Evaluation of Working memory model

Strengths

The working memory model is high in face validity, this means that the model seems plausible. In this case, it seems plausible because it fits with everyday experience of manipulating information when solving problems, with short term memory as a dynamic process rather than a static store. For example, Baddeley (1997) suggests that mentally counting the number of windows in your house (or flat) demonstrates the operations of working memory. Normally a person will imagine each room in turn, forming a mental image of each window (Visuospatial sketchpad), they will count using the phonological loop to rehearse the numbers and this will all be coordinated by the Central Executive.

Another advantage of the Working Memory model is that verbal rehearsal is not necessary for all types of information – just verbal strings - this fits better with our everyday experience. For example, we do not have to rehearse everything that happens to us because events are processed by the episodic buffer.

Empirical Evidence

Baddeley (1975) found that participants' memory span for visually presented one-syllable words was greater than for polysyllabic words (words with more than one syllable). This suggests that the articulatory loop is only able to hold limited number of syllable chunks. However, later studies showed that memory span for words like “Cricket” and “Bishop” was greater than for words like “Harpoon” and “Friday”. Harpoon and Friday have long vowel sounds and take longer to say, which suggests that memory span is limited by time (about 2 seconds) rather than number of chunks. This supports the idea that the articulatory control system is a time based rehearsal loop.

Research where verbal rehearsal is prevented by using articulatory suppression tasks has also supported the model. Baddeley and Lewis (1981), for example, visually presented sentences to participants while they performed an articulatory suppression task (e.g., repeating 'hi-ya' over and over again) and asked them to spot mistakes in the sentences. They found that sentences with syntactic anomalies (Words in the wrong order; e.g., The trees flew up into the birds) were detected less often than sentences with semantic anomalies(meaning does not make sense; e.g., 'The anxious woman was cycling by the man's laughter'). This seems to indicate that the articulatory loop is important in holding information regarding the order of words. It also shows that working memory is a complex dynamic system, rather than a static STM store.

Weaknesses of Working Memory Model

The most important weakness of the working memory model is that the functions of the central executive are vague and difficult to test. For example, Baddeley claims that the central executive has a limited capacity; however, how this capacity can be measured independently of the other components is unclear. Baddeley also claims that the CE can be divided into subsystems; however they have not yet been identified and it is difficult in practice to determine which processes that control the slave systems are part of the central executive and which are part of other systems, such as language

Visual spatial sketchpad example

Shepard and Feng (1972)

Aim - To discover the VSS

Procedure - Shepard and Feng used shapes such as the opening of a cube and asked P#s to imagine folding these flat shapes to form a cube with the grey area as the base and then to decide whether in the finished cube, the arrows on the open shape would meet.

Results - Found that visual images work in a very similar was to real life perception.

Conclusion - The VSS is used in tasks like the above for the temporary storage and manipulation of visual patterns and spatial movement. They require concentration. People are likely to find it difficult to do two tasks simultaneously

Baddeley, Thomson and Buchanan (1975)

Aim - To find out the existence of the phonological loop (Inner voice)

Procedure - P's were given word list for brief exposures and asked to write them down in serial order.

• Condition A - 5 words from a pool of one syllable, English words such as palm, twice. 
• Condition B - 5 words from a list of polysyllables such as organisation, association. 

Results - Average correct recall over several trials showed that the shorter words had greater recall. This was called the 'word length effect' 

Conclusion - Lab experiment in repeated measures was well controlled but could be criticized that longer words are simply less familiar than shorter words and therefore harder to recall.

- Consent. 

Working memory model - The components.

Baddely and Hitch believed that the STM store in the Multistore Model was too simplistic: they thought that short term memory was not a passive store, but several active processes that manipulate information.

Central Executive

The central executive is considered the most important part of working memory, because it controls attention and coordinates the actions of the other components, it can briefly store information, but has a limited capacity. The central executive is modality free, which means that it can store information in any sense. This core component has two slave systems which work independently. 

Phonological Loop

The phonological loop consists of two parts, the articulatory control system and thephonological store:

• Articulatory Control System (The Inner Voice) The articulatory control system rehearses information verbally and has a time based capacity of about 2 seconds. It is helpful to think of it as the system that you use to mentally rehearse information by repeating it over and over again.
• Phonological Store (The Inner Ear) The phonological store uses a sound based code to store information, but this information decays after about 2 seconds, unless it is rehearsed by the articulatory control system. The phonological store receives its input either directly from the ears or from long term memory. If you imagine your favourite piece of music you are using your phonological store.

Visuospatial Sketchpad

The visuospatial sketchpad stores and manipulates visual information, input is from the eyes or long term memory. if you imagine an object and then picture it rotating you are using your visuospatial sketchpad.

