology students from North East London Polytechnic participated in this experiment.
The experiment was carried out in three sessions with eighteen participants in each group, randomly allocated into three groups of six, each group given a different set of critical words. Participants were informed at the beginning that they formed the control group of a perception-memory experiment; whereas other participants had a task with a substantial memory load. They were to monitor the lists and write down the last critical word.It was stressed that their performance should be virtually perfect. Prior to the presentation of the list each participant was given a card with the list number and critical letter for that list; once the list ended the participants wrote down on the card the last critical word, immediately placing the card into an envelope. In each session, nine lists were presented at each speed, randomized separately for each session. All participants were shown the same twenty seven lists (recorded on tape) of twenty one (one or two syllable concrete nouns) each list was preceded by the list number and the speed of presentation.
Three rates of presentation were used; slow-one word every two seconds, medium-one word every second, and fast-one word every half second. Lists presented at one speed in the initial session were then the presented in the two remaining speeds. A tone was presented at the same rate as the list words signaling the end of the list.
Two further within variables were involved; the first is i-values (nine were used) the number of intervening noncritical words and the second variable concerns the critical words, and whether they were replaced by another critical word or if they were reported as the last.Each list contained three critical words, equally distributed over i-value and presentation time; two were replacement words and one a report word. The order of i-values was essentially randomized. Effort was made to avoid confounding serial position and i-value of the replacement words; making the i-value not totally randomized. The first three serial positions were not used; besides the twenty seven lists participants were given three practice lists. Lists were presented with an intertribal interval of fifteen seconds.
Participants engaged in one-minute arithmetic a task, immediately following, during which time paper was distributed for a final recall test. A ten minute period was allowed for the free recall of any words; critical or noncritical. Results 26. 2 of the twenty seven report words were correctly identified and reported, no participants made more than two errors. The critical words were correctly perceived and retained over their respective intervals.
An analysis of variance revealed significant effects of reporting and presentation rate.The findings reflect the advantage of reported words over replaced words, and slower presentation. The only other significant effect was the second order interaction between rate, i-value and replacement reported. The most important finding is that the non significance of the i-value variable. Discussion Recalling an item from LTS remained essentially independent of its i-values. This result is clearly contrary to the idea that recall probability necessarily increases in direct proportion to the total amount if time an item has been thought about or attended to (Craik and Watkins).The presentation rate effected recall, the slower the presentation the more time the participants had to rehearse the word.
The lack of an i-value effect implies the use of a rehearsal process which was sufficient to retain a critical word in a highly accessible state during the presentation of non critical words; having no long term effects. The present results at first seem inconsistent with evidence indication that retention varies directly with the amount of rehearsal (Craik and Watkins). This discrepancy can be reconciled by postulating two distinct modes of rehearsal, maintaining and elaborate.
Contrary to the models of others, time in STM does not itself lead to long term retention. Experiment # 2 In a previous study it was found that while the immediate recall of data showed the typical large recency effect, the final recall data were at least well recalled in the final recall test. There are two possible explanations of this negative recency effect; terminal items receive fewest rehearsals in immediate recall and are poorly registered in LTS or the type of rehearsal, not the amount of rehearsal, which is critical for later recall.In the present experiment participants were induced to rehearse aloud the last few items in the list at least as many times as they rehearsed the first few. If the number of rehearsals is a critical factor, then negative recency in final recall should disappear, with a slight positive recency effect (Craik and Watkins). If the type of rehearsal is critical, and the extra rehearsals are to maintain, then these rehearsals should prolong the recency effect in immediate recall and not lead to a strengthened trace.
Then it is quality not quantity of rehearsal that is important.The recency positions in the final recall should not be affected by the extra rehearsal period in immediate recall (Craik and Watkins). Participants were given lists of words that they were allowed to recall at the end, other list consisted of a twenty second delay before recall. All participants were informed that they were to recall the last four words of each list. The participants were asked to rehearse aloud and tape recorded, later being counted. Method Sixteen paid University of Toronto undergraduates participated in this experiment, each tested individually.The participants were told that the experiment concerned the effect of rehearsal on immediate retention and that rehearsal must be spoken aloud. Instructions were for free recall but with additional instructions that recall of the last four words was important.
Three practice lists were presented prior to the twelve lists; randomized and presented visually; consisting of twelve common nouns; the first eight words were in lowercase, the remaining four were block capital letters, shown for three seconds, were used in each trial. Six lists were under immediate recall, while the remaining six were under delayed recall conditions.Recall was immediate for half of the trials, signaled by a loud tap.
The other half was given a twenty second delay; participants were encouraged to rehearse during this time, recall was signaled by a loud tap. Participants were not aware of which group they were in at the beginning of the experiment. Participants were given one minute for written recall on each trial. Participants were engaged in a two minute conversation prior to being asked to recall as many words as possible in a final recall test; ten minutes was allowed.Results.
In both conditions; the instructions regarding the last four words resulted in a boost in rehearsal for these particular words. In the delayed condition, rehearsal of the first eight words was increased slightly, rehearsal of the last four words increased substantially by an average of six extra rehearsals per word (Craik and Watkins). Conditions were essentially identical for both conditions; participants were able to maintain the words over the twenty second interval.
The final recall scores for both groups were virtually identical.The mean final recall level for the last four words is less in the delayed than in the immediate condition; despite the great increase in rehearsal for the last four words under the delayed condition. Discussion Time spent in STS does not by itself lead to long term retention.
Although participants were given twenty seconds to maintain items in STS it had no effect on later recall from LTS. Participants in the study were instructed to pay particular attention to the last four words of the list. Participants were not expecting a final recall test, given them no reason to indulge in the complex semantic associative coding.Participants were aware of when the list was ending because of the last four capital block words. Phonemic encoding was encouraged by the necessity to rehearse aloud. It is not easily determined whether there are genuine negative recency effects on lists with as little as twelve words. The present results imply that negative recency is not due to fewer rehearsals of the last few items (Craik and Watkins). In this experiment recency items were rehearsed at least as often as primacy items, the recall of recency items in final recall remained poor.
The type of rehearsal rather than the amount of rehearsal is critical for good long term retention. If the participants encode and maintain the last words at a phonemic level only, final recall will be poor; performance should be enhanced if the participants encode the words in a richer semantic associative fashion (Craik and Watkins). General Discussion The results reported are surprising.
Neither the prolonged maintenance of an item in STS (exp 1) nor a substantial increase in the number of overt rehearsals (exp 2) increased final recall performances (Craik and Watkins).It can be assumed that final recall performance depends on LTS, implying that maintenance of an item in STS does not lead to better registration. This finding is opposed to the conclusions of Waugh and Norman, Atkinson and Shiffirn, and Rundus; all postulate that longer residence in STS.
Or a greater number of rehearsals, led to better long term performance (Craik and Watkins). The present study in conjunction with past findings, demonstrate beyond all reasonable doubt that neither overt rehearsal not maintenance in STS is by itself sufficient to enhance long term memory performance (Craik and Watkins).Rehearsal can be broken down into a maintaining function and an elaborate function.
The above experiment on the role of rehearsal is related to the one being conducted in research methods in many ways. Participants in our study were shown a list of nouns then kept busy by a task designed by the students. Once the task was completed the participants were asked to complete a final recall of the items shown.
Participants were not aware that they were going to be asked to recall the items at the end of the experiment, unless they had previously taken the course.