Short-term Memory

1
Short-term Memory

• Duration
• Peterson and Peterson (1959)
• Aim To test how long short-term memory lasts
  when rehearsal is prevented.

2
Short-term Memory

• Procedure
• Participants were briefly shown a consonant
  trigram (i.e. 3 letters such as MWG or CGX).
• Participants were asked to count backwards in
  3s from a specified number to stop them thinking
  about the letters.
• After intervals of 3,6,9,12,15 or 18 seconds,
  participants were asked to recall the original
  trigram.
• The procedure was repeated several times using
  different trigrams.

3
Short-term Memory

• Duration
• Peterson and Peterson (1959)
• Findings
• Participants were able to recall 80 of
  trigrams after a 3 second interval.
• Progressively fewer trigrams were recalled as
  the time intervals lengthened.
• After 18 seconds, fewer than 10 of the
  trigrams were recalled correctly.

4
Short-term Memory

• Duration
• Peterson and Peterson (1959)
• Conclusions
• If rehearsal is prevented, information vanishes
  rapidly from short-term memory.
• Therefore, decay is the mechanism for
  forgetting in short-term memory.

5
Short-term Memory

• The key point about findings based on the
  Brown-Peterson technique is that items disappear
  from short-term memory only when rehearsal is
  prevented. New items can only take their place if
  existing items move on.
• Repetition serves as a method of continually
  reinserting the information into the short-term
  memory and thereby strengthening the memory.
• Without rehearsal, the duration of short-term
  memory is less than 30 seconds.

6
Long-term Memory

• Encoding in LTM involves semantic coding (see
  Baddeley, 1966) although can use other sensory
  modes (visual, acoustic, even olfactory?)
• Duration of LTM very difficult to study! How
  many of your classmates will you remember in 50
  years time? (see Bahrick et al, 1975)
• Capacity of LTM unlimited? Computer analogy
  Merkle (1988) number of synapses suggest between
  one thousand and one million gigabytes- small
  compared to modern computers yet secret is
  organisation..

7
Multi-store model

• This model arose from the information processing
  approach where memory is characterised as a flow
  of information through a system.
• The system is divided into a set of stages and
  information passes through each stage in a fixed
  sequence.
• The Multi-store model suggests that incoming data
  passes through a sensory store where it can be
  registered for very brief periods of time before
  decaying or being passed into a short-term memory
  store.

8
Multi-store model

• There are two types of sensory storage
• Iconic storage the store associated with visual
  information
• Echoic storage the store associated with
  auditory information
• Atkinson and Shiffrin believed that memory traces
  in the short-term memory are fragile and can be
  lost within 30 seconds unless they are rehearsed.

9
Short-term Memory

• Difficult to assess capacity of short-term
  memory two main ways are span measures and
  recency effects in free recall.
• Span measures e.g. digit span (Jacobs, 1887)
  require the participant to repeat back
  immediately a list of random digits in the
  correct order.

10
Short-term Memory

• Span of immediate memory generally seven, plus or
  minus two, whether the units are numbers, letters
  or words.
• Miller (1956) proposed that the capacity of short
  term memory depends more on the number of chunks
  rather than individual units.
• Primacy/recency effect create list of 20 words,
  let partner view it for 30 seconds and ask them
  to write down as many as they can remember in any
  order. (A- straight away, B-interference task)

11
Short-term Memory

• Primacy and recency effects measured in studies
  of free recall, participants recall words from a
  list in any order immediately after it has been
  presented.
• Recency effect last few items in a list are
  usually better remembered than items in the
  middle.
• Primacy effectfirst few items in a list are
  usually better remembered than items in the
  middle.
• Why?

12
Short-term Memory

• Recency effect based on information in the
  short-term store- suggests capacity is about 2 or
  3 items.
• Effect eliminated when interference task
  introduced between end of presentation and start
  of recall
• Primacy effect depends mainly on extra
  rehearsal, items no longer in short-term store?
• Also known as serial position effect (Galnzer and
  Cunitz, 1966)

13
Short-term Memory

• The recency effect suggests that the capacity of
  short-term memory is only 2 or 3 items, however
  memory span measures indicate a capacity of about
  7 items. One reason for this discrepancy is that
  participants carrying out a span task will
  rehearse as many items as possible, whereas those
  asked to learn a list for free recall rehearse
  only a few items at a time.
•  
• Both measures indicate that the capacity of
  short-term memory is strictly limited.

14
Short-term Memory

• Rehearsal
• There are different types of rehearsal. Craik and
  Watkins (1973) distinguished between maintenance
  rehearsal, where, for example, a word is repeated
  out loud a number of times to keep it in
  short-term memory, and

15
Short-term Memory

• Elaborative rehearsal, in which information is
  processed in terms of its meaning. They suggested
  that maintenance rehearsal may be enough to keep
  information in short-term memory, while
  elaborative rehearsal is the type necessary to
  transfer information from short-term memory to
  long-term memory.

