lecture 17 - encoding and retrieval Flashcards
(47 cards)
long term memory
diagram in notes of how its organised
its memory that endures for hours, months, yrs, decades
two types of LTM
- non declarative memory - memory you can’t declare, its implicit including things like procedural memory eg knowing how to ride a bike you can’t articulate it but can demonstrate the action. also includes simple classical conditioning
- declarative memory - memory that can be declared or explicit memory - there are two sorts of memory within this - semantic memory = memory for facts and episodic memory = memory of our personal past/events, it contributes to our sense of self, allows you to mentally time travel, peoples quality of episodic memory differ
Examples of episodic memory
- When did you last go to the beach?
- What did you have for breakfast this morning?
- Can you remember what you did for your 18th birthday?
birthday?
Factors which
affect encoding and retrieval
Encoding (getting information in) - how objects and events in world come to be represented in our memories
Storage (holding it) - consolidation
Retrieval (getting it back out) - happens via some search
memory is not a passive lookup process its an active reconstructive process - if the memories are degraded we might need to fill in the blanks even if the memory is intact - we can sometimes add details that seem to be true so there can be some inaccuracies or errors in memory
- Levels of processing
- Transfer-appropriate processing
- Organisation
- Context-dependent memory
- State-dependent memory
- Retrieval strategy
encoding
Levels of processing (LOP)
- Craik & Lockhart (1972) – the major determinant of an
item’s memorability is the LOP it receives at encoding.
shallow level of processing - your just looking at perceptual features eg how many vowels does the word contain
deeper level of processing - sound of the word, phonology of the word
in the deepest stage you would access the semantics
Craik & Tulving (1975)
Words presented visually and participants asked to
make three different types of judgments:
Question Yes No
1. Is the word in. TABLE table
capital letters?
- Does the word. crate. market
rhyme with
‘weight’? - Does the word
fit in this. orange roof
sentence? ‘The
man peeled the
afterwards they have an unexpected recognition test - unexpected as dont want ptps to be doing any different things at encoding except what they are asked to do. in test phase ptps are give some these studies items and they are randomly intermixed with some new items and are asked whether these words are old or new. found that when ptps access the semantics in the sentence task they recall far more than in the shallow encoding conditions
results graph in notes
Q1 = shallow level of processing
Q2 = deeper
Q3 = deepest level of processing
Greater the depth of processing the
better the subsequent memory.
- But….is it just that semantic judgments take longer to
make or are more difficult? - No. Craik and Tulving (1975) made the shallow task very
difficult.
before ptps saw a word they saw on screen cs and vs (c= consonant and v= vowel) and then see a word and then have to say afterwards whether the pattern fits the word - this is a shallow task
|CVCCVC| |WITCH|. |pattern fit?|
- Shallow words = 57 %, Deep words = 82%. Not due to
time/difficulty. - it tells us something about the processing of the words so the deeper that you process the words the better will be your subsequent recall or recognition of them - Generalisability - as a rule of thumb is probably one of the most generalisable findings in human memory research - its very robust and reliable and its useful for anybody that wants to improve their memory capacity
The limit of levels
Can a valid measure of processing depth be obtained
that is independent from the amount remembered? If
not, is it a circular argument? - as we often say that was processed in a deep manner because its led to better memory or we have got better memory became this item was processed in a deep encoding fashion = circular
------------------ / \ Deeper \ processing Better memory \ / -------------
- Do features get processed sequentially or in parallel? - the framework generally says you access the superficial features then the phonology and then the semantics so there this kind of sequential processing of info but that is probably not true in that many different features of a stimulus may be processed at the same time rather than in strict serial fashion - so things may be happening in parallel rather than sequentially
- Does deeper encoding always lead to better memory? mostly but not always eg you need practical practice of driving a car before your driving test
Transfer-appropriate processing
(TAP)
- The way in which you encode materials needs to
“match” the way in which you test memory. - Morris, Bransford, & Franks (1977)
Two encoding tasks (incidental learning):
Deep: ‘The ___ ran into the lamp post: car’ - does it fit in sentence task
Shallow: ‘Does it rhyme with fighter: writer’ - rhyme
Two recognition tests:
Standard: old words (car) &
new (fish) - differentiate between words previously studied and new words
Rhyme: series of words, has an item
been presented which rhymes with
that word (bar, lighter)
found for the standard test better recognition and memory for the semantic task vs shallow task
for rhyme task opposite happens you are better if you encoded things in the rhyme task and not the semantic task
this indicates that the way in which you encode material needs to match the ay in which your going to be tested
Why is deeper encoding better?
