Transcript Memory Judgments

Recapitulation I:
Causal reasoning in
science
 It is impossible to establish causal
relations without doubt since it is an
instance on inductive reasoning.
 Causal reasoning in science:
 Proposition of causal theories / models.
 Assessment of causal theories / models by
means of data and existing well established
theories.
 Problem: Causal models could be
empirically indistinguishable.
Recapitulation II:
Simpson’s paradox
 Summing of relevant variables in
contingency tables can result in
erroneous conclusions.
 It is not allowed to sum over variables
that are related to more than one variable of the table.
Recapitulation III:
Ecological fallacy
 Treatment of clustered
data:
 Ignoring the clustered
structure of the data
can lead to wrong
conclusions.
 Specifically, between
and within cluster
relationships may differ.
Recapitulation IV: Regression to the mean
110
120
130
Regression line always below main axis (cf. formula)
x
100
y
x
80
90
x
y
x
70
y
y
xy
80
90
100
x
110
120
y|x
Regression to the mean
120
130
Recapitulation V:
100
x
~
y
x
y
y
x
x
y
80
90
~
y
y
x
70
y
110
x ~
y
80
90
100
x
110
120
y
Regression artefact: Lord’s paradox
Concept: Lord’s Paradox:
ANCOVA for testing differences between two
groups (due to an intervention) while controlling for preexisting differences between
groups leads to a different conclusion than
one based on a t-test that compares pre-post
intervention differences between groups.
Comment:
ANCOVA is used to test for the effect of one or
more independent variables (the intervention)
while partialling out the effect of covariates.
Regression artefact: Lord’s paradox
Example: Lord’s Paradox:
2 Groups of students:
(i) Switch from secondary to Grammar
school.
(ii) Staying for one more year in secondary
school (switching after the year).
The two groups differ with respect to their
language capabilities: The switching group is
superior to the non-switching group.
200
Regression artefact: Lord’s paradox
Grammar School (N = 250)
100
Posttest
150
 Obviously, switching
has no effect on language performance.
 Regression lines
have different
intercepts (due to
regression to the
mean).
 ANCOVA indicates
difference, t-test
of pre-/post difference not.
50
Secondary School (N = 250)
50
100
150
Pretest
200
Regression artefact: Lord’s paradox
Conclusion: Lord’s Paradox:
 Experimental control of confounding
variables is superior to statistical
control.
 Statistical and methodological sophistication together with knowledge of
the subject should help to design
studies exhibiting the required
experimental control.
Memory Judgments
Cognitive Mechanism: Encoding:
 Adjustment to existing knowledge.
 Gist / meaning extraction.
 Drawing inferences.
 Activation of related information.
 Processing of value and emotional
aspects.
Encoding is not storage but the construction
and modification knowledge structures.
Memory Judgments
Cognitive Mechanism: Retrieval:
 Integration of information from
different sources.
 Consistency and plausibility checks.
 Filling in gaps due to inferences and
judgments.
 Subjective theories relevant.
Retrieval from memory consists in a process
of (re-) construction.
Memory Judgments
Cognitive Mechanism: Forgetting:
 Interference of information: Blocking
of retrieval by related information.
 E.g. intrusions of irrelevant related
information blocks the retrieval of
irrelevant information (=Output
interference).
Memory Judgments
Experimental paradigms for testing
memory errors:
 Elisabeth Loftus: Misinformation
paradigm.
 DRM paradigma (Deese, Roediger, &
McDermott) for eliciting false
memories.
Memory Judgments
Method of misleading post event
information:
 Misleading information after the
relevant event provided.
 E.g. misleading questions that need
not be incorrect but may only result
in specific conclusions.
Memory Judgments
Misleading questions:
Following to a film about a traffic accident,
participants were asked the following
question (with participants in different
groups received different wordings):
»About how fast were the cars going when
they (smashed / collided / bumped / hit /
contacted) each other?«
Post-event information influenced peoples’
testimonies, e.g. whether they saw
breaking glass.
Memory Judgments
Misleading information: suggestion
Jean Piaget reported that he had vivid memories
of an attempt to kidnap him out of the baby
carriage on the Champs-Élysée. He »remembered« the gathering of people, the scratches in the
face of his heroic nurse that had saved him, the
white push stick of the policeman, and the fleeing offender.
However, the event never happened. Many
years later the nurse confessed to have invented
the whole story.
Memory Judgments
Misleading information: Suggestion
The 43 years old deputy Paul Ingram was accused by his
daughters to have mistreated them sexually in their childhood.
Initially he denied vehemently all the accusations since he
had absolutely no memory about these events. However,
his colleagues, the officers, and the priest (Ingram was a
member of a fundamentalist church) assured him that he
will remember the events after having made a confession.
Following to long talks and examinations Ingram finally
confessed and stated that he had probably repressed his
memories of the events.
Memory Judgments
Misleading information: Suggestion
However, the officers believed that the sexual abuse has
taken place in the context of Satanism, and, in fact, in the
course of the examination Ingram »remembered« increasingly better the various events and accused further
people that denied vehemently to have committed the
criminal acts.
At the same time Ingram’s daughters »remembered«
further details of the Satanism like killing of babies, and a
mass orgy.
Memory Judgments
Misleading information: Suggestion
On waiting on his lawsuit Ingram was visited and interviewed by social psychologist Richard Ofshe. The latter
asked Ingram to remember how he has forced his son to
have sexual intercourse with his daughter before his eyes.
Comment:
This event follows a logic that is similar to the other accusations. However, none of Ingram’s daughters has ever
claimed that such an event had occurred, and also the son
denied it vehemently.
Memory Judgments
Misleading information: Suggestion
Ingram‘s reaction to the questions of the psychologist followed a predictable pattern: First he could not remember.
However, following to intensive visualization and praying
he developed vivid »memories« concerning the respective
event.
Comment:
Finally, Ingram was convicted of imprisonment for 20
years. At the time of Schacter’s report (6 years after the
conviction) he had not yet been released.
Memory Judgments
Comments on Suggestion:
 The examples demonstrate the
impact of suggestions on memory.
 Children are extremely prone to suggestens and, thus, in general not apt
as witnesses in trials.
Memory Judgments
Cognitive mechanisms:
 Replacement of original memory with
post-event information (has not turned
out to be valid).
 Reality monitoring: Ability to separate
real events from imagined, dreamed etc.
events.
 Source monitoring: Correct attribution of
the source where the information comes
from.
Memory Judgments
Cognitive mechanisms:
Suggestions and misleading information, as
well as imagery technics influence both
reality and source monitoring:
 Specifically, imagination technics are able
blur the borders between reality and
imagined events.
 Repeating information can make it more
difficult to correctly identify the source
where the information comes from (cf.
the case of Piaget).
Memory Judgments
Therapeutic interventions and false
memories



