4. Memory manipulation - gleneaglesyear12psychology

Research findings indicate that much of what we
recall from LTM is NOT an accurate representation
of what actually happened previously.
Memory includes some exact details but also
some additions that are logically plausible, things
that might have happened, filtered and shaped by
our thoughts, attitudes and beliefs as well as by
who we are as individuals and social beings.
Human memory is able to remember relevant
fragments of what we experience, but is prone to
errors and distortions.
Unlike a camera or audio recorder that captures a
perfect copy of information, the details of human
memory can change over time.
Without our conscious awareness, details can be
added, subtracted, exaggerated or downplayed.
Confidence in memory is no guarantee of its
accuracy, in fact there is little correlation.
Some researchers believe that strong confidence in
the memory of minor details may actually be a cue
that the memory is inaccurate or even false.
(Reisberg, 2007; Hockenbury & Hockenbury, 2006)
Why so many errors?
Memory is not simply recorded but actively constructed
To form a new memory you actively organise and encode
different types of information (visual, auditory, tactile etc.)
When you later attempt to retrieve this information, you
actively reconstruct or re-create the details of your
Over time recalling memories can change the content and
meaning of a memory
Your experience of memory is NOT a guarantee that it
reflects reality!
Bartlett’s research on
manipulation of memory
British psychologist Frederick Bartlett in 1932
first drew attention to the reconstructive nature
of human memory.
Bartlett believed Ebbinghaus studied human
memory in an artificial way. E.g. Using
nonsense syllables to control influence of past
experience excluded important variables that
impact on human memory in everyday life such
as the influence of our prior experiences,
attitudes and expectations.
Bartlett’s research on
 Participants read prose (story or essay) or looked at a
Participants were asked to recall and describe the prose
or draw the picture on several later occasions.
Each time the original stimulus material was remembered
slightly differently.
e.g. Unusual or unexpected events were described in a
more logical or sensible way (as if to match their personal
beliefs of what might likely be true).
Bartlett concluded that we tend to remember only a few key
details of an experience, and that during recall we
reconstruct the memory, drawing on our personal values,
beliefs, and expectations to make up and add missing bits
in ways that complete the memory in a logical or plausible
way. Usually done without conscious awareness of it
Loftus’s leading question research
Other studies have confirmed Bartlett’s conclusions and
extended his findings.
U.S. Psychologist Elizabeth Loftus and various colleagues on
eyewitness testimony.
Loftus found that eyewitnesses similarly reconstruct their
memories and their testimony is therefore not always accurate.
Many of Loftus’ studies involve showing participants a short film,
video or slides of an event such as a car accident and then
asking them specific questions about the scene they “witnessed”.
Sometimes information that was not present in the actual scene
or which contradicts the scene is introduced. At other times,
leading questions are asked.
Loftus’s leading question research
The way questions are asked during eye
witness testimony can influence memory
Shows how recall does in fact involve an
active reconstruction of events, that is open
to suggestion
Leading Questions – are questions that
have content or are phrased in such a way
that suggests what the desired answer is or
to lead to the desired answer.
Example of a “Leading Question”
e.g. To a witness of a car accident “How fast was the car going
when it ran the stop sign?”
This is a leading question according to Loftus (1975) because it
contains a preposition (i.e. Information that should or must be
true in order for the question to make sense).
The question presupposes or assumes there was a stop sign.
But what if there was no stop sign?
The witness might answer the question anyway because it was
a question about how fast the car was going and not a question
about the presence of a stop sign or whether the car ran the
stop sign.
Loftus proposed that the way the question was worded the witness
might add the new false information about the stop sign to their
memory of the event. Then they would be more likely to recall it
as a part of their reconstructed memory when answering a
question about it such as “Did you see a stop sign?” at a later
Loftus’s leading question research –
study no. 1
One of Loftus’ research on memory
reconstruction was conducted with US
colleague John Palmer (1974). Two lab
experiments that investigated the influence of
question wording on memory and how
information supplied after an event can distort a
witness’s memory for that event.
