Models of Memory
Atkinson and Shiffrin (1968)
3 components of memory: Each component is characterized by:
● Sensory memory ● Duration (how long it can hold info)
● Short-term (STM) ● Capacity (how much info it can hold)
● Long-term (LTM)
Multi-store Model of Memory
Sensory Memory Store
Does not process information
Detects info; holds until transferred to STM or lost
Visual, auditory, olfactory, etc.
Capacity is limited to our perception
Duration is short (just a few seconds)
Short-term Memory (STM)
Attention brings info out of sensory memory
Modal differences of input is erased
Capacity: 7 +/- 2 chunks of information
Duration: generally less than 30 seconds
Long-term Memory (LTM)
Rehearsal increases duration of STM; brings traces to LTM
Capacity: large amounts of information (potentially unlimited)
Duration: indefinite
Retrieval (not storage) is problematic
Support for the Multi-Store Model
Partial Report Technique - Sperling (1960)
Participants flashed a grid of alphanumeric characters
Whole-report- fill in all characters (35% accuracy)
Partial-Report- fill in one row(75-100% accuracy)
Stimuli data contained in memory temporarily and accessable
Support for the Multi-Store Model
Serial Position Effect - Glanzer and Cunitz (1966)
Tendency to recall first and last items of a list (not the middle)
● Primacy effect and recency effect
Participants asked to memorize list of 20 words (free-recall)
● 1st group started immediately after words were read
● 2nd group were delayed by filler task (counting, etc)
Support for the Multi-Store Model
Glanzer and Cunitz (1966) Continued
1st group (started immediately)
● Primacy and recency effect occur
2nd group (delayed)
● Primacy effect preserved
● Recency effect disappeared
Rehearsal explains results
Without it STM decays in 30 sec.
STM and LTM must be seperate
Any problems/criticisms of this model of memory?
Questions about Multi-store Memory
1. Are the memory stores really distinct and separate?
2. Are there really three memory stores, not more, not less?
3. Are sensory modalities within sensory memory just modalities?
Why not separate memory stores?
4. Is there a physiological basis for the memory stores or are they
just constructs?
5. Is rehearsal necessary and sufficient for the transfer of information
from STM to LTM? Can is succeed without? Fail with?
6. Does information only flow in one direction? For example, can LTM
influence what pieces of data are selected from sensory memory?
Criticism of Multi-Store Model
Levels of Processing - Craik and Lockhart (1972)
Recall is a function of the depth of our processing
The deeper our processing, the stronger its trace in LTM
Deep processing involves building the stimulus into a structure of
meaningful connections and associations (linking to prior knowledge)
Criticism of Multi-Store Model
Levels of Processing
Model (con’t)
Criticism of Multi-Store Model
LTM might not be a unitary store
● procedural, episodic and semantic memory
STM may have more components
● working memory model
Criticism of Multi-Store Model
Levels of Processing - Craik and Tulvig (1975)
Working Memory Model
Baddeley and Hitch (1974)
Focuses on structure of STM
Dual-task technique- subjects
asked to perform simultaneous
tasks (auditory and visual)
STM should be distracted and/or
unable to preform both tasks
Working Memory Model
Working
memory= central
executive and
manager
Visuospatial-
holds visual and
spatial info
Phonological
loop- inner ear
and inner voice
Support for Working Memory Model
Phonological similarity effect - Conrad and Hull (1964)
Participants to recall one of two lists of letters
● Phonetically similar: BDCGP
● Phonetically different: FHPRX
Rhyming lists were more difficult to remember
Traces of similarly sounding letters easier to confuse
Supports idea that speech material uses a sound-based storage
Support for Working Memory Model
Articulatory suppression - Baddeley, Lewis and Vallar (1984)
Participants to block “inner voice” rehearsal
● Repeat a sequence of sounds (ie. the-the-the or one-two-three)
Auditory inputs can enter but rehearsal impossible
Rhyming words difficult to remember
Visual inputs cannot be recoded into sounds
Recalling rhyming/non-rhyming words no different
Criticism of Working Memory Model
Strengths:
More sophisticated model of memory
Explains wider range of memory phenomena
Does not overemphasize role of rehearsal in memory
Criticisms:
Complex models are difficult to empirically test
Testing complexity makes model difficult to falsify
Exact role of components remains unclear (esp. central executive
and episodic buffer)
Ignores other memory structures (sensory and LTM)
