disagreement and the subsequent marriage. Furthermore, subjects who incor-
rectly recalled minimal disagreement between Bob and Margie were every bit as
confident of their mistaken recollections as they were of their accurate recol-
lections about other aspects of the story.
These results make sense if we assume that subjects did not activate a memory
record of the story, but instead used their belief that successful engagements
require agreement about whether to have children, in order to reconstruct the
story. If Bob and Margie are still married, then it would have seemed that any
disagreement about children must not have been very serious.
An intriguing implication of a reconstructionist approach to memory is that it
ought to be possible to create false memories—that is, memories of events that
never happened. Some researchers have suggested that some memories of sex-
ual abuse are actually false memories created by psychotherapeutic practices
that encourage clients to interpret certain psychological symptoms as evidence
of past abuse.
Eyewitness Memory and Reconstruction Reconstruction has been studied exten-
sively in the context of eyewitness memory. A variety of research has shown
that eyewitnesses tend to distort their memories of crimes and accidents based
on information they receive after the crime or accident.
For example, eyewitness memory research demonstrates what is calledphoto
bias. In one experiment on photo bias, subjects were first shown a film of a
crime and were later presented photographs of suspects. Later still the subjects
were required to pick the actual perpetrator out of a lineup. What happened
was that subjects tend to be biased towards identifying as the perpetrator any
suspect whose photograph they had recently seen, even when the person was
innocent of the crime (Brown, Deffenbacher, & Sturgill, 1977). Apparently,
when the subjects viewed the lineup, they recognized that they had seen one of
the suspects before, and erroneously assumed that it must be because the sus-
pect was the criminal.
Elizabeth Loftus, one of the most influential advocates of a reconstructionst
approach to memory, has conducted a variety of experiments in which subjects
are shown a film of an accident and are later asked questions about the film
(Loftus, 1979; Loftus, Miller, & Burns, 1978; Loftus & Loftus, 1980; Loftus &
Palmer, 1974). In one experiment, she asked one group of subjects leading
questions like ‘‘Did another car pass the red Datsun while it was stopped at the
stop sign?’’ when, in fact, the Datsun was stopped at a yield sign. (This partic-
ular experiment was conducted before Datsun changed its name to Nissan.)
When questioned again about the film, these subjects were much more likely to
claim they saw the Datsun stop at a stop sign than another group of subjects
not initially asked the misleading question. In some cases, memory was tested
by showing subjects two slides, a slide of a Datsun stopped at a stop sign and a
slide of the Datsun stopped at a yield sign. Most of the misled subjects selected
the slide displaying a stop sign, even when the misled subjects were offered a
substantial reward ($25) for remembering accurately. Incidentally, this experi-
mental paradigm usually contains a whole set of questions about various
details of the accident or crime. I am illustrating the paradigm with only one of
the questions that might be used. At any rate, the subjects were presumably
334 R. Kim Guenther