Therapeutic techniques used to »recover« lost
memories can result in false memories.
Initially Sigmund Freud believed that neurotic
symptoms were the result of sexual misuse in
childhood. Later on he abandoned this theory
attributing reports of sexual abuse to patients’
imagination.
Jeffrey Masson’s accusations that Freud had swept
the issue of sexual abuse under the carpet could
never be confirmed.
Memory Judgments
DRM (Deese-Roediger-McDermott)
paradigm: Experimental creation of vivid
false memories (associative memory
illusions).
 People learn lists with words that are grouped
around a lure item:
 e.g. Critical lure (never presented): sleep
List of items to be learned: Bed, rest, awake,
tired, dream, wake, snooze, blanket, doze,
slumber, snore, nap, peace, yawn, drowsy.
Memory Judgments
DRM (Deese-Roediger-McDermott)
paradigm:
 After each list, participants performed a free
recall test where they should reproduce as many
items as possible.
 After presentation of all 24 lists, a recognition
test was performed in which participants had to
indicate whether an item was OLD or NEW.
 In case of an OLD item people should also
indicate, whether they remembered details of the
event the item was presented (=remember
response).
Memory Judgments
DRM paradigm: Results

Free recall: Intrusion of lure items:
The lure items that were never presented were recalled
approximately as often as old items in the middle or the
list.

Recognition test: slightly more critical words than words
actually on the list were categorized as old.

Of the critical items actually produced in the recall task a
remember response was given in 73 percent of the cases,
i.e. participants believed that they recalled some details of
the presentation episode.
Memory Judgments
DRM paradigm: Interpretation
Implicit associative response:
The presentation of an associated item leads to
the activation of the critical lure item (by
means of semantic priming and perhaps in
terms of conscious deliberation: thinking of the
lure).
This increases the familiarity of the lure items.
Memory Judgments
DRM paradigm: Interpretation
Implicit associative response can explain the
high rate of intrusions in free recall.
However it can not explain the high rate of
remember judgments.
The later may be explained by means of source
monitoring: People had actually »seen« the
lures at the moment they reproduced in free
recall. They erroneously attributed to having
seen them on the list.
Exercises:
 Exercise 2-7: Ecological fallacy
 Exercise 2-8: Lord’s paradox