Strengths and weaknesses of the multi-store model

- There is a considerable body of evidence supporting the distinction between the different types of memory ( ie sensory, short term and long term). Much of the evidence is based of lab studies but there is also important case study research on people whose memories have been affected by disease or accident.
- The multistore model was important and influential when it was first proposed, however, subsequent research has revealed a number of weaknesses. The model is too simplistic (reductionist) and inflexible to explain fully the complexities of human memory. It is also very deterministic.

Encoding of LTM- Baddeley (1966)

Aim - Investigate the nature of LTM coding.

Procedure - Baddeley constructed a pool of short, familiar words for each of four catagories, and agreed with Conrad. Baddeley modified his experiment to test LTM. He extended the length from 5 words to prevent rehearsal by interrupting them after each presentation. Each list was presented four times and then recalled after a 20 minute interval.

Results - Under these conditions, he found that acoustic similarity had no affect on recall but that words with similar meaning were poorly recalled

Conclusion - LTM codes mainly semantically

Evaluation - Lab experiment so it is therefore artificial
- No serious ethical issues
- Consent is needed.

Encoding of STM - Conrad (1964)

The 3 main types of encoding 

1. Acoustic coding - the sound of a stimulus 
2. Visual coding - the physical appearance of a stimulus 
3. Semantic coding - the meaning of a stimulus

Conrad 1964

Aim - Investigate the nature of coding in STM

Procedure - Conrad presented p's with several series of 6 letters visually at a rate of 0.75 per letter. The letters used were selected from B,C, F, M, N, S, P, T, V and X. P's were told that they had to write down each list in the order they were given after the presentation. Conrad analysed errors rather than correct answers. 

Results - P's tended to make errors by substituting letters that sounded the same rather than looked the same ( B for V) 

Conclusion - Letters are encoded by sound even when presented visually. 

Evaluation - There are no serious ethical issues

- Consent is needed

- Artificial stimuli

- Non-representative 

Factors that affect the duration of LTM.

Experimental techniques - People seem to be able to remember things from a distant past much better if they are given certain cues instead of being asked to recall from scratch.

Depth of learning - People are likely to remember things for longer if they learned it well in the first place. Bahrick and Hall (1991) tested long-term memory for algebra and geomatry. People who took it to high education courses showed higher recall rates than those who stopped in secondary school.

Pattern of learning - Bahrick (1987) looked at people who had learned Spanish and found that vocabulary items learned over spaced sessions were retained for longer than vocabulary learned in intensive sessions.

Nature of material to be learned - some types of material are retained for longer than others. Conway et al (1991) tested open university psychology students and found that certain subject topics were recalled more accurately over time.

Factors that affect the duration of STM

Rehearsal - We can extend the duration of STM by rehearsing information. For example, if we look up a phone number in a directory, we tent to repeat it over in our heads in order to hold it in our memory for long enough to dial it correctly.

Intention to recall - It seems to make a difference whether we are making a conscious effort to recall material of not. Sebrechts et al (1989) proved this using familiar 3 English nouns. Correct recall fell by 1 per cent after 4 seconds.

Amount of information to be recalled - Murdock (1961) adopted the Peterson and Peterson technique but used either a single, three letter word such as cat or a set of three unrelated words such as pen, hat and lid. When he used three words as the stimulus, he found the same pattern of decline  as the original study, but when he used three letters, he found recall was remarkably resistant to decay.

Duration of STM - Peterson and Peterson (1959)

Aim - To investigate the duration of STM.

Procedure - Asked P's to recall strings of consonants (e.g FBK) selected. Recall delay was set to 3, 6, 9, 12 seconds during which rehearsal was prevented by participants counting backwards in threes from a target number. Each subject was tested a total of 8 times at each of the 6 delay intervals.

Results - While after a 3 second retention interval trigram about 80% of trigams were recalled, after 18 seconds only 10% were recalled.

Conclusion - information disappeared or decayed rapidly from STM if not rehearsed.

Evaluation - Consent must be gained.
- Trigrams are artificial and do not reflect on how we remember things in daily life.
- Trigrams presented in earlier trials caused confusion and therefore were recalled incorrectly.
- Possible that loss of information is down to capacity limitations rather than duration.

Factors that affect the capacity of STM

Reading out loud - Digit span increases if p's read the digits aloud instead of reading them subvocally. Baddeley (1999) suggested this because  the digits are also then stored briefly in the echoic store.

Individual difference - There is some evidence that individual differences affect STM capacity. For example, people who are highly anxious appear to have shorter spans (MacLeod and Donnellan, 1993)

Pronunciation time - A number of studies have shown that the capacity of STM is determined by time constraints rather than structural limitations. Naveh-Benjamin and Ayres (1986) compared memory spans for speakers of English with speakers of other languages. They found a direct relationship between span and pronunciation time. For example, Arabic speakers have a shorter digit span because they have longer words (takes longer to pronounce words compared to English speakers) 

Influence of LTM - Cowan (2000) believes that Miller might have overestimated the number of chunks that can be held in STM. He thinks that performance on span tasks is often affected by rehearsal and long term memory does not reflect the capacity of 'pure' STM and estimated that the capacity of STM is actually four chunks when such factors are controlled. Bower and Winzenz (1969) found that digit strings that were repeated within a series of immediate memory span trials become easier for p's to recall. This suggests that the strings have been gradually rehearsed and stored in LTM, thus temporarily increasing STM capacity.