16
Short-term Memory

• Craik and Lockhart (1973) proposed the levels of
  processing theory which assumes that attentional
  and perceptual processes operating at the time of
  learning determine what is stored in long-term
  memory.
• Various levels of processing range from shallow
  or physical analysis of a stimulus (e.g. letter
  detection) to deep or semantic analysis.
• Craik (1973, p.48), Depth is defined in terms of
  the meaningfulness extracted from the stimulus
  rather than in terms of the number of analyses
  performed upon it.

17
Memory Processes

• Hyde and Jenkins (1973) presented word lists
  related or unrelated in meaning and asked
  different groups of participants to complete one
  of the following five tasks
• 1) rate the words for pleasantness
• 2) estimate the frequency of each word
• 3) detect the occurrence of the letters e and
  g
• 4) decide whether the list words fitted sentence
  frames
• 5) decide part of speech appropriate to each word
  (e.g. verb, noun)

18
Memory Processes

• Half of the participants were told to learn the
  words (intentional learning) whereas the other
  half were not (incidental learning)

19
Memory Processes

• Results
• Test of free recall
• Recall 51 higher after semantic tasks than the
  non-semantic tasks with the associatively
  unrelated words
• Recall 83 higher after semantic tasks than the
  non-semantic tasks with the associatively related
  words
• Incidental learners recalled same number of words
  as intentional learners
• Conclusion nature of processing activity, rather
  than intention, at time of learning impacts on
  recall

20
Evaluation of Multi-store model

• Three different kinds of memory stores- model
  based on evidence that there are important
  differences among the stores.
• Claims
• Memory consists of 2 separate storage areas
• Information is moved from STM-LTM via rehearsal
• STM and LTM are unitary stores that cannot be
  subdivided

21
Evaluation of Multi-store model

• Evidence
• 1. Temporal duration (e.g. STM- 18seconds
  (Peterson and Peterson,)
• 2. Storage capacity (e.g. Miller (1956) 7/-2
  chunks of information and Glanzer and Cunitz
  (1966))

22
Evaluation of Multi-store model

• 3. Forgetting mechanism (e.g. STM-primacy/recency
  effects, different types of rehearsal)
• 4. Case studies
• Scolville and Milner (1957) HM- hippocampus
  removed to treat epilepsy- unable to form new
  long term memories
• Baddeley and Warrington (1970) studied KF brain
  damage from motorcycle accident- no problem with
  LTM but digit span only 2 items
• Task- find out about Clive Wearing

23
Evaluation of Multi-store model

• 5. Physiological evidence e.g. PET and fMRI scans
  show prefrontal cortex active for STM tasks
  (Beardsley, 1997) whereas hippocampus is active
  when LTM engaged (Squire et al, 1992)

24
Evaluation of Multi-store model

• Problems and limitations
• Rehearsal does not always lead to storage
  (Craik and Lockhart, 1972).
• If coding in the short-term memory is mainly
  acoustic, how do we understand language?
• How does the model explain flashbulb memories
  (Brown and Kulik, 1977)?
• Some types of information are not amenable to
  rehearsal, e.g. smells and yet we remember them.

25
Evaluation of Multi-store model

• Problems and limitations
• Are STM and LTM unitary stores?
• Warrington and Shallices study of KF found
  short-term memory forgetting of auditory
  information much greater than visual stimuli
• Tulving (1972) injected participants with
  radioactive gold and instructed to think about
  events or facts- different parts of brain active
  during retrieval of episodic and semantic memories

26
Evaluation of Multi-store model

• Memory oversimplified in this model?
• Tendency to minimise the active involvement of
  the individual in learning and remembering.
• Q. Describe and evaluate the Multi-store Model.
  (12)

27
Working Memory

• Baddeley and Hitch (1974) argued that concept of
  STM should be replaced with WM. WM consists of
  three components
• Central executive modality-free component of
  limited capacity (like attention)
• Articulatory or phonological loop- holds
  information briefly in a phonological or
  speech-based form (in revised model divided into
  speech perception and speech production)

28
Working Memory

• Visuo-spatial sketch pad specialised for spatial
  or visual coding
• Central executive most important other systems
  used for rehearsal for specific purposes

29
Working Memory

• Baddeley, Thomson and Buchanan (1975) studied the
  articulatory loop.
• Asked participants to recall sets of five words
  in the correct order. Word-length effect better
  with short words than long words. Indicates
  limited capacity of articulatory loop.
• If each component in working memory has a limited
  capacity and is independent of the other
  components then predictions can be made.

30
Working Memory

• Dual-task technique used to test these
  predictions. If two tasks using the same
  processor are carried out concurrently,
  performance on one or both tasks will be
  impaired.
• Example of how theories on abstract concepts can
  lead to hypotheses and controlled research.

