A study investigating the effects of categorisation on recall

A study investigating the effects of categorisation on recall

1. Introduction

Background research
The cognitive approach to psychology studies the processes the mind uses to deal with information and looks at areas such as language, learning, perception and memory. Cognitive psychologists commonly use models to explain information flow. These models are abstract ways of representing how the mind deals with information rather than defining separate areas of the brain for each aspect of memory. The information processing model uses the analogy of a computer system – information is received and processed in various ways by the mind before being passed into memory. Within the study of memory, there are three main processes:

Encoding > Storage > Retrieval

Encoding is the process of perceiving and understanding input. Storage is the way in which we commit information to memory. Retrieval is the process used to access information that is not currently in conscious memory. William James, an early psychologist, identified two types of memory - “primary memory” and “secondary memory”, which are now called “short term memory” and “long term memory”.

Atkinson and Shiffrin’s “multi-store model of memory”
This theory states that there are three distinct memory stores - sensory, short term and long term. The amount of attention paid and “rehearsal” of information affects likelihood of this information passing first into short term and then into long-term memory.

Sensory memory has a very limited duration of around a second for visual store and 2 seconds for the acoustic store. Only information attended to is passed into short term memory – we would be swamped by sensory information otherwise.

Short term memory has a limited storage capacity and a very short duration. Short term memory can be lost by decay or displacement as new information is added to the store. Miller (1956) suggests the “magic number” 7 plus or minus 2 - that is between 5 and 9 bits of information can be retained in short-term memory. By organising this information in short term memory, between 5 and 9 “chunks” of information can be passed into long term memory. These chunks can be acronyms, words, phrases or anything else that links the information together into a meaningful structure. By chunking information, a much greater quantity can be stored. If information is suitably rehearsed it is more likely to pass into long-term memory.

Long term memory has an effectively infinite capacity and memories can last a lifetime. Memory can decay over time, or there can be problems of retrieval, where the memory is there but cannot be recalled. Information in long term memory is more likely to be in the form of semantics, organised by general meaning rather than in greater detail. In a study by Tulving and Pearlstone (1966) participants recall of word lists of 12, 24 and 48 words in categories of 1, 2 or 4 words was tested. The answer sheets were used with and without category headings to measure the effects of categorisation on recall. Participants showed a significant increase in words recalled when the category headings were present on the answer sheet. This study indicates that organising information in memory increases the amount of recall.

Rationale
Previous studies such as Tulving and Pearlstone (1966) imply that organisation of data into categories affects the amount of recall of this information. This experiment was an attempt to approximately replicate this study.

Aims
The aim of the study was to ascertain whether categorisation of words would increase the number of words recalled from prepared word lists.

Hypotheses
The null hypothesis
Any variation in the number of words recalled between the categorised and uncategorised word lists is due to chance factors.

The alternative, one tail hypothesis
There will be a significant difference in the number of words recalled between the categorised and uncategorised word lists such that the number of words recalled from the categorised word list will be higher.

2. Method

Method and Design
The study used the experimental method to try and establish a cause and effect relationship. The variables being tested can be manipulated when this method is used and extraneous variables reduced. Reliability is increased, as the study is replicable.

This is a preview of the whole essay

A repeated measures design was used, meaning fewer participants were required and participant variables were reduced. To reduce order effects, counterbalancing was used in the experiment, with half the participants hearing the uncategorised list followed by the categorised list and the remaining half hearing the categorised list first.

Variables
The Independent variable was the method of presentation and this was operationalised by presenting two lists of words – the two conditions of the independent variable were whether the words were categorised or uncategorised. The dependent variable was the number of words recalled. This was operationalised by the amount of words written down by the participants.

Participants
The target population of the participants was family and friends of the experimenter. An opportunity sample was used, as the target population was insufficiently defined to use other sampling methods such as random or stratified sampling. A self-selecting sample would have needed more time than was available for this study. The sample contained an equal number of male and female participants between the ages of 18 and 40. No participants were under the influence of alcohol or illicit drugs. English was the first language of all participants. IQ of participants was not measured and therefore not controlled.

Materials
Two lists of words were used in the experiment, words organised into categories (see appendix A) and uncategorised words (see appendix B). Participants were provided with written instructions (see appendix C). Recalled words were recorded on blank A4 paper - a sample answer paper is included in appendix D. Familiar, recognisable English words were used with no difficult or unusual words included. The 20 words in each list were comprised of 4 groups containing 5 words each.

Procedure
The participant was seated in a comfortable, quiet room and given the sheet of printed instructions to read. When the participant was clear on the procedure the experiment was commenced. The first word list was read out loud by the experimenter at approximately one every 2 seconds. At the end of the list, the participant then wrote down as many words as they could remember. When they had finished the second list was delivered orally as before. The participant again wrote down as many words as they could remember. Upon completion of the experiment the participant was fully debriefed as to the purpose of the experiment and any questions answered.