45 volunteer uni student participants shown 7
clips of car accidents. Clips were short excerpts
from driver education films ranging from 5 to 30
seconds long.
Loftus’s leading question research –
study no. 1
Participants were asked to write a description of the
accident after viewing each clip.
Also participants were asked to answer some
specific questions about the accident, including a
critical question requiring an estimation of speed of
the cars involved in each collision.
There were 5 conditions in the experiment, with 9
participants randomly assigned to each condition.
Each condition had a different verb used to complete
the critical question (therefore different versions of
the question were asked).
Loftus’s leading question research
– study no. 1
All participants asked the question – ‘About
how fast were the cars going when
they________ each other?’
The gap was filled with
Smashed (Condition 1)
Collided (Condition 2)
Bumped (Condition 3)
(Condition 4)
Contacted (Condition 5)
Loftus’s leading question research
– study no. 1
To control the potential influence of the order in which the clips were
viewed, the clips were presented in a different order to each group of
The most “intense” verbs resulted in the highest speed estimates and
the least “intense” in the lowest speed estimates.
See table on next slide.
Differences in speed estimates were found to be statistically significant at
Loftus and Palmer suggested that the different estimates could have been
influenced by the verbal label used to characterise the intensity of the
car crash.
They also suggested that results could have been influenced by response
bias e.g. Being uncertain about the exact speed, participants may
have adjusted estimates to fit with their expectations of the researcher.
Loftus’s leading question research
– results study no. 1
Participants estimates of speed were
influenced by how the question was asked
Loftus’s leading question research –
study no. 2
150 volunteer uni student participants randomly assigned to 3
groups (conditions) again viewed accident footage – this time a
1 minute film that included a 4 second scene of a multiple car
Participants asked ‘About how fast were the cars going when
they________ each other?’
Gap filled with:
(Condition 1)
Smashed (Condition 2)
One group not asked about the speed of the cars (Condition 3)
Differences in mean scores of each group were found to
statistically significant.
Loftus’s leading question
research –study no. 2
One week later all participants asked 11 questions
about the accidents without viewing the film again.
The critical question ‘did you see any broken glass’
was asked to all 3 groups
This question was “hidden” among the other
questions that served as “distracter” questions and
placed in a random position on the participant’s
question paper.
There was no broken glass in the film
Loftus’s leading question research –
study no. 2 results
Mean speed estimate (MPH)
Verb condition
Control (no
Loftus’s leading question
research –study no. 2 results
Participants ‘recollection’ of seeing glass
was influenced by how the original speed
question was asked a week earlier.
Although most participants accurately
reported not seeing any broken glass, more
participants who had been given the verb
“smashed” (16) reported seeing broken
glass than did those who had been given the
question with the verb “hit” (7).
Loftus’s leading question
research –study no. 2 results
Loftus and Palmer suggested in the first part participants
formed a memory of the car crash they had witnessed
after viewing the film.
They had integrated their memory of the accident with the
new false piece of information added after the event –
smashed or hit - which caused some to remember broken
glass that wasn’t really there. They had reconstructed a
new distorted memory.
Suffered source confusion – unable to distinguish real
info from false info given after the event. It arises when
the true source of the memory is forgotten or when a
memory is attributed to the wrong source.
Conclusion – memories can be altered by post event
exposure to inaccurate info during questioning
Other studies
Numerous other studies by Loftus and others
have since found that the memories of
eyewitnesses are reconstructions, instead of
exact replicas, of the events witnessed.
These studies also confirm Loftus’ findings
that eyewitness memories can be altered by
post-event exposure to inaccurate information
introduced during questioning.
See Fig. 8.3 on page 396 “Model of memory”
that integrates new information.
