Transcript Modal Model and Encoding and Retrieval

Modal Model of Memory
PSY 421 – Fall 2004
Overview
• Methods of conducting memory
experiments
• Typical memory findings or results
• Definitions with memory research – a
summary
• Long-Term Memory and the Serial Position
Effect
• Problems for the Modal Model
• Encoding and Retrieval
Memory Methodology
• The following is the typical procedure for a
memory experiment:
– Stimuli/items (letters, numbers, words, pictures)
are presented with the instruction that subjects
are to study these words for a later memory test
(order of the items may be important)
– A time period passes during which the subject
must remember the stimuli and may be asked to
also perform another task (counting backwards,
viewing unrelated items, solving math problems)
Memory Methdology, continued
– Subjects are asked to remember the stimuli that they studied
• Free Recall test – write down all the items from the study list that
you remember
• Cued Recall test – here are some cues to help you remember the
items from the study list
• Serial Recall test – remember the items from the study list in the
order that you studied them
• Recognition test – here are some of the items you studied among
other items called distractors – chose the items that you studied
(called targets)
• Relearning – number of trials to be able to remember all the items
correctly (Ebbinghaus)
• Reconstruction – attempting to remember items and other elements
associated with the items to "rebuild" the original event from
memory
– Performance is recorded as accuracy (number of items
remember correctly), sensitivity (number of hits and false alarms
from a recognition test), or speed (amount of time it takes the
subject to remember one or all of the items)
Demos
Typical Memory Results
• Free Recall – subjects remember some but not all
of the items that they studied (performance is
better or worse depending upon the type of
independent variable manipulated (IV))
• Cued Recall – subjects remember more items that
were studied than if they did free recall (but
performance depends on the IV)
• Recognition – easiest of all memory tests, subjects
remember items well to the extent that the targets
(studied items) vary from the distractors (nonstudied items)
• Order of memory – when the order of the studied
items is important, subjects typically remember the
beginning and the end items better than the middle
items – primacy and recency effects
Similar Terminology in Memory
• Sensory Memory
– Iconic or Visual Sensory Memory
– Echoic or Auditory Sensory Memory
• Short-Term Memory or Primary Memory or
Working Memory
• Long-Term Memory or Secondary Memory
• Memory Processes
– Acquisition = encoding = learning = input
– Retention = storage = holding information
– Retrieval = remembering = responding = output
• Memory Tests
– Explicit Memory Tests = Direct Memory tests
– Implicit Memory Tests = Indirect Memory tests
New Definitions
• Long-Term Memory
– Episodic Memory – memory for events and experiences
(including peripheral contextual info)
– Semantic Memory – memory for knowledge or meaning
of information
– Other Characteristics
• Duration – limitless
• Type of information coding – all modalities
• Capacity - limitless
• Serial Position Effect
– Primacy – memory for earlier items in a list
– Recency – memory for later items in a list
• Encoding – taking in information - acquisition
Serial Position Curve
1
0.9
Probability of Recall
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1
2
3
4
5
6
7
Serial Position
8
9
10
11
Serial Position Effect
• STM and LTM
– Multiple-stores explanation for serial position effects
(Waugh & Norman, 1965)
• Recency occurs because of rehearsal and the items at the end
of the list are still in STM (dump of STM)
• Primacy occurs because beginning items were rehearsed the
longest and have already been transferred to LTM
– STM and LTM are not different (Crowder, 1993; Nairne,
1992)
• Memory in the short-term is using memory for a short period of
time; memory in the long-term is using memory over a longer
period of time; just because we use our memories for different
periods of time does not mean there are 2 different forms of
memory
• Primacy effects can still be shown without rehearsal
• Recency effects can be found with information from LTM
(presidents; Roediger & Crowder, 1976)
More Problems for the Modal Model
•
1. The modal model predicts that if there is a delay between study and test and
rehearsal is prevent with a distractor task, then recency will decrease b/c words
that were in STM have faded away (or been forgotten) (Glanzer and Cunitz,
1966).
•
Opposing Evidence (Bjork & Whitten, 1974) - if a distractor activity is added after EVERY item,
including the final item (called the continual distractor paradigm - subjects are continually unable to
rehearse b/c of distraction), the recency effect COMES BACK, even though there was no rehearsal for
any of the items. [For more discussion of this topic - see pgs 72-76 of your memory text)
•
2. The modal model claims that STM codes information on the basis of how it
sounds and LTM codes information on the basis of its meaning
•
Opposing Evidence - there is evidence of acoustic interference in LTM (Dale & McGlaughlin,
1971) and evidence of semantic confusion errors (were items that mean the same thing are confused
with each other) in STM (Shulman, 1970, 1972).
