In a repeated measure design 26 students of psychology were asked to participate in an experiment in order to examine how chunking can be related to a specific amount of recall letters and to a specific amount of recalled numbers. Results of the correlation test indicated the existence of a vary strong positive correlation. In this paper, we review some of the basic mechanisms of chunking in human cognition and we suggest some possible directions for further research.
A various number of research studies have demonstrated the importance of memory in humans. According to psychologists memory is the process by which people encode, store and retrieve information. Although short-term memory is the memory store in which information first becomes meaningful there is not enough evidence for the specific process by which sensory memories are transformed into short-term memories. Some theorists support that the information is first translated into images or graphical representations and others believe that the transfer occurs when the sensory stimuli are changed to words (Baddeley and Wilson, 1985). What is clear, however is that short-term memory has incomplete representational capabilities. In fact, the specific amount of information that can be held in short-term memory has been identified as seven items or chunks of information with variations up to plus or minus two chunks (Miller, 1956). Consequently chunk is considered to be as a meaningful grouping of stimuli that can be stored as a unit in short-term memory. However, there is an issue as to what constitutes an item. According to George Miller an item can be individual letters or numbers or it may consist of larger categories much more complicated such as words or sentences. A large body of evidence indicates the existence of mechanisms by which chunks are created, stored and retrieved. The way in which chunking has been associated with human cognition is described by a specific family of computational programs called EPAM and CHREST.
EPAM (Elementary Perceiver and Memorizer: Feigenbaum and Simon, 1962) is a computer model which posits that learning is gained by the increase of a discrimination network allowing access to long-term memory. The general organization of EPAM consisted of a limited short-term memory (between 3 and 7 items in short-term memory), a discrimination network and attention mechanisms. The external information provided by this network is organized into the appropriate chunk (in EPAM chunks= nodes in the discrimination network) through patterns of perceptual tests. For each learning process is necessary a specific amount of time (e.g. 8 seconds in order to create a new chunk). EPAM has been used in several domains in order to explain the different aspects of verbal learning such as the tip of the tongue phenomenon and the learning of spelling by children. It has also been used to explore the role and impact of perceptual chunking such as the way in which letters are grouped into words and it has triggered off the development of the chunking theory based on the acquisition of chess expertise first proposed by Chase and Simon (1973).
Based on one's past experience Chase and Simon carried out a comparison between expert and inexperienced chess players in order to find out differences in their memory task (the De Groot recall task) and perceptual task. They defined chunk as a sequence of pieces forming perceptual and semantic units in a less interval of two seconds. According to the chunking theory groups of pieces that have a various number of connections are more probably to be recalled. Chase and Simon based on chess relations claimed that expert chess players have gained more and larger chunks than inexperienced chess players. These findings were carried out in a computer program called MAPP (Memory Aided Pattern Perceiver; Simon and Gilmartin, 1973) which led to the estimate that expertise in chess would require about 50.000 chunks in memory.
A further extension a EPAM theory is considered to be CHREST (Chunk Hierarchy and Retrieval Structure) which involves more complex patterns of self organization and adaptation. Similar to EPAM, CHREST posits that learning derives from the growth of a discrimination network even though there is a small difference in the settlement of the network such as the presence of action chunks in addition to perceptual chunks. Meanwhile other two new strategies have been added to CHREST (Gobet, 1996; Gobet et al. 2001). The first strategy adds a new link called lateral link between the existent nodes. This lateral link may either demonstrate that there is a kind of similarity between the two nodes or that there is a cause and effect condition between these two nodes. The second mechanism involves the production of retrieval structures called schemata. Schemata which are also known as templates are created automatically. The main components of templates are: (a) a core which encodes fixed information and (b) slots which encode diverse information. The main role of templates is to demonstrate how higher level structures can be constructed from chunks and to produce more rapid LTM encoding mechanisms. CHREST has provided a various cluster of applications of perceptual chunking such as language acquisition, verbal learning, problem solving and expertise behavior.
However chunking theory has addressed adverse criticisms by more recent studies. Cowan's research supports that attention has a finite capacity,the retrieval of information is limited by this and it amounts to only3-5 chunks. These conditions are strictly linked with the fact that LTM and rehearsal are not capable to associate stimulus items into chunks of an unknown size(Cowan,2001)
Holding(1992) based on expertise behavior identified five limitations of chunking theory and suggested SEEK (Search, Evaluation, Knowledge) theory. According to Holding: (1) storage into LTM requires less time than chunking theory proposes (2) because of their size chunks are not able to reflect conceptual knowledge (3) memory for random positions hae underlined that individual skill abilities have an impact on the storage of chunks into LTM (4) the estimated number of 50.000 chunks necessary to reach expertise in chess and in any complicated task is not reasonable and (5) verbal knowledge and processing are omitted by the chunking theory (Gobet and Simon, 1998a).
All research studies have concluded that chunking is an important process in learning because helps the learner to construct more specific patterns of information by grouping them into more comprehensible units. The general goal of the study was to examine how chunking can be related to a specific amount of recalled letters and to a specific amount of recalled numbers. Therefore, we test the hypothesis that there is not nay relationship in chunking between recalled letters and recalled numbers.
In a repeated measure design 26 students of psychology, 12 males and 14 females were asked to participate in an experiment in order to examine how chunking can be related to a specific amount of recalled letters and to a specific amount of recalled numbers.
Materials and procedures
A computer display type Toshiba Satellite L500-13T was used in order to present 10 slides of a series of letters and 10 slides of a series of numbers to the participants. Subjects were supplied with pens and paper for writing down their answers. Finally a class of the British Hellenic College was occupied for the attainment of the study. 10 slides of series of letters were presented successively to the participants. Each participant's task was to recall as many letters they could and once the slide has gone off they had to write them down in an interval of 10 seconds. The duration of each slide was 5 seconds. After the end of the first experiment a second one has taken place with the same procedure but with a small difference. 10 slides of series of numbers instead of letters were presented to the subjects.
Data were collected and a correlation test was conducted in order to test the relationship in chunking between recalled letters and recalled numbers. The results have shown that the total score in recalled letters was mean=76 and s=8.416 in comparison with the total score in recalled numbers whereas the mean=73,92 and s=16,543. Positive strong correlation was found in chunking between total number of letters recalled (r=1,p<0,01) and total amount of numbers recalled (r=0.720, p<0.01)
The main purpose of this study was to examine the degree in which chunking is related to the retrieval of a specific amount of letters and numbers. Our first hypothesis was statistically significant:that is, chunking is closely related with the specific amount of recalled letters and numbers. The present study`s result show a strong positive correlation,thus means that in chunking the retrieval of a total score in letters is associated with the retrieval of the total score in numbers. These results and their interpretation are compatible with the results and interpretations of some previous investigations which support that when related pieces of information are provided in chunks(complex chunking) students are more capable to build on schemas and to make the necessary associations in order to facilitate the process of retrieval.
We now address some limits of our study and we suggest possible directions for future work. One limitation is that we have not used a projector in order to make easier the show of the slides for the participants. Another weakness that we recognize is that the subjects were not asked if they had encouraged some problem with their eyes a cause of the subdued light that existed in the experimental room. Finally we have not considered factors such as motivations and emotions which are the central traits of attention.
As we have mentioned these limitations suggest some possible directions for further work. Several studies have discussed the effects of chunking mechanisms for education. Strategies should be found in order to develop chunks that will help students to monitor their limited resources such as small STM capacity and slow learning rates. Consequently a further empirical validation is necessary in order to improve the educational system of our country.