Murdock (1962) explored the “primacy and regency effect.” This is when participants recalled words remembering earlier and later items but tended to forget the words in the middle of the list. Glanzer and Cunitz (1966) backed this up but also suggested that our memory is split into different departments.
Paivio (1969) suggested the most powerful influence on recall of words is how easily they can invoke a mental image. He proposed a dual – coding theory to account for the effects of images on verbal learning. So does the brain have a filing cabinet like system that attempts to sort out into hierarchies or categories of knowledge? Collins and Quillian (1969,1972) devised the hierarchical network model of semantic memory. This is concerned with our memory for words and their meanings.
This report is a scaled mimic by the work of Bower et al (1969). He presented two groups of participants with identical sets of words to learn. One group had organised words by conceptual hierarchy, the other group’s words were random. Bower demonstrated that there would be higher recall when words were arranged in a conceptual hierarchical form. He claimed that recall was improved because of word association and the recognition of them being on the response sheet. With the random word list, participants could not establish a similar rule to help them generate words.
Method
Participants
The total number of participants was twenty-four, twelve female and twelve male. The participants ranged in age from 16 to 75 years - the mean average being 51. Employment status included employed, unemployed and mothers.
Preparation/Materials
Before the trial was conducted a hierarchical stimulus sheet was prepared. This consisted of 3 countries with the top parent “PLACES”. These countries were then categorised with 2 states (or counties) per country. Each state or county was again sub-categorised with 2 cities (or towns) appearing in that state (or county). All in all a total of 22 words was created (see appendix 1). A second stimulus list was then created. Taking the 22 words generated for the hierarchical list a random list was created by duplicating the same words but presenting them in a uniform but random fashion (see appendix 2). A briefing statement was prepared for the participants so, while they should not be aware of the aim of the experiment, they would be clear as to there parcipitation. Materials used to conduct the experiment were a coin (see procedure); a pen, 2 sheets of blank A4 paper per participant (these would become known as the response sheets) and a stopwatch to monitor and control word exposure and recall times.
Design
An opportunity sample was selected. To randomise the participants as much as possible, 12 experimenters each selected 2 participants- one male, one female. On top of this an independent measures design was applied. This is where the experimenter ensures that no participant is exposed to more than one part of the experiment. This design was applied to reduce fatigue and order effects and also to minimise the chance of a participant double guessing the aim of the experiment and consequently adapting their response. The independent variable is what changed across the two conditions. The controlled conditions, were the time allocated for each participant, equal allocation of female and male participants to each list of words, whether the participant is given the random list of words or the hierarchical list to memorise and finally that the experiment took place in a quiet room. The independent variable is the factor that is manipulated and controlled. In this case it is whether the list of words is random or hierarchical. The dependent variable is what is affected by the independent variable. It is also measurable. This is how successfully people remember. It is the difference, if any between the recall rate of randomised words and the hierarchical words. It is measured by recording how many words were recalled.
Procedure
Conducting The Experiment
12 experimenters each selected 2 piers as participants – one male, one female. The first participant was asked to toss a coin to randomly select which stimulus list they would be exposed to. The second participant would be the alternate. Once the stimulus list was established for each participant the methodology would be the same. In the controlled environment of a quiet room the participant was briefed as to the form of the trial. The following instruction brief was read to the participants and then the trial conducted as prescribed in the brief.
“ This is a two trial experiment. During each trial you will be given the same sheet of words, which you should try and learn. You will be given 60 seconds to study the sheet, it will then be removed and you will be given a sheet of paper on which to write down as many words as you can in any order. You will then have 2 minutes in which to recall the words. You will then be presented with the sheet of words again and have another 60 seconds to study them. The words will be taken away and you will have a further 2 minutes in which to write down the words.”
The time for exposure to the stimulus sheet and recall there-of was monitored and controlled using a stopwatch. When all 12 experimenters had conducted their trials the results were pooled for analysis.
De-brief
The participants were thanked and told their score. It was carefully explained that their results were confidential and would be kept anonymous.
Table Showing Relationship Between The Randomised - List Performance and The Hierarchical – List Performance
RESULTS
The probability level of the results was analysed using the Mann Whitney test. The interval-scale data was collected and formulated into two tables of raw data (see appendix 7, 8). These comprised a table of scores relating to recall of words from the randomised list and recall of words form the hierarchical list. The scores were converted to ordinal- scale data and analysed using a two-tailed Mann-Whitney test (see appendix 9). The lowest value being rank 1 and the highest rank N. Ranks were shared for those who scored the same. U is the observed (i.e. the calculated) value of the Mann Whitney statistic. The observed value of U at 0.05 level of probability was 54; the critical value was 37 (U= *, N1 = *, N2 = *p> or <0.05). It was therefore necessary to reject the research hypothesis. The results indicated a greater than 5% probability that the null hypothesis was correct and for this reason the null hypothesis was retained. Any difference in recall was purely due to chance. The experiment failed to provide significant evidence that word recall was improved through a hierarchically structured presentation.
