“The
effects of hierarchical schemes in encoding and free recall of word lists from
memory: effective learning strategies”
Researched
and written by Lesley Mackenzie-Robb
Introduction
The cognitive processes of memory have long fascinated theorists and
researchers, and are perhaps one of the most difficult areas of psychological
research but one which can yield a great deal of important information to aid
in the development of effective learning strategies. A main issue is that research can only reveal
indirect information (Eysenck & Keane, 2000).
One of the most influential theories is Tulving’s Encoding Specificity
Principle, based on the assumption that context matters in recall and
recognition (Eysenck et al, 2000).
Baddeley & Godden (1980) agreed with this notion with respect to
recall memory, but showed that recognition was not context dependent. Of particular relevance to this study is that
by Bower, Clark, Lesgold & Winzenz (1969) which investigated the effects of
hierarchical retrieval schemes in recall memory, comparing the results of these
with those using randomised word lists.
The experimenter-imposed organisation generated an intrinsic context for
the to-be-remembered words, finding that recall was up to three times better
when word lists were in organised schemes.
Broadbent, Cooper and Broadbent (1978) reported similar results when
they used hierarchically organised word lists, words in a matrix and unorganised
word lists in recall tests, as did D’Argembeau, Comblain & Van der Linden
(2005) who found better free recall performance effects with positive
self-referent information, and suggested that such effects are a result of
control of memory retrieval as opposed to initial encoding.
Bower et al (1969)’s study used a repeated measures design with, in all
but one instance, a modality difference between stimuli input (visual) and
output (spoken). This raises some issues
in the light of the influential Working Memory Model (Baddeley & Hitch, as
cited in Eysenck et al, 2000) which differentiated between visual and audio
processing. An important issue
identified by Bower et al (1969) is that of pairwise association, or the
difficulties of developing semantic hierarchies that are widely meaningful and
sensible.
This study addressed these two issues by comparing recall performance
between words in a highly structured, non-ambiguous hierarchical organisation
with the same words in a randomised list, using an interference task to prevent
sub-vocal rehearsal. Modalities of input and output were the same,
and there was no opportunity for between-experiment rehearsal. The use of a non-ambiguous hierarchical scheme
was predicted to have an effect on recall accuracy compared with that achieved
using randomised word lists.
Method
Design
The study used an experimental, independent measures design, with two
groups of participants: a control group (Group A) and an experimental group
(Group B). The independent variable (IV)
of interest was a list of 19 words with two conditions: Group A was presented
with an unstructured, non-categorised randomised word listing, while Group B
was presented with the same words organised into a non-ambiguous hierarchical
scheme. The accuracy of recall
performance (DV) was measured using a ratio scale (correctly recalled word = 1,
incorrect or no-recalled word = 0).
Participants
The participants ( N = 30) were mature Psychology students engaged in
part-time higher education courses through a College Higher Education Centre. An open invitation to take part in the study was
made through the students’ tutor. Participants received no reward for taking
part in the study.
Materials
Each participant was given two A4 sheets of white paper at the start of
the experiment. One of the papers
contained the list of printed words: one set presented words in a hierarchy
containing no linking lines or circles, the other presented the same words in a
randomised single list, both of which can be found in the Appendix.
The second sheet was blank save for an “A” or “B” discretely printed on
the underside of the sheet. The
participants used this sheet to write down their recalled words.
Procedure
The experiment was run on three separate occasions with different sets
of participants. On each occasion,
participants were randomly assigned to Group A or B by blindly selecting a
ticket printed with an A or B. Participants were provided with clear verbal
instructions by the researcher on what they had to do, but were not informed of
the purpose nor focus of the experiment to control for participant bias.
Participants were given two minutes in which to study the words on a
sheet of paper. At the conclusion of
this period, participants were instructed to turn over their word papers, and were
immediately tasked with counting backwards from 50 out loud as an interference
task to prevent sub-vocal rehearsal of the learned words. Participants were then given a two minute
period in which to write down as many of the words that they could recall using
the second blank sheet of paper.
On all three occasions, participants in both Groups took part in the
experiment simultaneously in the same classroom environment. Participants were arranged within the room in
such a way as to avoid any information exchange, and were specifically
instructed not to make contact with each other during the experiment.
Participants were not debriefed at the end of the experiment to avoid
introducing a confound with subsequent groups.
Results
The results were calculated for each group on the basis of one correct
word = 1, a no-word or incorrect word = 0.
