Multi-store model of memory

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

  Model:

  
• Sensory registor - capacity

  Very large storage of information from each senses

  Has a mechanism for selecting relevant information and discarding the rest. Known as attention and determines which information is transferred to the STM
• Sensory registor - duration

  Extremely brief - less than 1/2 a second
• Sensory registor - coding

  Codes accordingly to the sensory source

  e.g info from the eyes is coded visually (iconic memory) and sound based info is coded acoustically (echoic memory)
• Sensory registor

  Information enters the system through our senses.

  Everything we see, hear, touch, taste and smell enters our sensory registor so takes in the information from all our senses and there is a store for each sense.
• Sensory registor study - Sperling (1960)

  Briefly displayed visual arrays containing three rows of 4 letters to participants.

  Found that participants could recall 4 or 5 letters from the 0.05 seconds of arrays but reported being aware for the presence of more letters.
• Sperling further experiment on sensory registor

  Trained participants to recognise 3 tones, high tone indicated the top row of letters, medium tones indicated the medium row and low tone indicated bottom row.

  Sterling presented participants with a series of displays for 0.05 seconds each again and the sounded tone immediately afterwards.

  Participants recalled on average 75% of letters of the cued rows.

  Suggests sensory registor has a large capacity but decays rapidly so duration is short
• Short term memory (STM) - capacity

  Between 5 and 9 chunks of information

  7+/- 2
• STM - coding

  Mainly acoustic
• STM - duration

  Very short - about 18 seconds
• STM capacity study - Jacobs (1887)

  Measures digit span. He read out lists of digits starting with 4 digits and asking participants to read them back to him.

  If correctly recalled he read out 5 digits, repeating the procedure and increasing the number each time until participant could not correctly recall digits.

  This determined the participants digit span. Jacobs found the mean digit span for numbers was 9 and for letters is 7. Supporting the idea that capacity is 7 items
• Evaluation for Jacobs STM capacity study

  • Very old study, may not have been carefully controlled, could have been extraneous variables like distractions
  • Since 1887, study has been replicated many times and the finding confirmed. Meaning they are reliable, Jacobs study is a valid test
• STM duration study - Peterson and Peterson (1959)

  Participants briefly shown a trigram (3 consonants) and asked to recall it after a period of 3,6,9,12,15 or 18 seconds. Given an inference task of counting backwards in 3's between initial presentation of trigram and recall, to prevent rehearsal.

  Findings were that when time between presentation and recall increased, successful recall decreased.

  After 6 seconds, there was 40% recall and after 18 seconds, there was a 10% recall.

  Suggests if rehearsal is prevented, information in STM can only be retained for up to 18 seconds.
• STM coding study - Baddely (1966)

  Participants shows various lists of words and asked to recall them in the correct order. When tested immediately after presentation, it was found more mistakes came from mixing up words that were acoustically similar (cat, can, cab) rather than words that were semantically similar (big, huge, great)

  Concluded that participants made more errors with acoustically similar words because they said to themselves that they are likely to get confused with similar sounding ones. Supporting the idea that main form of coding is acoustic.
• Evaluation of research into short term memory

  • Carried out in high controlled lab conditions, good control over extraneous variables. Standardised procedures mean that each research study could be easily replicated by other researchers.
  • However, it was a highly artificial environment which could be different to everyday environments. Therefore lacks ecological/external validity.
• Long term memory (LTM) - capacity

  Potentially unlimited capacity

  Very difficult to research capacity of LTM
• LTM - duration

  Potentially lasts almost a whole lifetime
• LTM - coding

  Mainly semantic
• LTM duration study - Bahrick (1975)

  Cross sectional study using 392 ex-high school students aged 17-74. Had graduated from their high school anywhere from 2 weeks to 57 years ago.

  2 procedures: free recall and photo recognition

  Free recall - participants has to free recall names of their class mates

  Results 15 years after leaving = 60% accuracy. Results 48 years after leaving = 30% accuracy

  Photo recognition - participants had to sort 50 photos into those who were or were not in their class

  Results 15 years after leaving


  = 90%. Results 48 years after leaving=70%
• LTM duration study Bahrick conclusion

  Suggests memories can be held in long term memory without distortion for a very long period of time. However many years there may be some degeneration as a result of old age, shows how duration of LTM is many years.
• Evaluation of Bahricks LTM duration study

  • Only one type of LTM was being investigated, findings can not be generalised to other types of information, study doesn't tell us anything about duration of semantic memories.
  • Attempted to control how much contact the participants had with the yearbook but couldn't properly control amount of contact - extraneous variable.
  • Compared to other memory research, it has high external validity as it tests real life memory.
• LTM coding study - Baddely (1966)

  Participants were presented with one of 4 sets of words and asked to recall them.

  1. Acoustically similar (man, mad, cap, can, cat, map)
  2. Acoustically dissimilar (pit, few, cow, pen)
  3. Semantically similar (great, large, big, huge)
  4. Semantically dissimilar (good, huge, hot, safe, thin)

  When participants were asked to recall after 20 minutes, they were most likely to make errors in semantically similar words.

  Shows that words with similar meanings caused confusion, suggests info is coded semantically.
• Evaluation of Baddely's study

  • Lacks ecological validity as it doesn't resemble the way that memory is used in everyday life.
  • Good control over independent variable (the word lists), provides good evidence that coding in long term memory is semantic
• Evaluation of MSMM

  • Shows STM and LTM are different. For example in the STM we tend to mix up acoustically similar words whereas in LTM we tend to mix up semantically similar words.
  • Lack ecological validity - most studies that support MSM have very little meaning, artificial.
• Evidence from brain damaged patients showing MSM is an oversimplified model:

  • Clive Wearing - unable to lay down new long-term memories but can hold a brief conversation and has some short term memory. Supporting MSM, suggesting that LTM and STM are separate stores.
  • KF - motorcycle accident. His STM was reduced to only 2 digits but could still form long term memories. Supports distinction between STM and LTM. But does not explain how he can form LTM without a normal STM. As MSM assumes that information must flow through the STM to the LTM
• Multi store model suggests that information comes from one of the five senses and enters a sensory register

  A small fraction of this information is paid attention to and enters the short term memory (STM) If no attention is paid to it, the information quickly decays
• Three things can happen to the information in the STM:

  • Rehearsal helps keep the information in STM (maintenance rehearsal)
  • The information is rehearsed and goes into the LTM (elaborate rehearsal)
  • If information is not rehearsed it is forgotten
• The memory stores are unitary (each has just one part)