Bernstein,
Godfrey, Davison, and Loftus. “Conditions affecting the revelation effect for
autobiographical memory.” (2004)
Introduction:
This study examines revelation effects on childhood
autobiographical memory by altering levels for processing of words during
study. This is done because of the assumption that autobiographical memory
errors are errors in familiarity attribution. The revelation effect is the
tendency for a person to say an item is “old” in a recognition task if the
target word is degraded or obscured and later revealed. This is usually done by
having a participant unscramble an anagram which will increase the confidence
of truth from participants that a life event did in fact happen during their
childhood. The experimenters were interested in discovering if pre-exposure
increased fluency and familiarity of words later used in the recognition task
to increase confidence ratings.
Experiment 1:
Participants first counted vowels of 48 words or short phrases
during an exposure phase. Twenty-four of these words later appeared at test.
Next, the participants were trained on solving anagrams. The anagrams were
unconstrained meaning that the sentence did not provide clues to the solution,
and they were based on a formula for solving. (The formulas
are provided on p. 456.) Participants then had to rate whether the
solved anagrams had occurred to them as a child on a scale from 1 to 8. The test phase was based on the “Life Events
Inventory” in which half of the phrases were shown in tact and half were
anagrams. These anagrams were constrained and could be solved easily. Of the
list items, half had been seen during the exposure phase. Finally, participants
took a recognition test where they had to state where certain words appeared in
the experiment based on four categories.
Unscrambling words in the context of life events increased
participants’ confidence of occurrence in childhood. The effect was seen for
both old and new items. The standard revelation effect was seen. However,
recognition performance was terrible. No prior exposure effects were seen.
Vowel counting has no direct effect on childhood autobiographical memory.
Experiment 2:
The same procedure was used with visualization used during the
exposure phase instead of vowel counting and ratings of image vividness.
Unscrambling words now only had an effect when previously seen. No revelation
effect for new words occurred, and overall there was no revelation effect.
There was also no effect of prior exposure. Old anagrams had higher life ratings
than new anagrams. Visualization significantly increased recognition
performance when compared with vowel counting. Visualization has no effect for
childhood autobiographical memory.
Experiment 3:
The same procedure was used as before but instead of
visualization, participants freely formed sentences using words during
exposure. There was no recognition test. All of the effects of experiment 2
were replicated. The differences observed from the past three experiments are
attributed to familiarity misattribution between exposure words and life event
anagrams.
Experiment 4:
The same procedure as the past few experiments was used except
that participants had to form sentences about childhood experiences from
exposure words. There was no interaction between prior exposure and
unscrambling. No overall revelation effect and no effect of prior exposure. By
making exposure relevant to task of judging life events, the main effect of
unscrambling in exp. 1 and unscrambling and prior exposure in exp. 2 and 3
disappeared. Participants no longer misattributed familiarity to childhood
history.
Discussion:
Solving an anagram in the context of a life event will increase
a participant’s confidence that the life event actually happened to them. But,
this increased confidence required prior exposure. If prior exposure was
incidental, then there was a revelation effect for both new and old items. But
if the prior exposure was deliberate, then the revelation effect only occurred
for old items. Finally, if the prior exposure was directly related to the life
event task, then the revelation effect was no observed. The authors show
problems with current criterion models because both of the models state that
anagrams should cause a more liberal criterion judgment for new words, but the
results from this study oppose that thought. These challenges arise from the
current studies ability to alter revelation effect by changing the level of
processing on exposure words.