Transcript Memory--retrieval

Memory--retrieval
For later. . .
Try to remember these words. . .
• Measures of memory
• Relative difficulty of different measures of
memory
• Role of cues in memory retrieval
• Retrieval and prior knowledge
Measures of memory
•
•
•
•
Free recall
Cued recall
Recognition
Savings in relearning
Measures
Free recall: Minimal information from experimenter:
experimenter simply says “Remember” and the context
is usually implied, occasionally described.
Cued Recall: Experimenter also gives part of the
information, or some related information.
Recognition: The to-be-remembered information is presented,
along with other stuff (distractors) and the subject must
distinguish new from old.
Savings in relearning: Subject learns some material to a
criterion, and the # of trials required is noted. Later subject
relearns the material to the same criterion; if fewer trials are
required, that is savings.
Difficulty
An important principle is this:
Whether or not it appears that someone remembers
some material depends on the way that you measure
their memory.
That is, it can appear that someone has forgotten some
material, but if you give them a different test, it is clear
that they remember it.
Measures--difficulty
In general, free recall is hardest,
then cued recall, then recognition.
Shepard & Recognition memory
Subjects see 512 words, then take a 2choice recognition test--get 88%
Subjects see 612 brief sentences (e.g.,
“The truck swerved to avoid the limping
deer”), then take 2-choice recognition test-get 88%
Final experiment = colored pictures cut
from magazines.
Shepard’s results
Percent correct
100
75
50
25
0
Immed.
2 hours
3 days
Delay
7 days
120 days
Caveat
Although it’s true that recognition memory is
usually very good, you have to bear in mind
that the difficulty of a recognition task
depends on the distractors.
Sensitivity
Although we can’t talk about the absolute
sensitivity of free recall, cued recall, and
recognition, we can talk about the relative
sensitivity (usually). IN GENERAL
Free recall< cued recall< recognition.
Hart, 1965, 1967:
When free recall fails, subjects recognize answer
50% of the time
Recognition more sensitive than free recall
Tulving & Pearlstone, 1966:
Condition 1: Free recall, then cued recall
Condition 2: Cued recall, then cued recall
Cued recall  50% more remembered than free.
Cued recall more sensitive than free recall
Nelson, 1978
When recognition fails, subjects relearn the
material faster than new material
Savings more sensitive than recognition
Why the difference in sensitivity?
It feels like memory success is some
combination of memory strength and
the sensitivity of the test.
Tests are sensitive to the extent that they
give you good cues.
Free recall: context (time and place)
Cued recall: context + partial information
Recognition: context + all information
The importance of cues.
Try to recall the words to “Star Spangled Banner”
Does # of cues (sensitivity) explain
everything?
No, what’s actually important is whether the
cues make you think of material the same
way at encoding and retrieval.
0.5
Proportion Correct
0.45
0.4
0.35
0.3
Shallow Encoding
Rhyming
0.25
0.2
Deep Encoding
0.15
Recognition
Rhyming
Meaning
Type of Test
Transfer Appropriate Processing
The importance of cues
We can extend transfer appropriate
processing: cues are important not just if they
are about sound (rhyming) vs. about meaning:
they are important if they make you think
about different things about material, even if
all the cues concern meaning
Barclay et al.
This effect works even if you don’t change
the word, but just emphasize different
properties of the word!
E.g., “The moving men struggled to get the PIANO up the stairs.
Cue: something that makes music (avg. 1.6 words/10)
Cue: something heavy (avg. 4.6 words/10).
We can take this effect one step further:
if a recall cue is better at making you think
as you did at encoding than a recognition cue,
you should be better at cued recall than
recognition.
Recognition failure of recallable
words
Participants see word pairs:
glue: CHAIR
ground: COLD
They are told they need only remember the words
in capital letters, but the other words might help.
Recognition failure of recallable words
They get two tests:
Recognition test:
CHAIR
NURSE
Cued recall test:
glue:
ground:
SHELL
Recognition failure of recallable words
Key finding: there are some words for
which people fail to recognize the
word, but then are able to recall it.
Why?
“glue-chair”
“chair”
The point: recognition failure of
recallable words effect is consistent
with the idea that different measures
of retrieval lead to better or worse
recall because they provide different
cues, and the cues are more or less
likely to make you think about
material the way that you thought
about it at encoding.
Test
Try to recall the word list that you
heard at the start of the class.
Why do people remember “sleep?”
• They might think about “sleep” at encoding
• They know the list is composed of words
related to sleep, so at retrieval, “sleep”
seems like a plausible list member.
Retrieval & prior knowledge
We can generalize from this “sleep” example
to say that memory is almost always a
combination of the actual event plus relevant
prior knowledge.
What type of prior knowledge?
Schemas
A memory representation of a type
of event, characteristics generally
true of the event, not of a specific
event.
Schemas
• What’s in your schema of a doctor’s visit?
Things that conflict with the schema are attended to
and remembered. Things that are not part of the
schema but are irrelevant are not.
Schemas
• Would it stand out if the nurse didn’t say
“The doctor will be with you shortly” after
leaving you in the exam room? Would you
perhaps remember that she did say that
week later?
At encoding: schemas make atypical things
stand out, and they are memorable.
At retrieval: schemas make it seem likely
that typical things happened, even if they
didn’t.
Retrieval success varies with cues. Cues
vary in the extent to which they make
you likely to think about material the
way you did at encoding.
Most memory researchers believe that all
memory is a reconstruction. You
combine what happened with what you
believe probably happened and that’s
your memory.