False Memories
Participants shown footage of
impossible events
Children shown film of meeting bugs
bunny at Disneyland which is
impossible as bugs is a Warner
Brothers brand
When questioned later the subjects
exposed to the film were more likely to
‘remember’ such an encounter than
the control group
Read Box 8.1 on page 396
False Memories
Participants read stories
about them getting lost in
the shopping centre as
 Told the info comes from
 About 25% of participants
after reading the story
report recalling the event
Elaboration and organization
Elaboration enables effective encoding
The more associations that are made
between new information and that
already in memory the more likely the
information will be retrieved
Making it meaningful makes it easier to
Improving memory
Observing unusual events (e.g. Teacher on
top of the flagpole at school) is often
enough to ensure a scene is remembered.
But often we must learn information that is
more difficult and this requires conscious
effort. Mere exposure, even very frequent
exposure, to information is often insufficient
to produce efficient remembering.
Attention must be given to the information,
the it must be organised and integrated into
the information already stored in LTM.
Context and State dependent cues
Retrieval can be enhanced if the conditions under
which the information was originally learned is
recreated – either the external environment
(context) or the learner’s internal environment
This approach is based on the encoding
specificity principle. (Tulving, 1983)
Context dependent – environmental cues in a
particular setting (context) where a memory was
formed that act as retrieval cues to help access
the memories formed in that context. E.g. sights,
sounds, smells of a specific location.
Experiment on context dependent
Goddon and Baddeley (1975) wanted to compare
the efficiency of land training and underwater training
of deep sea divers.
Divers with a list of 40 unrelated words in either of
two settings - on the beach or under about 5 metres
of water.
Divers then asked to recall words either in the same
environment or the alternate one.
Results showed divers recalled up to 20% more
words when words were learned and retrieved in the
same context.
Context dependent cues
Eyewitnesses taken back to the crime scene to recall
apparently forgotten information.
The environmental cues at the crime scene act as
additional retrieval cues that assist recall of the
Studies on students taking exams in the same room as
work learned showed any differences are sufficiently small
so as not to be of concern for students.
However, this is only when the learning environment and
the testing environment are similar (i.e. Different rooms In
the same school)
If the testing environment is substantially different from
the learning environment, the differences in performance
are likely to be more noticeable.
State dependent cues
State Dependent cues are associated
with an individual’s internal physiological
and/or psychological state at the time the
memory was formed, which act as retrieval
cues to help access those memories.
Information is more likely to be
remembered if the person is in the same
state as when they learned it e.g. drunks
hiding an object often have difficulty finding
it when sober but when drunk again can
recall where they hid the object.
State dependent cues
It seems that aspects of the psychological
and physiological state (in the previous
case, intoxication) may become encoded
with new memories.
State dependent retrieval involves better
recall of information when the physiological
and/or psychological states or learning and
retrieval match.
This DOES NOT mean that intoxication
improves memory – it actually impairs
memory as it interferes with encoding.
State dependent cues
Your mood also provides state-dependent
retrieval cues.
We seem to associate good or bad events with
their accompanying emotional state.
The emotional states become retrieval cues
when we feel good or bad again, and they trigger
memories that are consistent with the mood.
However, there is a complicating factor:
 When happy we recall happy events, but when we feel
“down” our minds can become flooded with sad or
negative events of the past which in turn darken our
interpretations of current events.
Mnemonic devices
Many mnemonic devices were developed in
ancient times.
Mnemonic devices are techniques for
enhancing or improving memory.
They make use of information already stored in
They don’t simplify information, they make it
more elaborate.
More information is stored, not less.
Mnemonic devices
The additional information makes the material
easier to locate and retrieve because it has
enhanced organisation in LTM.
Tend to organize new information into a
cohesive whole, so that retrieval of part of the
information generally assists retrieval of the
rest of it.
The better the information fits with what we
already know, the easier it is to retrieve.
Mnemonic devices
Each mnemonic device is based on a
particular kind of elaboration, encoding
or rehearsal strategy to enhance
Some mnemonic devices emphasise
logical organisation, others rely on the
use of visual imagery to elaborate the
Mnemonic devices
Acronyms are pronounceable words formed
from the first letters of a sequence of words.
