Remembering and forgetting Flashcards
How can we measure forgetting
recall test
recognition test
What is a recall test
recall of events, story recall (hard to score)
free recall of lists of nameable items (% recalled),
cued recall e.g. paired associates (% recalled)
serial recall (% in correct position)
What is a recognition test
ability to discriminate “old” from “new” items (% correct)
Who was Ebbinghaus (1885) and what did he do?
learned many lists of 13 nonsense syllables to criterion (2 correct serial recalls), and then relearned each after a variable interval
Forgetting is orderly, explain
forgetting, measured appropriately, can often be described by a simple mathematical function of the retention interval (here a power function)
What causes forgetting?
orderliness of forgetting might suggest some inevitable decay process: loss from storage.
but this can’t be the whole story:
- information not recalled now may be recalled later
- further prompts or cues may succeed in eliciting recall
so some cases of “forgetting” due to retrieval failure not loss
moreover, some memories show essentially no loss over time
Why does a longer retention interval not necessarily increase forgetting?
no forgetting of school class-mates over 30 years (assessed using yearbooks (Bahrick et al., 1975, J.Exp.Psy:Gen)
“flashbulb” memories (e.g. JFK assassination, 9/11 attack)
but forgetting of former students by teachers does increase with interval —>
Why?
teachers subsequently encounter many more students
so, forgetting attributable to interference from other similar memories?
How can we test ‘interference’ v decay theories?
under normal circumstances, retention interval — time in storage — is confounded with the number of other experiences accumulated during the interval.
so, control the interval, vary intervening experiences
if forgetting due to interference, then p(recall) should decrease with more exposure to similar stuff, with time held constant
How is some forgetting clearly attributable to interference?
‘paired associate” learning (1940s,1950s): participant must learn (say) 10 arbitrary pairings between “stimulus” and“response” words.
P learns List 1 to criterion
then learns List 2 to criterion
test on either List 1 or List 2
later recall of List 1 worse when List 2 was learned afterwards - retroactive interference
later recall of List 2 worse when List 1 has been learned before - proactive interference
implication: retrieval difficulty increases when other similar material has been learned, holding retention interval constant.
How are time v intervening (similar) experiences predictors of event forgetting?
at end of season, two rugby teams recalled games played: clear forgetting over a season (though some games are more memorable)
each player missed some games:
- if control for time, the number of games played during the interval is a significant predictor of forgetting
- but not vice versa.
in general, there is ample evidence that retrieval failure is increased by interference from similar material
What factors influence p(retrieval)?
processing at encoding/acquisition
consolidation after encoding
(more on) Interference from other memory traces at retrieval
similarity of encoding and retrieval contexts
[underlying theme: memory is an associative system, not a container]
What is organisation at acquisition
deliberate rote rehearsal does increase later recall
hence the primacy effect in free recall – first few items get more rehearsals
but mere rote rehearsal is a relatively ineffective learning strategy
What do incidental memory exps show?
processing the meaning of, and actively organising, the material are effective learning strategies
Mandler (1967) study
groups 1 and 2 sorted words on cards into 2-7 categories of their own devising
group 1 were also told to try to learn the words. Group 2 were not
no difference in a later recall test (if control for N of categories).
group 3, who just placed the cards into columns while trying to learn the list, remembered less than Groups 1 and 2.
Moral of organisation at acquisition
organising the material is what produces effective acquisition, not effort to learn (by itself)
What is depth of processing at acquisition?
Craik and Tulving (1975) study
Craik and Tulving (1975) study
showed a series of unrelated words, and gave one of three orienting tasks:
is it written in upper/lower case?
does it rhyme with X?
does it fit into a sentence(e.g. “The man broke his ____”)
later unexpected recognition test
Moral of depth of processing at acquisition
processing the meaning is better than processing surface form
(unless: details of surface form are what you are required to remember – “transfer-appropriate processing”)
What do mnemonics demo?
the power of appropriate elaboration at acquisition
one-is-a-bun mnemonic for sequence learning
first, learn the rhyme
to learn a sequence: form a vivid image representing the rhyme item for position X in interaction with what you want to remember at position X
remembering: Just go through the rhyme
Method of loci
first, memorise a route around a familiar building or garden, so that you can walk around it in your “mind’s eye”
learning the speech: At each point in the route, form a striking image representing the idea you want to mention at that point
remembering: Traverse the route, reading and interpreting the images.
What does learning form?
associations
What does retrieval do to associations?
activates them
Why do mnemonics work?
they bind ideas to a pre-established framework which organises them (e.g. serial order, in the case of the method of loci and ‘one-is-a-bun’)
imagery encourages formation of rich nexus of associations between the frame “hook” and the concept attached to that hook.
forces exhaustive retrieval attempts
General moral of associations, retrieval and mnemonics
effectiveness of learning involves forming associations among representations that already exist in the mind, including representations of
elements of the new experience/fact
elements of the context
elements of prior knowledge
Retrograde memory loss after concussion
John Terry (at that time at Chelsea) “I remember walking out for the second half and nothing else until waking up in the ambulance on the way to the hospital”.