Discussion
The results demonstrated no significant difference in the number of words recalled across the two conditions. The two-tailed hypothesis was therefore rejected and a null hypothesis was retained. We were hoping for results that supported Bower’s study because his work was adopted as our model. There could have been a variety of reasons for why the experiment did not concur. Bower and Winenz suggested that if participants were asked to read out loud it would improve their recall ability. If we had applied this method then better results may have been achieved. The results may have failed to realise our expectations due to inherent errors in our approach. Perhaps choosing participants from our pier groups was a mistake. For a true insight into someone’s memory the participant has to be receptive at the time of testing. The relaxed atmosphere that exists around friends and family may have had a negative effect on the participant’s concentration levels. The trials were supposed to be conducted as formally as possible but it would be hard to totally discount the fact that the participants may have been simply going through the motions. Would the results have been different if the same participants were used but in environment of say a job interview? Fatigue or a rushed approach of the participants may also have a played a factor- was the trials conducted straight after work? Late at night? Or while the participants mind was elsewhere? Some of the participants may have a very high IQ others may be low. If we had more time the experiment could have been improved using a matched pairs design. This would have improved the subject variables and the consistency across the two conditions, therefore would produce more reliable results. Validity could also be subjected to the experimenter’s variables. I had no control over where and how the other experimenter’s applied their tests. It could have been a noisy room a television or radio may have been on in the background, or the participant may have talked through the experiment. To reduce the effects of extraneous and participant variables good sampling and good control measures are needed.
In addition, the results supported Atkinson and Shiffrons theory of the two distinct and separate memory stores and the existence of the rehearsal loop. The participants were given 60 seconds to learn the words, it is reasonable to assume that the acoustic rehearsal occurred and some of the words were transferred into the LTM. It would have been interesting to see what words all the participants had remembered, if the results showed that the items at the beginning and the end of the list were recalled more frequently it would have supported Glanzer and Cunitz. The items in the middle of the list would have been the poorest to recall. This is further evidence of the existence of separate memory stores.
Suggestions for further research on this subject may include an experiment based on Paivio’s original work. According to Paivio and Csapo (1973) free verbal recall is generally higher for items presented as pictures rather than just words. He required participants to encode words and pictures verbally. The results of these experiments yielded a higher recall of pictures rather than words. It would be interesting to see if visual aids could provide meaningful context and therefore aid learning. I would extend my research and conduct the experiment again but see if there is any difference between randomised words and pictures with words. This would support the theory that image and verbal codes are independent. The superiority of pictures in free recall can be explained by the duel-coding theory.
If the results were insignificant it is necessary to look at other theories of memory of memory recall. Collins and Loftus (1975) revised the idea of semantic hierarchies by proposing a network model that was interconnected in semantic clusters (see appendix 10). They assumed storage systems in computers might be similar to ones in humans. The models concepts are represented in nodes. Nodes are connected with other nodes via links. It is referred to as the semantic network. The strength of the links between the nodes depends on the semantic relatedness between them. For example BMW and car have a strong link, whereas BMW and banana have a weak link. The notion of spreading activation refers to the process by which the activation of a given node also activates others to which it is linked. This could explain why the results were not significant in our research. Some participants may not be knowledgeable about geography and therefore the links to word association will be weak. If the list had been words that most people could relate to for example food the test may have gained more positive results. This has come under some criticism due to the problems defining attributes. There are too many concepts difficult to define. Consequently another idea was developed which is known as the schema theory. This is based on the idea that knowledge is stored in the memory organised in a set of schemas. Each incorporates all the knowledge of a given type of object or event that we have acquired from past experience. They are linked together into related sets organized hierarchically. Schank and Abelson (1977) proposed people have scripts that are used in our comprehension of everyday events. Scripts represent knowledge about complex event sequences. They can be broken up into hierarchical organised scripts and subscripts. This again could show why our results are weak because if the participants have not acquired knowledge from a past experience in reference to countries there will be poor memory recall.
Memory is key to how we function and interact with our environment and those around us. Research into memory development is an important field. We are constantly bombarded with information and if we can learn to retain this information and recall efficiently then it would enhance us all. One day technological advances may allow us to see into the brain. A system may be developed to decode the millions of tiny electrical impulses that flash around our heads and reveal more to us than we were previously aware. However as technology advances the need to retain information is being taken away from us. Everything we need to know is just the click of a button away. Computers are everywhere and accessing information is becoming easier all the time. Some people don’t even know their friend’s telephone numbers any more. It’s just scroll to name and dial. It is important that we don’t let this over reliance on technology to take over one of the primary functions of our minds therefore a balance needs to be kept to maintain a healthy and active mind.
