Multi-Store Model

• Atkinson and Shiffrin's Multi-store model of memory (1968)

  • Represents how memory is stored and transferred between different stores.
  • 3 stores - Sensory register, STM and LTM.
• Sensory register - Contains a sub-store for each of the 5 senses. Receives information from all of our senses and so has a large capacity, but a duration of half a second. Information will pass from the sensory register to STM if we pay attention to it.
• STM is being acoustically encoded (Baddeley) and having a capacity of 7 +/- 2 items, and a duration of 18-30 seconds (Petersen). Maintenance rehearsal occurs when we repeat the new information to ourselves, keeping it in the STM. Prolonged maintenance rehearsal allows it to go to the LTM. No rehearsal causes forgetting.
• LTM is described as being semantically encoded, having unlimited capacity and a very long duration (over 46 years - Bahrick et al). In order to remember information, retrieval must occur which is when information is transferred to back into the STM and continues to pass through a maintenance loop.
• Strengths of the multi-store model of memory:

  • Acknowledges separate stores - Qualitative differences between STM and LTM and shows them as separate stores. For example, STM is acoustic, while LTM is semantic and has a longer duration, Therefore, the model portrays and accurate view of the differences between the two types of memory, as supported by Baddeley and Miller.
• Limitations of the multi-store model of memory:

  • Limited explanation - Tulving et al found different types of LTM (procedural, semantic and episodic). The MSM doesn't represent this because it sees LTM as a single store. This also doesn't represent that some LTM can be retrieved unconsciously (eg - procedural) while others must be retrieved consciously (eg - semantic), which is not reflected in the universal process of information being consciously transferred to the STM during retrieval.
• Limitations of the multi-store model of memory:

  • Alternative explanations - MSM suggests that the amount of maintenance rehearsal determines the likelihood that the information will pass into the LTM, whereas Craik and Watkins (1973) suggest that it is the type of rehearsal which is more important. They suggest that elaborative rehearsal, instead of prolonged rehearsal, is needed to transfer information from the STM into the LTM, by making links with existing knowledge.
• Limitations of the multi-store model of memory:

  • Limited accuracy - Incorrectly represents STM as a single, unitary store. For example, Shallice and Warrington found that their amnesiac patient KF had poor STM recall for auditory stimuli, but increasingly accurate recall for visual stimuli. This, alongside KF being able to differentiate and recall both verbal and non-verbal sounds, suggests that there may be multiple types of STM.
• Strengths of the multi-store model of memory:

  • Supporting evidence - Glanzer and Cunitz (1966) asked participants to free recall a word list in any order. Recall was stronger for words at the start and end of list. This suggests there are separate STM and LTM stores; first words enter LTM (primary effect), most recent words enter STM and are recalled (recency effect), middle words in STM were displaced by later words.
• Strengths of the multi-store model of memory:

  • Supporting evidence - Sperling (1960) investigated capacity by showing a 3x4 grid of letter to trained participants for 1/20th of a second and asked them to recall one row. Recall was over 75% suggesting all rows were contained within the capacity of the iconic store (sensory register is large).
  • Investigated duration - When asked to recall all the letters, participants could only recall 4-5 letters, suggesting letters faded from register before they could be paid attention to. This suggests duration of sensory register is less than 1 second.
• Strengths of the multi-store model of memory:

  • Supporting evidence - Baddeley (1966) - 4 lists of 10 words to different groups.

  A - Acoustically similar

  B - Acoustically dissimilar

  C - Semantically similar

  D - Semantically dissimilar

  • Immediate recall was worst for list A. After 20 minute recall was worst for list D. This suggests coding in STM is acoustic because recalling list A was difficult due to similar sounding words.