Elaboration: semantic processing allows a richer more elaborate code. - allows items to the be more readily retrieved as when we produce a more elaborate memory it usually links to many different concepts so they you can access it in a variety of ways
- Craik & Tulving (1975). Judge whether a word would or would not fit into a sentence:
Simple: ‘She dropped her pen’
Complex: ‘The little old man hobbled across the castle courtyard and dropped his pen in the well’. - ptps remember better when there is a more complex sentence - when there is more elaboration around the word
The one who THINKS over his
experiences most, and weaves them
into systematic relations with each
other will be the one with the best
memory…
James, 1890, p. 662
Why is deeper encoding better? 2
Distinctiveness: semantic processing allows participants
to encode more unique features from each word relative to surface features.
* Prediction: If deep LOP make memories more distinct,
then false recognition should be lower for words
subjected to deep than shallow encoding.
Tested by Gallo, Meadow, Johnson, & Foster (2008)
Organisation
when we try to remember things sometimes they could be linked in some way and we can use that to help us remember
- Tulving and Pearlstone (1966)
List 1: pink, green, blue, apple, cherry, plum, lion, cow, rabbit
List 2: table, river, gun, iron, dentist, hand, square, shirt, pen
Which one is easier to remember? - Tulving (1962)
found list 1 is easier to remember
Repeatedly presented with the same set of words to
remember. Words not designed to be related to each other.
Participants found ways of linking the words and tended to
remember words in clusters. People’s natural instinct is to
organise and this helps memory.
Bower et al. (1969)
Participants needed to remember a list of 112 minerals.
Can organisation help?
Minerals presented organised or scrambled
graph in notes
in test one the organised group recalled over three times as much material as the scrambled group
materials being in an organised format aids retrieval
it doesnt matter if you know your going to be tested or not the critical thing is what you do with the materials
Linking to what we know can be
beneficial but beware of distortion
Bower, Karlin and Dueck’s
(1975) “droodles” - meaningless patterns was presented to ptps and they were asked to free recall on a blank sheet of paper - ptps were poor at doing this. when they were given an interpretative label of the droodle their recall was greatly improved
Carmichael, Hogan and Walter
(1932) – the label influenced
people’s drawings. - ptps were given some figures and labels they were ambiguous objects and ptps were given different labels for the objects than each other and asked to draw the object. they found the label biases the perception and storage of the objects - in recall you get a figure thats more in line with the label underneath than what the actual figure was
Prentice (1954) found that the bias arose at
retrieval and not encoding.
retrieval
Context-dependent memory
- Incidental context can be important to retrieval.
- Godden & Baddeley (1975) tested deep-sea divers.
- 40 unrelated words. - some ptps encoded those words on land and some encoded them underwater - some were tested and water and some were tested on land - some had matched conditions some didnt
- recall was better if learnt and recalled on land or learnt and recalled underwater
- recall was worse for mismatching conditions
- Effects increase in size as the delay between encoding and retrieval gets longer.
- You can mentally
reinstate the context.
State-dependent memory
Also occur when the learner’s internal environment is
changed by alcohol, drugs, caffeine.
* Sometimes only demonstrated when free-recall used
and not recognition.
* Can also be physiological states: Miles & Hardman (1998) examined aerobic exercise.
* Encoding conditions: rest on exercise
bike or pedalling (had to have Heart Rate: 120-150)
* Free-recall: rest or pedal.
* 20% better recall when there
is an encoding-retrieval
match.
some of the effects for context and state and generally bigger when you have free recall where you have no cues for retrieval rather than recognition which creates quite small effects. good thing as mean memory is not totally reliant on being in same state and context.
Retrieval strategy - order
The ways in which we interrogate our memories can be
important for what we retrieve.
* Whitten & Leonard (1981) demonstrated that retrieving
names of teachers from early schooling years was more
accurate when starting with later years and moving to
earlier ones, rather than vice-versa.
* Why? It is easier to successfully retrieve details from
more recent episodes which may provide more prompts
for the earlier more difficult names.
order in which we retrieve info can be important for what we remember and the accuracy of what we remember
Retrieval strategy - perspective
Anderson & Pritchert (1978) participants read a story
about boys missing school and hiding out in the house of
one of them. Given details of the house.
* During reading: asked to adopt perspective of a burglar or a
homebuyer.
* Later test both groups recalled a similar amount but bias
towards what was relevant to their perspective.
* Second recall attempt: same or different perspective.
