Working Memory

Working-Memory
Postscript
Portable Document Format

Working memory , the more contemporary term for short-term memory, is conceptualised as an active system for temporarily storing and manipulating information needed in the execution of complex cognitive tasks (Baddeley 1986) (e.g., learning, reasoning, and comprehension). Experimental evidence has shown that the working memory is of limited size (Miller 1956), and hence, due to the high conceptual demands, complexity of laboratory experiments and potential information overload associated with problems solving, in both chemistry and physics, there are clear instructional implications.

Definition of working memory

Miller 1956

Miller, G., The magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information. Psychological Review, 63(2), 81-97, 1956.

In this paper Miller presents findings which indicate that short term memory has a capacity of 7±2 ``chunks'', where a ``chunk'' is an arbitrary unit of information.

Baddeley 1994

Baddeley, A., The Magic Number Seven: Still Magic After all These Years?. Psychology Review, 101(2), 353-356, 1994.

This paper provides a review of Miller's original paper (Miller 1956) on working memory, and presents, with references, related works stemming from his ideas.

Baddeley 1986

Baddeley, A., Working memory. Oxford: Clarendon Press, 1986.

This book presents Baddeley's theory of working memory.

Baddeley 1992

Baddeley, A., Is Working Memory Working? The Fifteenth Bartlett Lecture. Quarterly Journal of Educational Psychology, 44A(1), 1-31, 1992.

Presents the philosophy underlying the working memory model and then illustrates it by giving a brief review of the model and some of the findings that have resulted from it.

Ashcroft 1994

Ashcroft, M. H., Human Memory and Cognition. 2nd Edition. New York: Harper Collins, 1994.

Ashcroft describes the working memory as ``the mental workplace for retrieval and use of already known information''. He points out that the short term memory implies a static short-lived store which is limited in the amount of work that it can perform. The more information to be held, the less processing can occur and vice versa.

Kellett 1978

Kellett, N. C., Studies on the Perception of Organic Chemical Structures. Ph.D. Thesis University of Glasgow, 1978.

Proposed the I.C.C.U.D hypothesis which relates Information Content, Conceptual Understanding and Difficulty. Here pupils who lack conceptual understanding may perform reasonably well when information load is low, but as information load increases their performance decreases, causing complaints of difficulty. This can most easily be appreciated with the ``Concorde'' representation (Johnstone 1980) (Figure 9.1)

Johnstone 1980

Johnstone, A. H., Nyholm Lecture: Chemical Education Research: Facts, Findings, and Consequences. Chemical Society Reviews, 9(3), 363-380, 1980.

Presents (Figure 9.1) a pictorial representation of the relationship between concept understanding, information load, and perceived difficulty.

Figure 9.1: The relationship between concept understanding, information load and perceived difficulty: At the tip: information load is low, perceived difficulty is low and concept understanding is high. Moving from the tip to the tail: Information load increases, perceived difficulty increases and concept understanding decreases. View a Larger Image Here

Johnstone and Kellett 1980

Johnstone, A. H. and Kellett, N. C., Learning Difficulties in School Science - Towards a Working Hypothesis. European Journal of Science Education, 2(2), 175-181, 1980.

In this paper the authors acknowledge the interaction between conceptual knowledge, ``chunking'' and perceived subject difficulty. As a students conceptual understanding increases they are able to create larger ``chunks'' of information and thus reduce the information load. They also present strategies by which the teacher can reduce information overload and thus facilitate conceptual understanding.

Laboratory and Working Memory

Johnstone and Wham 1982

Johnstone, A. H. and Wham A. J. B., The Demands of Practical Work. Education in Chemistry, 19(5), 71-73, 1982.

In this paper the authors suggest that the undergraduate laboratory is a poor learning environment. They believe that this is due to an overload of the student's working memory. The reasons for students response to, and suitable strategies that a lecturer can employ to reduce, working memory overload are summarised in Figure 9.2

Figure 9.2: The Effects of Practical Work on the Working Memory. View a Larger Image Here

Working Memory: Experimental Evidence for Finite Capacity

The following definitions are provided here so as to reduce working memory load during the discussion of subsequent papers.

Z:

An integer corresponding to the number of thought steps necessary to solve a problem (i.e. is a measure of a problems complexity or demand). It should be noted that the number of steps varies in accordance with experience/ability. In general, students require more steps than an experienced practitioner. (see ``chunking'', (Johnstone and Kellett 1980)).

X:

An integer corresponding to an individuals working memory capacity (7±2, (Miller 1956)).

Y:

Represents, schema, tricks, stored knowledge and techniques that may be brought to bear on a problem to reduce its Z value.

Facility Value (FV):

is the fraction of problem solvers who were able to solve a given problem correctly, and is measured on a scale from 0 $\longrightarrow$1.

Digits Backwards Test (DBT):

In this test a series of numbers are presented to a student to which he or she responds with the same numbers but in the reverse order (i.e 5279 $\Longrightarrow$ 9725). This can be carried out verbally or as a written test.

Figure Intersection Test (FIT):

In this test, students look at shapes and then shade in, on another diagram, the common area of overlap of all the figures. As the number of figures increases the task becomes more complex.

