Computer-assisted enhancement of memory in children
Supervisor: Dr. C. Singleton
We are grateful to Dr. Singleton who kindly provided a copy of this unpublished research of which the Abstract and Discussion are provided here.
The study investigated the possible effects of computer-aided memory training. The experiment involved 56 children from two age groups: 9:00 - 10:00 and 10:01 - 11:05.
For memory training the Mastering Memory Program (Mitchell, 1996) was used in two conditions:
- with the experimenter in an active role of a teacher / therapist providing suggestions regardingmemory enhancement strategies;
- where the subject used the program on its own.
Short-term memory was assessed by means of 7 of the British Ability Scales - II tests:
- Recall of Objects Verbal Immediate,
- Recall of Objects Spatial Immediate,
- Recall of Digits Forward,
- Recall of Digits Backwards,
- Recall of Objects Verbal Delayed,
- Recall of Objects Spatial Delayed.
(Elliott, Smith & McCulloch, 1996)
The same tests were administered twice: before and after training to the experimental groups, and at approximately the same times to the control group.
The outcomes of the statistical analysis of the results of the tests indicated that both groups which had the memory training obtained significantly higher scores on 3 of the BAS-II tests than the control group.
There were no significant differences between the two experimental groups, although, the descriptive statistics suggested greater benefit for the group which used the Mastering Memory Program together with instructions from the experimenter regarding mnemonic techniques.
The purpose of this project was to examine the effects (if any) of the computer-aided memory training in children. The main findings indicated that for some of the memory tests taken after the training there were significant differences between the 3 groups of children.
The test which has showed consistently significant differences among all 3 groups was the BAS-II 'Recall of digits Backwards' test. There was, however, a 3-way interaction (for group, sex and age) found for this test (F=6.8 df=(s,55), p < 0.05). Nonetheless, the significance level (p<0.01) of the main effects found and the nature of the data justify the acceptance rather than rejection of those effects (Howell, 1989). The means for the groups on this test suggest that the 'instructions' group had the comparatively highest mean scores (66.5), followed by the 'no instructions' group (55.7), and the control group with the smallest mean (of 35.5). The simple interpretation of such results could be that it was the training which was the cause of such differences, where the Mastering Memory Program used as a tool rather than on it own had the biggest effects. This, however, must be treated with caution in the context of: the nature of the 'Digits Backwards' test, the nature of the program used for the training, and the results obtained from other BAS-II memory tests.
Beech & Singleton (1997) claimed that even though the digits span tests (both backwards and forward) are not without their problems, they are considered by many practitioners to be the best indicators of short term memory (STM). The 'Digits Backwards' test involves immediate recall of verbally presented digits in the reverse order, a process which puts heavy demand on working memory, which holds the digits while one 'turns them round' in order to present them (verbally) in a backward fashion. Such a process can be regarded as a complex one, which perhaps requires a skill or learning of the best method (e.g. realisation that to achieve the aim one has to first store the digits in STM and only then attempt to 'read them out' backwards). It involves metamemory to a much greater extent than the other tests.
The nature of the Mastering Memory program is such that it presents pictures or names of objects and then demands immediate recall in the way of choosing the items from available options. It is of great importance to note that the 'instructions' group did not use the 'only vocal' modality at all (they did use it together with the visual presentations), and the 'no instructions' group used the 'only vocal' modality very occasionally, by very few children (unfortunately, there are no data regarding frequencies). Thus the Mastering Memory program involved primarily visual and visual/auditory memories, whereas, the digit span tests involved exclusively auditory memory.
Another important issue in interpreting the results obtained for the 'Digits Backwards' test is the results obtained for other BAS-II tests, especially, the 'Digits Forward' test for which no significant effects were found. The main difference between those two tests is the further processing of information in the 'Digits Backwards' test. As suggested earlier such processing requires certain skills, therefore, it may be mre susceptible to practice effects. As such the effect could have been due to the practice of memorisation processes experienced by the children from the experimental groups. However, it is very unclear how this could have happened considering that: the other BAS tests did not involve working memory to the same extent (i.e. did not demand information to be held in temporary store while other processing was carried out); and that the Mastering Memory program had been used essentially in its visual modality while the digits tests require auditory memory.
