People have been found to misrepresent past events by combining features from two separate events to form a composite event that was never experienced. Known as a memory conjunction error, this type of memory error has been shown to occur for words, sentences and even landmarks from separate city maps.
The
present set of experiments used pictures of faces as stimulus materials
to illustrate the occurrence of memory conjunction errors with more ecologically
relevant materials. In the first experiment, participants viewed pictures
and then completed an immediate recognition test. In the second experiment
participants viewed faces and completed a recognition test following a
24-hour delay.
The
manipulation of delay was of interest because of the theoretical explanation
of memory conjunction errors that has been offered by Reinitz. Reinitz
assumes that perceptual features are easier to encode than the information
of how individual perceptual features relate to one another. Additionally,
the relational information is assumed not to persist in memory as long
as the perceptual features. Memory conjunction errors occur when the relational
information for the composition of a set of features is lost. When relation
information is lost, features from different events can be combined to
form composite events.
A second variable of interest was the temporal proximity of two faces. Recent research cited in this article suggests that the temporal proximity of events to one another influences the likelihood that features will be joined together from separate events to form a new composite event. A discrimination mechanism has been proposed to exist that allows for individual traces to be distinguished from one another to prevent memory conjunction errors. When a memory cue activates multiple traces these traces must be identified as separate in order to keep their component features from becoming entangled with one another. This mechanism distinguishes the separate traces based on when an event was encoded.
96
participants participated in this study.
Line
drawings of faces were used in the present study that contained two sets
of distinct features. There were 8 eye-nose sets and 8 hair-mouth sets,
which were completely crossed with one another to form 64 unique faces.
There
were four trials in which two faces were presented adjacent to one another
for 16 seconds. There was a
15-minute retention interval between the presentation of the faces and
the administration of a recognition test.
Recognition
Test
2 – Target Faces
2
– Simultaneous conjunction faces (features were intermixed from two faces
that were presented at the same time)
2
– Near conjunction faces (features from two faces that were presented in
adjacent pairs were intermixed to make a new composite face)
2
– Far conjunction faces (features were intermixed from faces from face
pairs that had an intervening face pair between their presentation)
Results
Rates
of acceptance of the different recognition items (remember there are only
two instances of each type – and as always 50 % accuracy could merely represent
guessing)
The
procedure was the same except that was a 24-hour delay between presentation
and recognition testing.
Results
After
a delay there was no difference between the mean acceptance rates for targets
and simultaneous conjunction faces. The results were as follows: target
= simultaneous conjunction faces > near-con = far con
Discussion
The
researchers view their results as suggesting that temporal markers allow
for different traces to be distinguished from one another. After a delay
the relational information is lost but not the temporal marker, which allows
for more temporally distinct traces to be distinguished from one another
and suppress rates of memory conjunction errors.
|
|
|
|
|
|
|
|