Specification: The multi-store model of memory: sensory register, short-term memory and long-term memory. Features of each store: coding, capacity and duration.

Research investigating the MSM

The MSM of memory has been investigated extensively, and research has provided support for the model's different components. For example, research by Miller (1956) supports the idea that our STM has a capacity of 7+/-2 'chunks' of information; Baddeley (1966) supports the notion of different types of encoding in STM and LTM; Peterson & Peterson (1959) support the idea of limited duration in STM and Bahrick (1975)supports the idea of unlimited duration in LTM.

Miller (1956) 'The magical number seven, plus or minus two' Capacity of STM

Aim: To investigate the capacity of STM.

Method: Literature review of published investigations into perception and STM, from the 1930s to 1950s.

Results: This existing research suggested that organising stimulus input into a series of chunks enabled STM to cope with about seven 'chunks', and this was why more than seven digits, words or even musical notes could be remembered successfully. When we try to remember a phone number, which has 11 digits, we chunk the information into groups, for example: 0767...819...45...34, so we only need to remember four chunks of information and not 11 individual digits.

Conclusion: Organisation (or 'encoding') can extend the capacity of STM and enable more information to be stored there, albeit briefly.

Evaluating Miller (1956)

Miller’s (1956) theory is supported by psychological research. For example, Jacobs (1887) conducted an experiment using a digit span test, to examine the capacity of STM for numbers and letters. Jacobs used a sample of 443 female students (aged from 8–19) from the North London Collegiate School. Participants had to repeat back a string of numbers or letters in the same order and the number of digits/letters was gradually increased, until the participants could no longer recall the sequence. Jacobs found that the students had an average span of 7.3 letters and 9.3 words, which supports Miller’s notion of 7+/‐2.

Although Miller’s (1956) theory is support by psychological research, he did not specify how large each ‘chunk’ of information could be and therefore we are unable to conclude the exact capacity of STM. Consequently, further research is required to determine the size of information ‘chunks’ to understand the exact capacity of STM.

Finally, Miller’s (1956) research into STM did not take into account other factors that affect capacity. For example, age could also affect STM and Jacobs’ (1887) research acknowledged that STM gradually improved with age.

Peterson & Peterson (1959) Duration of STM

Aim: To investigate how different short intervals containing an interference task affect the recall of items presented verbally, and to infer the duration of STM.

Method: The participants were 24 male and female university students. The verbal items tested for recall were 48 three‐consonant nonsense syllables (such as JBW or PDX) spelled out letter by letter. These have since been named ‘trigrams’. There were also cards containing three‐digit numbers (such as 360 or 294). The researcher spelled the syllable out and then immediately said a three‐digit number. The participant had to count down backwards in either 3s or 4s (as instructed) from that number. This was to prevent repetition of the trigram by the participant. At the end of a preset interval of between 3 and 18 seconds a red light went on and the participant had to recall the trigram.

Results: Peterson & Peterson found that the longer the interval the less accurate the recall. At 3 seconds, around 80% of the trigrams were correctly recalled, whereas at 18 seconds only 10% were correctly recalled.

Conclusion: STM has a limited duration of approximately 18 seconds. Furthermore, if we are unable to rehearse information it will not be passed to LTM, providing further support for the MSM and the idea of discrete components.  

 Evaluating Peterson & Peterson (1959)

Peterson & Peterson used a sample of 24 psychology students, which is an issue for two reasons. Firstly, the psychology students may have encountered the MSM of memory previously and therefore may have demonstrated demand characteristics by changing their behaviour to assist the experimenter. Secondly, the memory of psychology students may be different from that of other people, especially if they had previously studied strategies for memory improvement. As a result we are unable to generalise the results of this study to non‐psychology students.

Furthermore, it could be argued that Peterson & Peterson’s study has low levels of ecological validity. In this study participants were asked to recall three‐letter trigrams, which is unlike anything people would want to memorise in their everyday lives. As a result we are unable to apply these results to everyday examples of memory and are unable to conclude if the duration of STM may be longer for more important information, such as a vital phone number. 

However, Peterson & Peterson’s study was highly controlled and took place in a laboratory of Indiana University. As a result Peterson & Peterson had a high degree of control for extraneous variables, which makes their procedure easy to replicate to test reliability.

Bahrick (1975) Duration of LTM

Aim: To investigate the duration of LTM.

Method: 392 American university graduates were shown photographs from their high school yearbook and for each photograph participants were given a group of names and asked to select the name that matched the photographs.

Results: 90% of the participants were able to correctly match the names and faces 14 years after graduating and 60% of the participants were able to correctly match the names and faces 47 years after graduation.

Conclusion: Bahrick concluded that people could remember certain types of information, such as names and faces, for almost a lifetime. These results support the MSM and the idea that our LTM has a lifetime duration (at least 47 years), and is semantically encoded.

Evaluating Bahrick (1975)

Bahrick’s research used a sample of 392 American university graduates and therefore lacks population validity. Psychologists are unable to generalise the results of Bahrick’s research to other populations, for example students from the UK or Europe. As a result, we are unable to conclude whether other populations would demonstrate the same ability to recall names and faces after 47 years.

Furthermore, Bahrick found that the accuracy of LTM was 90% after 14 years and 60% after 47 years. His research is unable to explain whether LTM becomes less accurate over time because of a limited duration, or whether LTM simply gets worse with age. This is important because psychologists are unable to determine whether our LTM has an unlimited duration (like the MSM suggests), which is affected by other factors such as getting old, or whether our LTM has a limited duration.

Finally, it could be argued that Bahrick’s study has high levels of ecological validity as the study used real‐life memories. In this study participants recalled real‐life information by matching pictures of classmates with their names. Therefore, these results reflect our memory for real‐life events and can be applied to everyday human memory.

Overall Evaluation of the MSM

Support for the MSM comes from the case study of Clive Wearing, who contracted a virus that caused severe amnesia (memory loss). Following the virus, Wearing could only remember information for 2030 seconds; however, he was able to recall information from his past, for example his wife’s name. Wearing was unable to transfer information from his STM to his LTM, but was able to retrieve information successfully. Wearing’s case supports the idea that memories are formed by passing information from one store to the next in a linear fashion and that damage to any part of the MSM can cause memory impairment.

Further support for the MSM comes from psychological studies. For example, Miller (1959) supports the idea of a limited capacity of 7+/‐2 chunks of information in STM; Peterson & Peterson (1959) support the idea of a limited duration in STM, of approximately 20 seconds and Bahrick (1975) supports the idea of a lifetime duration in LTM. All of these studies support the different elements of the MSM and therefore suggest that the model is an accurate representation of human memory.

Baddeley and Hitch (1974) developed the working memory model (WMM) as an explanation of the complexity of STM and a way of explaining some of the research findings that could not be accounted for by the MSM, for example parallel processing (multi‐tasking).

Finally, evidence from brain scans has shown that different areas of the brain are active when performing STM tasks (hippocampus and subiculum) and LTM tasks (motor cortex). The hippocampus is also involved in transferring short‐term memories into long‐term memories. This suggests that different brain regions are responsible for the different components of the MSM, supporting the idea that our memory is made up of discrete stores.