The argument for a suggested interpretation (that it was the memory training which contributed to the reported effects) can be based on the dual-coding theory, which suggests that pictorial stimuli generate simultaneous verbal and non-verbal memory codes (Paivo, 1986). In which case the Mastering Memory program would involve the auditory memory regardless of its visual modality. Although, admittedly the role of the articulatory process in working memory is far from being clarified (Massarro, 1995).
The above argument can be also supported by the fact that both Mastering Memory program and the digits tests involve sequential memory. Singleton, Thomas and Leedale (1996) proposed that STM deficits in dyslexic children do not involve general STM but memory which involves verbal coding and/or rapid sequential processing. If those types of STM are the ones that seem to underlie the cognitive processes, perhaps they are also the ones which can be enhanced more easily (which would explain why there were very few signifcant effects found on other BAS-II tests). If it is accepted that this finding was a result of the memory training, the differences in modality (of the training tool and test) may indicate the transferability of the learned skills, which would be a very important factor in the light of evidence suggesting that transfer of skills is very difficult. Also the fact that there were significant results on the test which involves working memory more than the other tests has a bearing on those theories regarding working memory which claim that STM and working memory are distinct (e.g. Siegal & Ryan, 1989), and/or those which view working memory as a temporary storage of information while other cognitive tasks are being performed (e.g. Baddeley & Hitch, 1974).
The other test for which significant differences were found was the 'Recall of Objects Spatial Immediate' test, where the 'instructions' group obtained significantly higher scores (mean=80.6) than the 'no instructions' group (mean=75.35), which in turn obtained higher scores than the control group (mean 57.26). The 'Recall of Objects Spatial Immediate' test requires the types of memory which are very closely related to those that the Mastering Memory program involves i.e. visual and sequential (or spatial). This perhaps allows us to suggest that the differences found were the effects of memory training. Such a clam can be further strengthened by evidence brought be Kail & Hagen (1977) that despite iconic memory being invariant developmentally there are age-related differences in the way children and adults maintain and transfer visual information. This evidence allows memory training to facilitate the acceleration of the developmental processes.
Another test where there were significant effects foud between the 'instructions' and the control group was the 'Recall of Objects Verbal Immediate' test. (The first group had a mean of 80.6, the latter 57.76). In terms of the modality of the test it involved presentation and then verbal recall. (The Mastering Memory program involved visual presentation and visual recall.) However, here again, one can claim on the basis of the dual-coding theory that both the test and the Mastering Memory program engaged both types of memory simultaneously. Which perhaps makes it possible to attribute the differences found to the effects of the training. The fact that there was no significant main effect on this test for the 'no instructions' group may indicate the importance of the role of a teacher/therapist in providing memorisation strategies.
Due to the fact that significant effects regarding age groups and both sexes were reported for just 2 different tests it is difficult to suggest that any of those two variables played any significant role in the experiment.
On the whole, the following facts which must be considered in the interpretation of the results:
1) some significant effects were found on 3 out of 7 BAS-II tests. (Recall of Objects Verbal Immediate, Recall of Objects Spatial Immediate, Recall of Digits Backwards; no significant results were found on: Recall of Digits Forward, Recall of Designs, Recall of Objects Verbal Delayed, Recall of Objects Spatial Delayed). A possible explanation for this is that all 7 tests involved different types of memory, and the fact that there were significant results on some but not the others is due to the match of the type of memory involved in the training.
2) the training involved 4 sessions per child only (with a 4 week break between two of the sessions). This, of course, considering the effects of practice theory (Baddeley, 1994) is very likely indeed, to have influenced the outcomes of the invesigation. Although, on the other hand, the Mastering Memory program manual makes a claim that in some cases just 5 sessions will be sufficient.
regarding the statistical analysis:
3) despite the attempts to make up equal groups regarding their scores of the BAS-II before the training, and despite the fact that there was not a significant difference found between them, there were quite big differences in the means of the (percentile) scores between the groups before the training.
4) despite attempts to make up groups with equal numbers of subjects in each one this was not the case.