Opposing Evidence - as you will see in the discussion of levels of processing, different types of
processing can occur in what is assumed to be STM (Craik & Lockhart, 1972)
•
•
3. The modal model assumes three separate storage units
•
Opposing Evidence - where does SM end, STM begin and LTM begin? There will always be
contamination of one store with another. Since information has to travel from STM to LTM and from LTM
to STM to be retrieved, how do we know if or what information is coming from LTM in retrieval? (Nairne,
1996).
•
4. Primacy comes from rehearsing information into LTM and recency comes
from dumping STM
•
Opposing Evidence - primacy can occur with no rehearsal (Greene, 1986; Neath & Crowder,
1996)
•
•
Opposing Evidence - recency can occur even when there have been long delays between study
and test (Greene, 1986; Nairne, Neath, Serra & Byun, in press, Neath, 1993)
Encoding
• How information gets into memory
• Can vary depending upon the nature of retrieval (if
known) – encoding and retrieval do not do well
independently of each other
• Factors that affect encoding
– Attention
– Repetition – multiple experiences with the item(s)
– Rehearsal
• Maintenance rehearsal – simple repetition
• Elaborative rehearsal – thinking about the meaning of the item
and associating the item to information already in memory
• Levels of Processing Theory
• Transfer-Appropriate Processing
• Encoding Specificity Hypothesis
Levels of Processing Theory
• Craik and Lockhart (1972); Hyde and Jenkins (1973)
• Memory structure (SM-STM-LTM) not important for this
theory
• Type of processing is important
– Shallow processing = processing features or physical
characteristics
– Deep processing = processing on the basis of meaning or
associations
• Results
– Deep processing leads to better memory performance
• Problems
– Define "deep"
– Only focuses on encoding, not retrieval or the interaction of the
two
Transfer-Appropriate Processing
• Considers both encoding and retrieval – makes
this different from Levels of Processing
• Processing leads to better memory performance
when the processing at retrieval (test) matches the
processing used during encoding (study)
• Morris, Bransford & Franks (1977)
– Study - Deep (sentence fit) vs shallow processing
(rhyming)
– Test – standard recognition or rhyming recognition
– Results
• If subjects in the "deep" condition took the standard recognition
test, they remembered more than subjects in the "shallow"
condition (taking the standard recognition test)
• If subjects in the "shallow" condition took the rhyming
recognition test, they remembered more than subjects in the
"deep" condition (taking the rhyming recognition test)
Encoding Specificity Hypothesis
• Recollection of information depends on the interaction between
properties of the encoding information and properties of the retrieval
information
• Endel Tulving – developed ESP
• Thomson and Tulving (1970)
– Study – targets were presented alone during study or were presented with
a weak cue
– Target = flower; weak cue = fruit; strong cue = bloom
– Test – target words were presented alone, with a weak cue, or with a
strong cue (never studied)
– Results
• If a target word was NOT paired with a cue at study, it was best remembered
when it was paired with a strong cue
• If a target word was paired with a weak cue at study, it was best remembered
when it was paired with a weak cue at test.
• Previously, it was thought that strong cues were best for improving
memory performance. Thomson and Tulving showed that performance
depended on if the weak cue was presented at study – if it was, then
performance was best when the weak cue was again paired with the
target word
Context Effects in Memory
• A change in performance when information is
studied in one context and remembered in a
different context
• Environmental Context – incidental aspects of
the environment, not necessary for encoding or
retrieving information
• Semantic Context – gives meaning to items
(e.g., bank – money or river)
• Example research:
– Abernathy (1940) – classroom change
– Smith (1979) – classroom change
– Light and Carter-Sobell (1970)
Context Dependent Memory Effects
• Memory can be worse when tested in a new or different
context relative to when context stays the same between
study and test
• Works for recall memory tests – memory is worse with a
change in context from study
• Does not work for recognition memory tests –
performance is not different with a change in context
• Explanations
– In recall, you are recalling/remembering the item and all that is
associated with it, which includes the context
– In recognition, the item (target) is given back to you at test so you
do not need additional information (like context) to help with
remembering the item
– How can you get context effects in recognition – with response
bias – subjects tend to be more conservative in their responses
with a change in context from study to test.
State Dependent Memory Effects
• Psychological/emotional state or pharmacological state
(not a US state)
• Application – the reason why we don't remember our
dreams better is because when we are awake, we cannot
capture the "state" that we were in while dreaming (Prince,
1910)
• If you learn something while you are calm and are tested
in a state of shock (e.g., earthquake) then you will have
difficulty remembering what you have learned (Susokita,
1934)
• Happy and neutral moods – best performance if you
match the moods between study and test (Bartlett &
Santrock, 1979)
• Alcohol, marijuana, nicotine – best performance if you
match the pharmacological state between study and test
– Performance is the absolute worst if you study in the
pharmacological state and test in a sober state (cues are different)
– but not vice versa