The use of the hierarchical word scheme had an effect on free recall
accuracy in Group B (
= 15.2, SD = 3.01) compared with Group A (
= 12.93, SD = 3.11), however the frequency distribution for Group B was negatively
skewed whereas that for Group A was normal.
This has implications in respect of the accuracy of the measures of
central tendency in describing a typical score.
Figures 1 and 2 below illustrate this discrepancy.
Figure 1:
Distribution of Group A accurate recall scores Figures
2: Distribution of Group B accurate
as a function of
their frequency. recall
scores as a function of their frequency.
To determine the significance of the effect, the data was subjected to a
Mann-Whitney U inferential statistical test.
This revealed that there was no significant effect of hierarchical word
schemes on free recall accuracy (Mann-Whitney U (n1 = 15, n2 = 15), = 70.5, p > 0.05). The null hypothesis cannot
therefore be rejected.
Discussion
This study sought to investigate the effects on non-ambiguous
hierarchical word schemes on free recall accuracy compared with recall using a
randomised word list. The between participants
experimental design used identical conditions for each of the two groups, with
participant bias controlled. The
results, although revealing a slight tendency towards a more accurate level of
recall when using the hierarchical word scheme, did not reveal any significant
effects when calculated statistically.
This is at variance with the findings of Bower et al (1969), although
there are differences in the methods and materials used.
What is of interest is the way
in which the participants recalled the words.
Without exception, all participants in Group B recalled their words in
an exact replica of the hierarchical structure of the to-be-remembered list
although no prior guidance was given on the structure of the words, suggesting
that the participants used the hierarchy as a retrieval plan which is
consistent with Bower et al (1969)’s findings.
Why did they not achieve a much higher level of accurate scores than
those working with randomised lists? Potentially,
the answer may lie in the notion that the Group B participants spent more time
memorising and subsequently recalling the structure
of the hierarchy, than the words themselves. If this were to be the case, then the
hierarchical structure could be said to have represented a cue overload. Conversely, those who recalled from the
randomised word list typically listed their words out in a single or dual
column pattern, often with the order of recall mirroring the order of the
to-be-remembered list. This may explain
why more participants in Group A were able to recall between 11 and 13 words
than those in Group B, whereas those in Group B displayed more consistency in
accurate recall with a greater number of words.
Two further observations can be offered to explain the phenomena. First, the participants were all students,
and the experimental environment was their own classroom. It is possible to hypothesise that these
participants are more likely than non-students to be competent and rehearsed in
the use of learning strategies and information recall. Even those students provided with randomised
lists could arguably possess higher than normal learning and recall
capabilities. Secondly, the trend in the
scores from Group B showed an obvious consistency towards higher volumes of
accurately recalled words, suggesting that the use of hierarchies in learning
and retrieval strategies is effective in generating greater efficiencies.
In conclusion, although this study has not revealed a significantly
improved volume of accurately recalled words using the non-ambiguous
hierarchical method, it does support the notion of semantically meaningful
hierarchies in the development of efficient learning and performance
strategies. It would be useful to repeat
the study with a different set of non-ambiguous hierarchical schemes of words
to determine the wider effectiveness of these strategies, and also to repeat
the study with non-student participants.
References
|
Baddeley, A. & Godden, D. (1980).
When does context influence recognition memory. British
Journal of Psychology, 71, p
99 - 104. |
|
Bower, G.H., Clark, M.C., Lesgold, A.M. & Winzenz, D. (1969). Hierarchical Retrieval Schemes in Recall of
Categorized Word Lists. Journal of Verbal Learning and Verbal
Behaviour, 8, p. 323 - 343. |
|
Broadbent, D.E., Cooper, P.J. & Broadbent, M.H. (1978). A comparison of hierarchical retrieval
schemes in recall. Journal of Experimental Psychology: Human
Learning and Memory. 4, (5),
p. 486 - 497. |
|
D’Argembeau, A., Comblain, C. & Van der Lindern, M. (2005). Affective valence and the self-reference
effect: influence of retrieval conditions.
British Journal of Psychology, (96), p 457 - 466 |
|
Eysenck, M. W. & Keane, M.T. (2000). Cognitive Psychology: A Student’s Handbook, 4th edn.
Hove: Psychology Press. |
Appendix
Word
List
Condition
A: Randomised List.
World
Ottowa
Condition B: Hierarchical Schema
|
|
|
|
World |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Ottowa |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|