An experiment to see if clustering of words improves word recall

We use cookies to give you the best experience possible. By continuing we’ll assume you’re on board with our cookie policy

This study tried to replicate the work of Cofer, Bruce and Reicher (1966), however the results of our data analysis for no significant difference between the groups, this may be explained choice of sample.


Memory, one of the most important processes the brain conducts, it tells us everything about ourselves, what we know, how we do things. Much work has been done into the study of memory, its process and how to improve ones own memory. This side of psychology is of keen interest to cognitive psychologist. In the context of psychology memory is the capacity to retain and recall information. This in itself is further divided two main types of memory, long term memory (LTM) or short term memory (STM)

Theories of memory processing such as the multi-store model, which is based on the assumption that there are three ‘stores,’ which make up human memory, these stores are the sensory register, short term memory and long term memory. The difference between them is how much memory is stored by each. The sensory register is a short acting memory store for the sensory record of the stimulus. There is no meaning to the stimulus at this point. From this store the stimulus information passes to the next store, the short term memory store. Peterson and Johnson (1971) showed that information in the STM is stored in an acoustic format and also that the duration that information can be stored in the STM is limited. Peterson and Peterson (1959) estimated this duration to be 6-12 seconds before the memory starts to decay and also that STM capacity is limited.

For the memory in the STM store to become a Permanente memory, the information must pass to the long term memory store, which is apparently limitless.

This model of multi-store memory was challenged by Craik & Lockhart (1972), who replaced it with the concept of levels of processing. According to this approach incoming information is analysed by a limited capacity processor, which can handle a variety of information. Once analysed the resulting memory trace which is passed on to the LTM. Research indicated that there are three levels of LTM, being episodic memory, semantic memory and procedural memory. Episodic memory retains information about events that have happened to us in our live. Semantic memory deals with rules, language and other symbolic material. Procedural memory is as it seems, it deals with how we do certain things and the correct order to do it in, and for example the procedure one must follow to correctly start a car. In this case it’s your procedural memory that tells you that you put the clutch down in the car before you put the car into gear.

The work of Craik & Lockhart (1972) on the levels of processing approach found that by increasing the levels of processing recall of the information would be greatly improved. That the deeper the analysis of the information the stronger the memory becomes. Further research into the effect of levels of processing by Cofer, Bruce and Reicher (1966) found that semantic encoding provided better recall in a free recall task then episodic encoding. This was found when they compared recall scores of their participants, finding that words in category based word blocks, as apposed to random word blocks, were recalled better.

The aim of our experiment was to see if semantic encoding of words would lead to better recall of the words compared with episodic encoding, therefore replicating the findings of Cofer, Bruce and Reicher (1966) to see if they still hold merit.

The null hypothesis I shall employ for this experiment is that I predict no difference in recall between participants who were given category word blocks or randomly mixed word blocks.

The experimental hypothesis I shall employ for this experiment is that I predict participant who are given category instances in a blocked manner will do better in recalling the words then participants who are given randomly mixed words.



The design employed in this experiment is an independent measures design, using an unrelated sample. The major disadvantage of employing this method is that differences that occur between groups my not be due to the manipulation of the independent variable but due to independent difference between the groups. However the advantages of the design are that there are no order effects, which is important in memory experiments.

The Independent variable that was manipulated the level of processing that occurred, whether it was semantic encoding (category instances) or episodic encoding (randomly mixed word). The dependent variable for this experiment is the number of correct words recalled. Spelling and order of the words recalled was not taken into account.


There were 19 participants that took part in this study, 10 in one group and 9 in the other. All participants are aged between 18-30 years old. The method of sampling employed was an opportunity sample, where the participants are all first year psychology students studying at City University.


The materials used for this study were a set of standardised instructions and a word list of 36 words. This was presented in 6 blocks of 6 words, either randomly assorted or assorted in category instances. See appendix for example. They were also provided with paper and pencil to recall the words after.


Firstly standardised instructions were read out to the participants explaining what they were to do, Gave the participant a sheet with the word blocks on it. 10 participants were given the category instance blocks and the remaining participants were given the randomly assorted blocks. They were then asked to memorise the list of word that they had been given, they had a period of one minute and thirty seconds to do this. Participants were then asked to wait 30 seconds and then write down as many words as they could recall onto the record sheet.


The mean average of condition A (category instances) is 22.2 words recalled accurately and the mean average for group B (randomly assorted words) is 17.3, from looking at the mean on could infer that semantic encoding has lead to better recall but I shall employ the Mann-Whiney U test to see if the difference in mean results is significant. I have chosen this test because I have used an independent measures design. This gave U a value of 25 at P? 0.025. The critical value for U is 20 at P? 0.025 as it is a one tailed experiment. As the observed value of U (25) is greater then the critical Value of U (20), the results of the study are not significant and therefore I must reject my experimental hypothesis and accept my null hypothesis.


From the analysis of the raw data I was no able to accept my experimental hypothesis. Though the mean averages for group A was higher than that of group B in employing the Mann-Whiney U test the results were not significant. One of the reasons this could have occurs is because there was no interference task employed. Another reason is that because the participant are all psychology students they are aware of memory improving techniques and some participants in group B may have used these during the experiment, thus confounding the results.

To improve the study I would have like to have used a interference task and as further research I would like to compare the results of psychology students against non academic participants.

Tagged In :

Get help with your homework

Haven't found the Essay You Want? Get your custom essay sample For Only $13.90/page

Sarah from CollectifbdpHi there, would you like to get such a paper? How about receiving a customized one?

Check it out