Acronyms can be used for remembering other
types of information.
e.g. Roy G Biv (red, orange, yellow, green,
blue, indigo, violet) – colors of rainbow or
visual spectrum
Mnemonic devices
A rhyme is a phrase or a string of words (such
as a jingle) often with an emphasis on similar
sounding key words.
e.g. “i” before “e” except after “c”
They organise information by associating the
information with a particular rhythm (sound)
and with rhyming words.
We know an error in retrieval has occurred when
the rhyme is ruined or the rhythm broken.
Mnemonic devices
Acrostics, or the first letter technique
involve making verbal associations for
items to be remembered by constructing
phrases or sentences using the first letters
of the information to be remembered.
e.g. “Every Good Boy Deserves Fruit”
to denote the lines on a treble staff in
music. - E, G, B, D, F.
Acrostics are useful when you have to
remember the information in sequential
Mnemonic devices
The peg-word method consists of
memorising a rhyme or jingle that has
mental “pegs” or “markers” on which you
“hang” the items to be remembered.
Very effective for a list of information to be
remembered in a particular sequence. E.g.
Shopping items, errands, historical events.
Mnemonic devices
1 – Bun
2 – Shoe
 3 – Tree
 4 – Door
 5 – Hive
The rhyming words are the retrieval cues (the “peg”
words), then you must associate the information to be
remembered with the rhyme by using visualisation.
Visualisation is particularly effective when the image of
the peg-word object and the image of the object to be
recalled interact rather than merely being paired side by
The peg-word method has been shown through
experimental evidence to significantly improve recall of
information from semantic LTM.
Mnemonic devices
Narrative chaining involves linking
otherwise unrelated items to one
another (“chaining”) to form a
meaningful sequence or story
See Table 8.3 on page 405
Narrative chaining research –
Bower and Clark, 1969
Given a word list to learn
 Experimental group – told to use
narrative chain
 Control group – given no instructions on
memory techniques
 narrative chaining group recalled 93% of
 Control group recalled on average 13%
Narrative chaining research –
Bower and Clark, 1969
The sample was good
 The variables were clearly defined
 Extraneous variables were controlled (research
 The results good – experimental group performed
significantly differently to the control group
Ah but how do we know this difference is a real
one…. Not just some random event?
Research on narrative chaining
Participants who used narrative chaining remembered 6
times more information than participants who learned by
simply repeating the words to themselves. (Loftus, 1980)
Narrative chaining adds organisation and
meaningfulness to otherwise meaningless material.
Particularly useful when needing to remember information
in a particular order.
Research has found that it is a useful technique for
people with memory impairment as well as people with
normally functioning memory. (Wilson, 1995)
A narrative will not help though, if it hangs together so
loosely that you cannot remember the story (Matlin, 2002)
Method of Loci
The Method of Loci uses a well-learned
sequence of locations as a series of retrieval
cues for the information to be recalled.
Particularly useful when needing to remember
a list of items in a particular order.
Mentally “move” through the locations
sequentially without difficulty.
Involves a mental image of the items to be
remembered and visually linking them with a
particular location.
Method of Loci
If any item is an abstract or unusual concept
that is hard to visualise, it must be changed
into an object that can be visualised instead.
E.g. Justice = courtroom
The method of loci elaborates information to
be learned by linking it with memory cues that
are very familiar.
Experimental findings indicate that it can
improve memory by a factor of 2 to 3 times
over serial recall without the use of a
mnemonic device.
Key-word method
A mnemonic device for learning LOTE that
involves visual imagery as a way of
encoding information to improve retrieval.
e.g. Spanish word for “horse” is “caballo”
pronounced “cab-eye-yo”. “Eye” could be
chosen as the key word to help remember, then
a horse with a very large eye could be
Massed rehearsal – a couple of extended
study sessions over a very short period of time
Spaced or distributed rehearsal – spacing
out study sessions over an extended period of
Research suggests that if long-term retention
of information is required, spaced rehearsal is
a more effective strategy.
Read Box 8.6 on page 411 about effective
studying using the SQ3R method.

similar documents