Improve your digit-span performance by 'chunking'.

The digit span task exercises your verbal working memory. Scientists refer to working memory as the cognitive system that allows the temporary storage and manipulation of information. According to one influential cognitive theory, this system has specialised components, one of which, the 'phonological loop', underlies verbal working memory abilities (Baddeley & Hitch 1974). The phonological loop is comprised of a verbal storage system and a rehearsal system. If you do this task, you may find yourself mentally rehearsing the string of digits as they appeared on screen; this is the rehearsal system in action. It allows the visual inputs to be recoded so that they can enter your short term verbal store and it also refreshes decaying representations (that is, any item that is about to be forgotten).

Verbal working memory is involved in many everyday tasks, from remembering your friend's telephone number while you enter it into your phone, to understanding long and difficult sentences. Think about it; how could you understand a whole sentence if you couldn't remember the words at the beginning long enough to connect with the words at the end! Verbal working memory is also thought to be one of the elements underlying intelligence (or 'IQ'); thus, the digit span task is a common component of many IQ tests, including the widely used WAIS (Wechsler Adult Intelligence Scales). Performance on the digit span task is also closely linked to language learning abilities; improving your verbal memory capacities may therefore help you to master a new language or to expand your vocabulary.

An average adult is thought to have a digit span of 7 (plus or minus 2) items. One of the best studied methods for improving verbal memory is through the use of 'chunking' strategies (Miller 1956), in which items are recoded into meaningful units or 'chunks'. For example; 2 6 7 8 9 7 (6 items) is easier to retain when remembered as 26 78 97 (3 items). In one study, by training a volunteer to use complex chunking strategies over the course of 20 months, scientists were able to increase digit span from 7 to a massive 79 items! (Ericcson et al, 1980). We recently studied the underlying brain activity involved in this process. When recoding strategies were used to remember digit sequences, increased activation was oberved in lateral prefrontal and posterior parietal cortex. On this basis, we have hypothesised that this prefrontal-parietal network underlies strategic recoding in working memory (Bor et al, 2004, 2006)




source - http://www.cambridgebrainsciences.com/browse/memory/test/digit-span

Digit Span

source -  http://www.cambridgebrainsciences.com/browse/memory/test/digit-span

Capacity

LTM - It is generally accepted amongest researchers that LTM has an unlimited capacity. It is possible to lose things from LTM through processes such as decay and interference but the loss does not occur because of capacity limitations.

STM - One of the earliest systematic attempts to measure the capacity of STM was carried out by Jacobs (1887) called the digit span technique which is where P's repeat back strings of digits in order of presentation. Jacobs found that on average people could recall about 7 digits in this immediate serial recall task and other research has supported this. George Miller wrote the famous article named 'The magical number seven, plus or minus two', in which he proposed that we can hold about seven items in our STM, but that there is a range of capacity between 5 and 9 items. Miller believed that our immediate memory span is determined by the number of 'chunks' of information we can hold rather than the number of individual letters or numerals. Many psychologists have criticized this idea of 'chunks' as it is too vague. Simon (1974) said that the span of the 'chunk' depends on the amount of information that is being processed.

Glanzer and Kunitz (1966)

Aim - Existence for the difference between STM and LTM

Procedure - Two groups of P's were presented with the same list of words. Group A were tested on immediate recall (nonsense syllables)

- Group B recalled after 30 seconds in which they completed a distactor task.

Results - There was some evidence of the primacy effect (P's could recall items from the start of the task more accurately because they'd been transferred to the LTM) but the recency effect disappeared. 

Conclusion - Millers suggestion of a STM capacity of 7 is affected by the primacy effect, and the true capacity of STM is only 3 or 4 'chunks'

Evaluation - It isolated the affect of STM recall whilst minimising the effect of LTM influences but, lacks mundane realism.

-  word lists do not comprise a reflection of the memory tasks used in everyday life. 

- Lacked ecological validity 

Sperling (1960)

Aim - To find out the existence of sensory memory

Procedure - Using a chart, participants were asked to look at the chart for 50 milliseconds and the asked to recall how many of the letters they could remember.
(2nd stage) Sperling then got the p's to recall single rows of letters when particular tones were heard (high tone, for top row, medium tone for middle row etc)

Results - First instance, P's could remember 4/5 letters although they were aware of more. Second instance, P's on average recalled 3 items from the row indicated

Conclusion - Although in theory P's should remember more than the average 4 items, it is thought that the image of each item fades during the 50ms and the time it takes to report back recalled items.

Evaluation - Lab exp. therefore high levels of control, lack of ecological validity as it is artificial, but easily replicated.
- Consent is needed