Controls
The experiment was conducted in a variety of environments but variables were as consistent as possible. The time of day was limited to the evening and temperature kept to within normal limits. Noise was kept to a minimum during the experiment with televisions or radios turned off. When other people were present this was limited to the participants partner or children. Experimenter effects were reduced by minimising interaction before and during the experiment. A number of ethical issues were addressed during the study. The participants were not deceived as to the purpose of the experiment and were all fully debriefed immediately afterwards. As stated in the instructions participants were free to stop the experiment and leave at any time if they felt uncomfortable. All data collected was anonymous and confidential – no names or other identifying features were recorded on the answer sheets or anywhere else. No racist, sexist, obscene or otherwise inflammatory or offensive words were used in the two word lists, in an attempt to protect the participants.

3. Results

Table 1 – Summary table of results

Summary table commentary
A significant difference was found between the total number of words recalled in the categorised and uncategorised conditions. The words recalled from the categorised list exceed those recalled from the uncategorised list by a ratio of 3:2.

There is also a corresponding difference in the mean scores between the two conditions, although the mean is subject to influence by unusually high or low scores and can give a distorted picture of central tendency.

The median is less likely to be affected by extremely high or low scores and presents a more accurate overview of the central tendency of participant’s scores. However the two totals still have a ratio of 3:2.

The mode is not a particularly useful measure in this study. In the categorised condition the most common value of recall is 12, which is significantly higher than the 9 in the uncategorised conditions.

The range is greater for the categorised condition than the uncategorised condition. This is to be expected given the totals.

Additional graphical description of results

Fig 1. Mean number of words recalled

Fig 2. Median number of words recalled

Fig 3. Mode number of words recalled

Descriptive statistics commentary
Similar points to those made above can be made regarding figures 1/2/3. However the bar charts do show in a more obvious, dramatic fashion the extent of the differences in the scores between the two conditions. In all cases it may be seen that there is a significant increase in words recalled from the categorised word list compared to the uncategorised word list.

Relationship of results to hypothesis
From the commentaries above it can be seen that there is a 50% difference between words recalled from the categorised and uncategorised lists. It is unlikely that a difference of such a magnitude could have arisen by chance. Therefore the null hypothesis that any variation in the number of words recalled between the categorised and uncategorised word lists is due to chance factors is rejected. Consequently the alternative hypothesis, that there will be a significant difference in the number of words recalled between the categorised and uncategorised word lists such that the number of words recalled from the categorised word list will be higher, is supported.

4. Discussion

Validity & Suggestions for improved validity
This study has face validity, as the number of words written down is likely to be the amount of words remembered. The repeated measures design of this experiment means that each participant was tested in the same environment at the same time for each of the two conditions. The independent variable appears to be influencing the dependent variable as nearly all participants scored higher on the categorised list than the uncategorised list.

The independent variable was operationalised by the word lists being presented in a categorised or non-categorised format. The organisation of words into groups is a form of categorisation. The dependent variable - the amount of words recalled - was operationalised by the participants writing down the words remembered on the answer sheet. This was an effective way of measuring words remembered. An alternative may have been to have the participant vocalise the words recalled but this is unlikely to have had any effect on the results.

Support for the previous research into the effects of categorisation on recall by Tulving & Pearlstone provides this study with concurrent validity.

Ecological validity may be lacking in this study due to the presentation of information using word lists instead of in a more realistic form such as a short story or passage of text. However a list of words is more reliable and more accurately scored than a passage of text or a story. When using a passage of text as stimulus material, operationalising the independent and dependent variables would become difficult. Demand characteristics may have been influential during this study as most participants guessed the purpose of being tested twice. Ecological validity could be improved by testing participants in a more natural situation, but then extraneous variables would not be as controlled and reliability may suffer.

Reliability & Improving reliability
This study used the experimental method – this was the most appropriate as cause and effect can be inferred. Other methods such as case study or correlation study would have been inappropriate in this case.

A repeated measures design was used with counterbalancing to reduce order effects such as boredom or familiarity. Using an independent measures design may have reduced order effects and demand characteristics as the participants would not have been tested twice, however this would introduce more participant variables. By using a matched pairs design, the benefits of the independent measures design could be obtained without the same degree of participant variables. Matched pairs would have been the ideal design for this study but required time and resources beyond those available.

The participants were all friends and family of the experimenter and are not representative of the general population. The target population is a narrow, biased and indefinable group, which ruled out the possibility of using methods of sampling such as random, stratified or systematic. A more clearly defined target population such as students attending a certain college or employees of a certain company would allow less biased sampling types to be used and would probably provide a wider selection of participants who would be more representative of the general population. A larger sample would also be preferable as only 20 participants were tested.