What is phase 1 of consolidation?
after traumatic brain injury (e.g. concussion) or ECT, there is often retrograde memory loss spanning many minutes, even hours (way beyond duration of “working memory”),
disruption of process of consolidation of memory trace in hippocampal/medial temporal cortex system
What does consolidation of novel traces suffer from?
interference from consolidation of further novel traces
sleep improves memory for material learned in last few hours (Jenkins & Dallenbach, 1924)
alcohol and barbiturates impair learning (disrupt consolidation)
but improve memory for material learned just before! (“retrograde facilitation”)
What is phase 2 of consolidation?
over a longer timescale (months, years), recent LTM traces are more vulnerable to hippocampal damage than older traces
amnesic patients with damage to hippocampus/ medial temporal cortex typically show gradient of retrograde amnesia over years – older memories better preserved, more robust [“Ribot’s law”, 1882] – e.g. HM
over time, re-activation of distributed representations in cortex from hippocampal trace reinforces direct links between elements of those representations, creates robust traces no longer dependent on hippocampus.
What is associative interference at retrieval?
E.g. In paired-associate experiments, can vary similarity of stimulus terms, response terms, etc.
interference maximal when the same (or similar) stimulus terms are used for each list: competition between two associative links from the same retrieval cue
What is the ‘fan effect”
in fact retrieval as e.g. of associative interference
S learns 0-4 new “facts” about each of set of famous people, e.g.
Napoleon had a wart on his nose
later, true/false RT measured for test statements, e.g.:
Napoleon was Emperor of France (actual true)
Napoleon had a wart on his nose (experimental true)
Napoleon was six feet tall (false)
Mitigating the effects of associative interference
the fan effect seems paradoxical: the more you know about Napoleon, the harder it is to retrieve any one fact about him?
in the Lewis & Anderson experiment, the N facts learned were unrelated.
it has been claimed that if the facts are thematically related, the fan effect is eliminated. E.g.:
the princess went to the dockyard
the princess christened the ship
the princess broke a bottle of champagne (Smith et al, 1978 – but see Reder & Anderson, 1980)
the thematic relationship enables the learner to form associations between the separate facts using pre-existing knowledge schemas, which provide multiple retrieval paths.
Moral of mitigating the effects of associative interference
when studying, work to create multiple links among the facts you are learning about a topic, and to prior knowledge
What is remembering as reconstruction vulnerable to ?
associative intrusions
Bartlett’s (1932) “War of the Ghosts” story experiment
we interpret what we see and hear via learned “schemas” [schemata], or “scripts” – knowledge of typical patterns or event sequences
when we try to remember, we recover only fragmentary associations, from which we reconstruct the event/fact, filling in the gaps using
general knowledge schemas (in semantic memory) E.g. a restaurant meal —> a waiter (as part of the “script”)
fragments remembered from other episodic sources.
False memories
the “recovered memories” controversy: “Complete or partial memory loss is a frequently reported consequence of experiencing certain kinds of psychological traumas, including childhood sexual abuse. These memories are sometimes fully or partially recovered after a gap of many years…within or independent of therapy…..Although clear memories are likely to be broadly accurate, they may contain significant errors…Sustained pressure or persuasion by an authority figure could lead to the retrieval or elaboration of memories of events that never happened.” (British Psychological Society working party on recovered memories, 1995)
too cautious! It is easy to create false memories: “recognition”, “recall”, and “recollection” of events that did not happen, because:
source amnesia (retrieval of information coupled with inability to remember its source) is common,
recall is reconstructive (Bartlett): fragments of actual experience recovered from memory get combined with other information in memory whose source is lost (e.g the implicit or explicit suggestions of an interrogator or therapist).
Loftus’ ‘EWT’ exps
E.g. Loftus & Palmer (1974):
100 Ss see film of car crash
then answer a series of questions, including “About how fast were the cars going when they smashed into/hit each other?”
group 1 (“smashed into”)
group 2 (“hit”)
one week later, they answer some more questions, including“Did you see any broken glass?” [There was none.]
group 1 16/50 say “yes”
group 2 7/50 say yes (p
False memories of realistic and emotionally salient material
1992 El Al crash in Amsterdam: 10 months later : “Did you see the television film of the moment the plane hit the building?” >50% said “yes”, some with incidental detailsCrombag, Wagenaar, Van Koppen (1996, Applied Cognitive Psychology)
Loftus’s “lost in the mall” study: ~25% of subjects “remembered” (after several probing interviews) being lost in a public place as a child. Loftus & Pickrell (1995, Annals of Psychiatry)
Initial interview on memory for real childhood events included a strong suggestion of a false one, e.g. - “At a wedding reception when you were 5, when you ran into the table and spilled the punch bowl on the bride’s parents”. No memory reported then, but in later interviews 20-40% of participants reported this as a memory, and up to a half of these described “clear” recollections, with incidental details.
What are context effects and ‘encoding specificity?’
as a general rule, information is more easily retrieved if tested in the same context in which it was acquired.
E.g.: environmental context (Godden & Baddeley, 1975):
divers learned word lists either on dry land or under water.
they were tested on land or under water
in between, all subjects moved between environments
Eich, Weingartner, Stillman & Gillin (1975)
sensitivity of retrieval to congruence with the“internal” context at the time of learning is sometimes calledstate-dependent learning
similar effects of induced sad and euphoric moods (e.g. Teasdale & Russell, 1983).
encoding-specificity is causal in the maintenance of depression: negative memories are more accessible in the depressed state, and their retrieval reinforces the depression.
the power law of forgetting
if an appropriate measure of “trace strength” is used,
forgetting functions are linear in log-log plots
what is the principal determinant of forgetting?
the time in storage