31
Working Memory

• Hitch and Baddeley (1976) tested two predictions
  of the working memory model
• 1) If two tasks make use of the same component,
  they cannot be performed successfully together.
• 2)If two tasks make use of different components,
  it should be possible to perform them as well
  together as separately.

32
Working Memory

• Participants carried out verbal reasoning
  task-involved deciding whether a set of sentences
  provided a true or false description of the
  letter pair that followed it (e.g. B is followed
  by A AB).
• Assumed that this task makes use of the central
  executive.

33
Working Memory

• At the same time as carrying out the verbal
  reasoning task, participants were given an
  additional task (one of four conditions) before
  each trial
• the repeatedly
• one, two, three, four, five
• a different random string of digits out loud
  every trial
• or there was no additional task (control).

34
Working Memory

• It was assumed that the first two conditions
  would make use of the articulatory loop only
  because little thought or attention is involved.
• In contrast, saying six random digits involves
  the central executive as well as the articulatory
  loop.
• Predicted that only condition 3 should interfere
  with performance on the verbal reasoning task.
• Results reasoning performance slowed down only
  when the additional task was condition 3.

35
Working Memory

• Evaluation
• Working memory model concerned with both active
  processing and brief storage of information.
• Multi-store model minimises active involvement of
  the individual in learning and remembering.
• Therefore, working memory model relevant to
  everyday tasks such as mental arithmetic, verbal
  reasoning and comprehension.

36
Working Memory

• Evaluation
• Working memory model views verbal rehearsal as an
  optional process (one of two specialised systems)
  rather than the main process for retention of
  information.
• Yet, little known about central executive,
  visuo-spatial sketch pad and the interaction
  between these components.

37
Memory Processes

• How do we investigate the effects of learning
  processes on long-term memory?
• Obvious method would be to ask participants to
  learn a list of words using different tasks (e.g.
  counting letters, using imagery etc..)
• If participants aware they are taking part in a
  memory experiment then they may be tempted to
  process the words more thoroughly than the task
  requires.

38
Memory Processes

• To control for this, experimenters deceive
  participants about the true purpose of the
  experiment and incidental learning can be
  investigated as opposed to intentional learning.
• In this way, psychologists can use the
  experimental method to investigate learning and
  memory.

39
Memory Processes

• Craik and Lockhart (1973) proposed the levels of
  processing theory which assumes that attentional
  and perceptual processes operating at the time of
  learning determine what is stored in long-term
  memory.
• Various levels of processing range from shallow
  or physical analysis of a stimulus (e.g. letter
  detection) to deep or semantic analysis.
• Craik (1973, p.48), Depth is defined in terms of
  the meaningfulness extracted from the stimulus
  rather than in terms of the number of analyses
  performed upon it.

40
Memory Processes

• Key theoretical assumptions
• Depth of processing of a stimulus has a
  substantial effect on its memorability
• Deeper levels of analysis produce more elaborate,
  longer lasting and stronger memory traces than do
  shallow levels of analysis
• Theory supported by findings of Hyde and Jenkins
  (1973) and other studies.
• Criticisms?

41
Memory Processes

• Original approach over-simplified- e.g. not only
  depth of processing but other types of processing
  may affect memory
• Elaboration of processing important (Craik and
  Tulving, 1975)
• E.g. amount of processing of a particular kind

42
Memory Processes

• Study
• Elaboration manipulated by varying the complexity
  of the sentence frame between simple (She cooked
  the ------)
• and complex (The great bird swooped down and
  carried off the struggling ------)
• Results
• Cued recall twice as high for words accompanying
  complex sentences
• Suggests that elaboration benefits long-term
  memory
• Later studies have found that type of elaboration
  and not just amount of elaboration is important
  (Bransford et al, 1979)

43
Memory Processes

• Distinctiveness
• Eysenck (1979) argued that memory traces that are
  distinctive or unique in some way will be more
  readily retrieved than memory traces resembling
  others
• Depends on context in which a given stimulus is
  presented

44
Memory Processes

• Processes occurring at the time of learning have
  a major impact on long-term memory
• Common sense? Studies of value in terms of
  identifying processes such as elaboration and
  distinctiveness as factors influencing learning
  and memory
• How to measure depth of processing? Difficult to
  determine whether certain tasks involve shallow
  or deep processing

45
Memory Processes

• Criticisms
• Morris, Bransford and Franks (1977) found that
  information is remembered only to the extent that
  is relevant to the memory test
• Study
• Participants had to process words in terms of
  their meaning or sound

46
Memory Processes

• Some tested by a rhyming recognition test-e.g. if
  word ship had been on the list then
  participants should select the word whip from
  the rhyming test
• Results
• Found that recognition memory was much better
  for words that had been processed in terms of
  their sound than those processed in terms of
  their meaning