In this study, categorisation of the words was into four categories. The results may have been different if fewer or more categories were used – Tulving and Pearlstone found differences in the number of words recalled between categories containing 1, 2 and 4 words.The test only used verbal material as a stimulus, so can only be used as a test how many words can be remembered when presented verbally. Results may vary if the words were presented visually. Different words were used on the two word lists, which may have affected the amount of words recalled, as some words may be easier to recall than others. Using the same words on the two lists would allow more accurate measurement of recall, but then order effects may have increased recall from the word list presented second due to extra rehearsal.

The method of presentation in this study was verbal. It may have been preferable to use mechanical means such as a Dictaphone or audio CD to present the word lists to avoid any variations in the delivery and to ensure consistent timing between each word within the test and also between subsequent participants. This would dilute experimenter effects during the delivery of the word lists and make the study more accurately replicable. If the study were to be repeated, testing all the participants in the same room would be preferable to reduce any differences in environmental variables such as distractions, noise and temperature.

Implications of study
While the methodology varies slightly from Tulving and Pearlstone’s study, the principle is the same and the results show the same significant increase in words recalled from the categorised word list. By organising the words into groups, the information was possibly, as Miller suggested, “chunked” allowing a greater quantity of information to be passed into long term memory. The median number of words recalled from the uncategorised list was 8.5, within Miller’s proposed number of “slots” in short term memory, 7 plus or minus 2.

The categorisation of the information gives the participant a cue to aid recall. Tulving and Pearlstone provided the category headings in one of the conditions of their study, which further increased recall. Category headings were not provided on the answer sheet in this study, but participants commented on the groups of words when recalling and it appears that the participants were using the category headings as cues to help recall.

Semantic errors were made on the answer sheets such as putting “hifi” instead of “stereo”. The participant knew the general form or function of the item they were trying to recall, but used the wrong name to label it. These errors support the Atkinson and Shiffrin model of memory and the theory that long term memory stores information in a semantic format.

Generalisation of findings
The findings of this study can only be generalised to the target population – in this case the friends and family of the experimenter. This is culture bound, biased and very narrow, so generalisation to people beyond those who took part is likely to be inaccurate as the target population is not representative of the wider population.

Application of study to everyday life
Using mnemonics to help remember otherwise unconnected information is a practical and effective use of Miller’s “chunking” in day to day life. Creating a mnemonic such as “every good boy deserves fudge” to remember the musical notes EGBDF turns a collection of letters into a distinctive, easy to remember sentence.

Organising information in a “mind map” or “spider diagram” is another effective way of increasing the quantity of information recalled. These diagrams involve the “chunking” of ideas and information into groups and hierarchies. This has applications in academic and work related situations where large amounts of data need to be recalled.

5. References

Atkinson, R. C. & Shiffrin, R. M. 1968. Human memory: A proposed system and its control processes. Cited in K.W.Spence and J. T. Spence (Eds.), The Psychology of Learning and Motivation: Advances in Research and Theory (Vol. 2, pp. 89-195). New York: Academic Press.

James, W, 1890, The Principles of Psychology, New York, Holt

Miller G. A., 1956, The magical number seven plus or minus two: some limits on our capacity for processing information, Psychological Review, Vol. 63, pp. 81--97.

Tulving, E. and Pearlstone, Z., 1966, Availability vs Accessibility of information in memory for words, Journal of verbal learning and verbal behaviour, 5, 381-391.

Other sources

British Psychological Society Code of conduct ()

Hill, G., 2001, AS level psychology through diagrams, Oxford, Oxford University Press

Jarvis, M., Russel, J., Flanagan, C., Dolan, L., 2000, Angles on Psychology, Cheltenham, Nelson Thornes

Appendix A - Categorised word list
Banana
Apple
Orange
Pear
Strawberry
TV
Video
Stereo
Fridge
Cooker
Parrot
Spider
Monkey
Lizard
Turtle
Jumper
Shorts
Jacket
Gloves
Shirt

Appendix B - Uncategorised word list
Bear
Bee
Chips
Coat
Computer
Giraffe
Iron
Kangaroo
Pasta
Radio
Rice
Scarf
Shaver
Skirt
Soup
Steak
Telephone
Trousers
Whale
Vest

Appendix C - Written instructions received by participants

This is not an intelligence test and all data collected is anonymous and confidential.

Procedure;

You will hear a list of words.
Please write down as many of the words as you can remember. You are not being timed, and may stop writing when you are ready
When ready to continue, you will hear another word list.
As before, please write down as many words as you can. Again, you are not being timed, and may stop writing when you are ready.

The procedure will be explained further on completion of the two word lists. You are free to leave at any time throughout the procedure. Do you have any questions?

Appendix D – Sample answer paper

Appendix E – Table 2 – Number of words recalled