More generally, the hypothesis that there will be significant effects of the memory training between the groups were confirmed only for 3, and rejected for 4 BAS-II tests. In those 3 tests, however, it was also confirmed that it was the group which had used the Mastering Memory program together with the instructions from the teacher / experimenter that exhibited greater enhancement of memorisation process than the group which used the Mastering Memory program on their own. Although, the 'no instruction' group did exhibit some significant improvement too. This suggests that even though using the Mastering Memory program with an active role of the teacher is most effective, the Mastering Memory program used on its own may have some memory enhancing effects. This is contrary to the claims in the Mastering Memory manual, which say that without the teacher / therapist no effects can be gained. One of the intriguing questions which arise from this finding is whether the 'no instructions' group would benefit equally to the 'instructions' group if the computer program rather than the teacher / therapist / experimenter suggested using different memory strategies. This seems to be a likely hypothesis considering that the role of the teacher / therapist advocated by the Mastering Memory manual is that of prompting and suggesting, and giving the child a chance of discovering 'how' on his / her own.
In the context of evidence from studies on the effectiveness of mnemonic techniques, presented earlier, (which is inconclusive), in regards to certain memory tass, the tact that the improvement in memorisation processes was not found on all of the administered tests is not surprising. But the fact that there were significant effects found on some of the tests is greatly encouraging. The evidence on mnemonic techniques suggests that they are fairly difficult to use and are even less likely to be adopted by children with 'below average' cognitive abilities (Kail & Hagen, 1977). The present study did not include any subjects with particular cognitive deficits and the groups wer matched according to their abilities. Yet, it would be extremely useful if further investigations were able to show that chldren with 'below average' abilities could benefit from the computer-assisted or even fully computerised memory trainng to the same (or perhaps even greater) extent as 'normal subjects'. This would be especially impotant considering that memory underlies many other cognitive functions, which are crucial in children's educational careers. Moreover, the finding that even the group which used the computer on its own without amy mnemonic instructions had exhibited some memory enhancement, suggests that fully computerised training might be feasible. This would make it possible to include memory enhancement on the curriculum in schools, without creating a burden on teaching resources. And also, as a consequence, would give the child more opportunities to practice.
The present study indicated also the importance of determining the type of memory being focused on. The research on dyslexia (Singleton et. al., 1996) clearly indicates that deficits of specific kinds of memory bear distinct consequences. Thus, it would be advisable to aim at the enhancement of a particular aspect of memory rather than 'general' memory skill, which as the results reported in this project indicate might be a difficult task.
Although this experiment has its imperfections, it has validated the attempts to devise a form of training which can facilitate short-term memory and take advantage of the capability of the computer. There is a need for further research which would clarify the long term effects of such a training, and identify the role of particular components of the training.
(given in the text above rather than the entire original piece of research)
Baddeley, A. D. (1994) Human Memory. Theory and Practice. Hove. LEA
Baddeley, A. D. (1982) Reading and working memory. Bulletin of the British Psychological Society. 35, 414-417
Beech, J. R. & Singleton, C. H. (Eds.) (to be published 1997) Psychological Assessment of Reading. Routledge
Elliott, Smith & McCulloch (1996) missing reference
Howell, D. C. (1989) Fundamental Statistics for the Behavioural Sciences. Belmont, California: Duxbury Press
Kail, R. V. & Hagen, J. W. (Eds.) (1977) Perspective on the development of memory and cognition. Hillsdale, N. J.: Erlabaum
Massarro, D. W. (1995) Verbal working memory; a view with a room. American Journal of Psychology. Vol. 148, No 1, pp. 123-155
Mitchell, J. E. (1996) Mastering Memory Manual. www.calsc.co.uk
Paivo, A. (1986) Mental Representations: A Dual-Coding approach. New York: Oxford University Press.
Siegal, L. D. & Ryan, E. B. (1989) The development of working memory in normally achieving and subtypes of learning disabled child. Child Psychology. 60, 973-980
Singleton, C. H.., Thomas, K. V. & Leedale, R. (1996) CoPS 1. Hull: Lucid Research Ltd