Thomas, A. K., Bulevich, J. B., & Loftus, E. F.
(2003). Exploring the role of
repetition and sensory elaboration in the imagination inflation effect. Memory
& Cognition, 31, 630-640.
Memory for recent and
childhood events can be influenced by imagination. Often, confidence that a fictitious event occurred increases
after that event has been imagined, a phenomenon called imagination
inflation. Source monitoring errors
also increase after imagining events.
Imagination inflation can be accounted for by both the source monitoring
framework and familiarity misattribution theory. The goal of this study was to examine which of these accounts
more for imagination inflation. If it
is driven more by perceptual detail, then actually imagining events would be key
and the source monitoring framework would be the best explanation. If inflation is driven more by general
familiarity characteristics, then imagining the event might not be necessary,
and familiarity misattribution theory would better account for the phenomenon.
According to the source
monitoring framework, characteristics of memories such as thoughts or images
are attributed to particular sources of past memories. We assess these and other characteristics in
order to determine the credibility of a memory. Perceived events usually contain more perceptual detail than
imagined events, which involve more cognitive processes. Perceived memories are often more clear and
people can better remember how they felt about it. Source misattribution errors occur when characteristics of the
memory are not encoded or when cues between two sources are similar.
Familiarity could also bias
recognition. Previous exposure can
increase familiarity, and this can affect how easily information is
processes. Ease of processing may
influence source judgments.
Participants may rate the familiarity of an event instead of their
confidence that it occurred.
The researchers hypothesized
that participants that are provided with more sensory information during
imagination will create more false memories because false memories are driven
by misattributing specific characteristics.
Previous research has shown that more vivid visual images result in more
false memories. This suggests that the
source monitoring framework best accounts for imagination inflation.
Experiment 1
Participants were shown
objects that they might need to perform or imagine tasks. They either performed or imagined 48 tasks
in the first session. 24 were
performed, 24 were imagined, and there were 24 that were not presented. Half of the actions were familiar and half
were bizarre. Participants were given
15 seconds to perform or imagine the action.
In the second session, all statements presented were imagined either 0,
1, or 5 times. They imagined 8 statements
from each of the first groups (performed, imagined, not presented) for each of
the conditions in the second session.
Participants used either simple or elaborate imagining. They then had to rate the vividness of the
imagination. In the third session,
participants had to answer questions based on what they remembered from session
1. They answered whether an action had
been presented. They then answered
whether it was performed or imagined, and made remember/know judgments on those
actions.
For actions that were not
presented in the first session, imagining the action more times resulted in
more false did answers. Elaborate
imagination also led to more false did responses. Imagination inflation was a little larger for familiar actions
than bizarre actions. There was also an
interaction between number of imaginings and type of imagining. False responses increased as imagining
increased and with elaborate imagining.
Participants almost never said they had performed an unpresented action
if they had not imagined it in session 2.
Elaborate imagining resulted in more R judgments, as did the type of
action. Familiar actions received more
R judgments. There was also an
interaction between type and number of imaginings (top of Table 1). For actions that were imagined but not
performed in session 1, results were similar to novel actions (more false
responses with elaborate imagining, number of imaginings, and an interaction
between type and number of imaginings).
Unlike results for novel actions, there was no significant difference
between type of action. As number of
imaginings increased, the proportion of R judgments increased. Familiar actions got more R judgments as well. There was no significant effect of type of
imagination (bottom of Table 1).
Repeatedly imagining an
action increased the likelihood that participants would claim to have performed
that action. Elaboration leads to
greater inflation, but simple imagining also causes inflation for both bizarre
and familiar actions. Elaboration could
cause greater inflation because it forces rehearsal of sensory information from
several modalites, so it comes to mind in a richer, less distinguishable form.
Experiment 2
This experiment sought to
determine whether repetition of the key object (as in elaborate imagining)
drove inflation. In the individual
imaging condition, participants were read one sentence and then had to develop
their own imagination script that involved two sensory modalities. Sessions 1 and 3 were identical to those of
Experiment 1, but session 2, participants either elaborately imagined,
individually imagined, or read the action.
In the text condition, they were to read and reread the sentence for 15
second, then report the serial position of the longest word.
For novel actions, both
elaborate and individual imaginings led to more false responses, as did
repeated presentation. Type of activity
was significant. Text presentation
resulted in less false responses than individual imagining, which were less
than elaborate imagining. There was an
interaction between type and number of presentations. R judgments increased as number of presentations increased and
also with type of presentation. (top of Table 2). For imagined actions, type and number of session 2 presentations
led to more false responses. Individual
imagination led to the most false responses, followed by elaborate, then by
text presentation. There was an
interaction between type and number of presentations. Imagination led to more R responses than text presentation
(bottom of Table 2).
Although repeating objects in
the elaborate imagination condition may have increased familiarity, it did not
account for inflation, because those who imagined individually were even more
likely to elicit false responses. In the
text condition, no elaboration did produce a small amount of false
responses. This suggests that global
familiarity could account for some, but not all, of imagination inflation.
General Discussion
Greater sensory imagination
did result in more false memories. When
sensory information was removed from activity, there was minimal inflation,
suggesting that false memories come from misattributed discrete memory
characteristics. The way actions are
processed seems more important than simple repetition of actions.
There was a significant
number of false imagined responses in both experiments. The researchers suggest that by imagining
actions in session 2, cognitive associations with the actions increased, so
participants thought they had imagined them in session 1. They may have also been biased into saying
more actions were imagined because imagining was so frequent, but this is
weakened in that actions performed in session 1 were rarely classified as
imagined.
More elaborate imagining led
to more R judgments, supporting the theory that the more vivid and distinct an
action is, the greater the likelihood that it is a true memory. These results are consistent with the source
monitoring account of false memories.