The following is an introduction to a series of papers, (Johnstone and El-Banna 1986), (Johnstone 1984) and (Johnstone and El-Banna 1989), on working memory overload. The individual papers are then listed with additional comments.

To investigate the overload of working memory a number of chemistry problems were assessed for complexity (Z) and a plot of facility value (FV) against question complexity (Z) produced, Figure 9.3. As can be seen, the plot has two clusters, one with FV$\gtrsim$0.6 and the other with FV$\lesssim$0.3 with the break occurring between Z=5 and Z=6. A sample size of $\sim$  20000 was used and as such contained pupils of differing ability, i.e. working memory capacities (X) of 7±2, and therefore the plot of Figure 9.3 should reflect this. That is, students of a given working memory capacity (X) would successfully answer questions of demand Z until their capacity was exceeded, at which point their performance would fall dramatically (idealised in Figure 9.4). The authors, used the DBT and FIT tests, to assess pupils working capacity (X). Since these tests were new to the students it was assumed that an individuals problem solving strategies (Y) would not be appropriate for these tests and would therefore provide an accurate measure of working memory capacity (X). Students, of varying ability (X), were then presented with questions of varying demand (Z) and plots of facility value (FV) against demand (Z) produced for students with a given capacity X. The authors present data from both secondary and tertiary education and conclude that the results are in general agreement with their hypothesis, but do admit that working memory capacity is not the only factor effecting a students performance.

Figure 9.3: The Correlation Between Facility Value and Question Demand. View a Larger Image Here

Figure 9.4: Predicted Performance in Students with Different Working Memory Capacities. View a Larger Image Here

Investigations into the overload of working memory

Johnstone 1984

Johnstone, A. H., New Stars for the Teacher to Steer By. Journal of Chemical Education, 61(10), 847-849, 1984.

Following an introduction, similar to that given above, this paper discusses overload of working memory and its relevance to both laboratory work and language. Ways in which overload can be reduced in the laboratory, and through appropriate ordering of curriculum content, are presented.

Johnstone and El-Banna 1986

Johnstone, A. H. and El-Banna, A., Capacities, demands and processes - a predictive model for science education. Education in Chemistry, 23(5), 80-84, 1986.

This paper expands on the introduction given earlier and in the light of their evidence offer the following:

• Learning demand (Z) must be kept below the working memory capacity (X) of the learner.
• Strategies (Y) should be taught/developed so that a student can operate beyond their capacity (X).
• Consequences for teaching and testing include:
  1. the re-examining of concepts for Z demand,
  2. the interlinking of concepts to promote Y strategies,
  3. the avoidance of ``noise'' in the laboratory,
  4. the realisation that high Z questions test both X and Y. The demand (X) may be too great, or strategies (Y) not yet developed, such that the students understanding of the subject is not suitably tested.
  5. the information density of text books and worksheets needs to be scrutinised.

They also note that the two independent psychological non-science tests (FIT and DBT) were predictors of student performance in conventional school and university chemistry examinations

Johnstone and El-Banna 1989

Johnstone, A. H. and El-Banna, A., Understanding Learning Difficulties - A Predictive Research Model. studies in Higher Education, 14(5), 159-168, 1989.

Expands on the introduction given above and presents a discussion of their results.

Johnstone et. al. 1993

Johnstone, A. H., Hogg, W. R. and Ziane, M., A working memory model applied to physics problem solving. International Journal Science Education, 15(6), 663-672, 1993.

In this paper the idea that the same question, presented in different forms, may be testing different skills is explored. In particular, the effects of working memory and field dependency on student performance in answering examination and tutorial questions is investigated. Those students who have difficulty in differentiating signal from noise are known as field dependent (see Topic 10, this place 10) They found that a physics problem can be presented in such a way as to reduce the noise input and, as a consequence, improve problem solving success for all groups, especially the field-dependent group.

Opdenacker et. al. 1990

Opdenacker, C., Fierens, H., Van Brabant, H., Sprut, J. and Slootmaekers P. J., Katholieke Universiteit van Leuven, Belgian and Johnstone, A. H., University of Glasgow, Uk, Academic performance in solving chemistry problems related to student working memory capacity. International Journal of science Education, 12(2), 177-185, 1990.

In this paper the correlation between working memory capacity and problem solving performance, as hypothesised by (Johnstone and El-Banna 1986), was investigated using two hundred and fifty undergraduate medical students. Again the DBT and FIT were used, with varying degrees of success, to assess the working memory capacity of students. In the discussion of their results they state that ``our results does not lead to a straightforward confirmation of (Johnstone and El-Banna 1986). However, we do find a moderate correlation between the size of working memory and problem-solving ability. This correlation was 0.3 (significance 0.01% )''. However they do point out that working memory capacity may only be one factor effecting problem solving ability: ``the correlation found is an estimate of the importance or weight of the working memory factor''.

Tsaparlis et. al. 2000

Tsaparlis, G. and Angelopoulos, V., A model of Problem Solving: Its Operation, Validity, and Usefulness in the Case of Organic-Synthesis Problems.. Science Education, 84, 131-153, 2000.

In this paper the problem solving model which is based upon working memory theory is discussed. The authors report that the model was more useful in the case of students without previous training and for those students who were field-independent (Topic 10).