* Those adopting a different perspective remembered more
items relevant to the new perspective.
* Retrieval improved because of a change in strategy
we can adopt a Perspective when recalling the past which guides and restrains our retrieval to things that are relevant to that perceptive or schema so we not be recalling things outside that perspective
cognitive interview - change perspective
Overview of today – Factors which
affect encoding and retrieval
Encoding
(getting
information in)
* Levels of Processing
* Transfer appropriate
processing (also at
retrieval)
* Organisation
Storage (holding it)
Retrieval (getting
it back out)
* Context-dependent memory
* State-dependent memory
* Strategy
Carmichael, Hogan, and Walter (1932)
presented the visual stimuli for subsequent recall. Each item was sufficiently ambiguous as to fit two
different verbal labels, for example a beehive or a hat. If participants were given a label at encoding, would the underlying concept of the label influence the way people remembered the picture later on? The answer is was very clearly yes. When participants were later asked to draw the stimuli from memory, their drawings were strongly influenced by the label they had been given. It is tempting to think of this again as a bias in the way in which the material was perceived and stored. However, a subsequent study by Prentice (1954) suggested otherwise. The encoding conditions were the same as for the Carmichael et al. study, but retrieval load was minimized by using recognition rather than recall. The label effect disappeared under these circumstances, suggesting that the bias occurred at retrieval rather than encoding; the appropriate information was stored but the difficult task of recalling by drawing led to an undue influence of the participants’ background knowledge underlying the verbal labels.
Bower, Karlin, and Dueck (1975)
in a study in which people were asked to recall apparently meaningless patterns or “droodles”. Free recall of these patterns was very poor. However, recall was greatly improved when each droodle was accompanied by an interpretative label. Bower et al. conclude that memory is aided whenever contextual cues during encoding arouse appropriate schemata. We will return to the powerful influence of schemata on encoding and consolidation later in this chapter.
As can be seen in these examples, Bartlett’s approach to memory focuses on how episodic encoding, and learning in general, occurs in the context of a person’s existing knowledge. More broadly, Bartlett shifted the focus away from quantitative factors such as total study time or number of repetitions during learning to the meaning underlying a stimulus.
meaning enhances episodic memory encoding
Partlett’s principal criticism of Ebbinghaus was that his attempt to separate memory from meaning by using nonsense syllables meant that he was studying simple repetition habits that were not especially relevant to the way in which our memories work in everyday life. A basic premise is Bartlett’s approach was that people seek to identify meaning in their experiences, which he described as an effort after meaning. The history of memory research since Ebbinghaus has very strongly supported Bartlett’s perspective, providing repeated demonstrations of the benefits of meaning and knowledge to how well experiences are encoded
early evidence of a role of meaning
Early Evidence for the Role of Meaning in Memory
- Critique of Ebbinghaus’s Approach:
Ebbinghaus aimed to exclude meaning using nonsense syllables.
Later research (e.g., Glaze, 1928) showed participants often associated syllables with real words (e.g., CAS → castle, cast), affecting recall.
More meaningful syllables were easier to remember (Jung, 1968), showing the influence of prior knowledge.
- Role of Language Habits:
Even with slow presentation rates, forming associations was hard unless syllables resembled real words.
Bartlett suggested that “repetition habits” and familiar language patterns aid memory (supported by Baddeley, 1964; Underwood & Schulz, 1960).
- Shift Toward Meaningful Material:
By the 1960s, word lists replaced nonsense syllables to better reflect the role of meaning.
Studies showed that pre-existing associations (e.g., bread–butter) improved paired-associate learning.
- Free Recall and Clustering:
Deese (1959) and Jenkins & Russell (1952) showed that words with strong associations were more easily recalled and often clustered during recall.
These findings highlighted the role of prior knowledge in encoding and retrieving episodic memories.
- Episodic Memory Relevance:
Though word lists seem artificial, recalling specific words from a list reflects episodic memory.
Each presented word creates a “mini-event” in memory, allowing controlled study of episodic encoding.
- Imagery and Memory:
Paivio (1965, 1969, 1971) demonstrated that imageable words are better remembered due to dual coding (visual + verbal).
Abstract words (e.g., hope) are harder to recall than concrete ones (e.g., apple).
Interactive imagery (e.g., crocodile biting a football) enhances recall.
- Role of Coherent Meaning:
Sentences or word sets with a shared, coherent meaning (e.g., List C: “Elephants terrified by flames…”) are recalled better.
Combines Bartlett’s emphasis on meaning construction and Paivio’s focus on imagery.