Loftus (2004) pointed out several limitations with Anderson et al. ’ s fi ndings.
First, there are numerous theoretical and methodological problems with brain
imaging techniques. For example, do increases in ‘ metabolic ’ activity (e.g.
Recovered Memories
185
blood fl ow) to an area of the brain necessarily indicate an increase in ‘ cogni-
tive ’ activity? Do decreases in metabolic activity indicate decreases in ‘ cognitive
activity ’ ? At what point, statistically, do we conclude that one part of the brain
has become ‘ more ’ or ‘ less ’ active? By creating an ‘ average ’ brain scan from
the scans of different participants, are we running the risk of masking important
individual differences in brain structure? Can we be sure that the brain areas
identifi ed are facilitating the processes under investigation (e.g. suppression),
or could they be inhibiting another response? Brain imaging research is still
in its infancy and many of these questions are a long way from being resolved.
Thus any data derived from brain imaging studies need to be treated with
caution (for detailed critiques, see Kagan, 2007 ; Uttal, 2001 ; see also Vul,
Harris, Winkielman & Pashler, in press).
Secondly, and irrespective of methodological issues to do with brain scans,
the effects of the T/NT paradigm appear to be quite fragile. As Garry & Loftus
(2004) noted, the degree of suppression in the Anderson et al. experiments
was not particularly severe – instructing participants to ‘ forget ’ the target word
(e.g. roach) led to a 10% reduction in recall. This meant that participants still
recalled about 80% of the target words. Importantly, another group of psy-
chologists have failed to replicate these fi ndings in three separate experiments
(Bulevich, Roediger, Balota & Butler, 2006 ; Wade, 2007 ). Such a fragile effect
is not convincing evidence that human being can block, consciously or uncon-
sciously, entire autobiographical episodes from their memory.
Finally, the nature of the stimulus material used in these studies of suppres-
sion does not allow us to generalize to the kinds of traumas which allegedly
results in repression or dissociation . Freudian repression and dissociative amnesia
allegedly result in the blocking from awareness of traumatic, threatening and
emotional information. As Garry & Loftus (2004) argued, word pairs (ordeal –
roach) hardly mirror the impact of this kind of material. This last criticism was
recently addressed by Depue, Banich & Curran (2006) in a replication and
extension of Anderson et al. ’ s work. They examined whether the suppression
effects found for word pairs would be replicated for more emotional material.
Thus, rather than using words as both targets and cues, Depue et al. used faces
as cues and either emotionally neutral or negative words, or pictures, as targets.
Participants fi rst practised recalling 40 face – word or face – picture pairs until
they could recall them with a high level of accuracy (97%). They then took
part in an experimental phase where they were shown 32 of the face cues.
Sixteen of these face cues were paired with an instruction to ‘ think about ’ the
associated word or picture targets, whilst the other 16 were paired with an
instruction to ‘ not think about ’ the associated word or picture targets. For
half of the face cues, these ‘ think ’ or ‘ no think ’ instructions were repeated fi ve
times and for the other half they were repeated ten times.
Depue and colleagues found that participants who were instructed to ‘ think ’
about the targets ten times recalled more of those targets in a fi nal test than
participants who were given the ‘ no think ’ instructions. Importantly, the ‘ no
think ’ instructions led participants to recall fewer word or picture targets com-
186
Handbook of Psychology of Investigative Interviewing
pared to baseline word or picture targets for which they had been given no
instructions. The emotional nature of the stimuli also seemed to magnify the
effect. Participants recalled more of the emotional word or picture cues after
ten ‘ think ’ instructions than they did of the neutral word or picture cues.
Similarly, participants recalled fewer of the emotional, compared to neutral,
word or picture cues after the
‘ no think ’ instructions. Thus, according to
Depue and colleagues, there is a ‘ cognitive control ’ process in the brain that
deals differently with emotional and non - emotional memories. When emo-
tional material is repeatedly processed (or thought about) it becomes more
accessible than neutral material, but when emotional material is repeatedly
suppressed (not thought about) it becomes less accessible. In a follow - up f MRI
study, the authors also found evidence of two neural mechanisms which appear
to be implicated in the suppression process (Depue, Curran & Banich, 2007 ).
So does this body of research provide evidence that people can suppress
(avoid thinking about) unpleasant or traumatic events? The stark answer is no.
The stimuli used in the Depue et al. studies represent an important method-
ological improvement over those used by Anderson and colleagues but still do
not provide a good analogue for the experiences of people who have been
sexually abused. The more serious problem with this body of research,
however, is that it is cited as supporting evidence for the notion that people
repress (i.e. are unable to remember) traumatic events. This is evident in the
opening lines of Depue et al. (2007) . The authors state that whilst there is
evidence that people actively try to suppress memories,
‘ others claim that
memory repression or suppression is a clinical myth in search of scientifi c
support ’ ( ibid. : 215).
This sleight of hand, in which suppression and repression were confl ated, is
problematic – the two terms are not interchangeable. Suppression refers to
cases where people actively try not to think about something, usually with very
limited degrees of success (Anderson & Green, 2001 ; Depue et al. , 2007 ).
We ’ ve all had the experience of cringing and trying to distract ourselves when
a memory for an embarrassing event suddenly comes to mind. Most of us try,
often with limited success, not to think about events that upset us and psy-
chological research shows that it is not a particularly effective strategy. This is
mainly because the rule ( ‘ I must try not to think about X ’ ) contains the thing
one is trying to forget. Thus, most of us cannot help but picture a white bear
when explicitly instructed not to (Wegner, Schneider, Knutson & McMahon,
1991 ). Thus it comes as little surprise that, even when explicitly instructed to
‘ forget ’ , most participants in the Anderson studies showed only a 10% reduc-
tion in recall. These are typical fi ndings in studies of suppression . Repression ,
however, is when an individual is allegedly unable to remember something
because the mind has unconsciously blocked out any memory of the event.
Years of psychological research have indicated that this is indeed a ‘ clinical
myth ’ (Kihlstrom, 2002 ; H
ayne et al. , 2006 ; see also Loftus & Guyer, 2002a;
2002b , for a discussion of the case of Jane Doe in which an allegedly ‘ repressed ’
memory was ‘ recovered ’ during a videotaped interview). If participants in the
Recovered Memories
187
Anderson or Depue research have diffi culty in suppressing non
- traumatic
words and pictures, it suggests that consciously or unconsciously ‘ forgetting ’
traumatic autobiographical events would be a far more diffi cult task (see
McNally, 2003 : 152).
Unfortunately, media reports of the Anderson and Depue work have pro-
vided the
‘ take home
’ message that the latest advances in technology are
showing that people can block out memories of traumatic events (Highfi eld,
2007 ; Mundell, 2007 ). Most non - psychologists are, understandably, not alert
to the critical distinction between suppression and repression . This has meant
that the research fi ndings concerning the former were sometimes interpreted,
or presented, as evidence of the latter. One result of such misinterpretation is
that it bolsters the belief that individuals are capable of unconsciously forget-
ting traumatic events. Individuals (therapists and their clients) may begin
searching for evidence of such memories with the a priori , but misguided,
conviction that they must be there to be discovered. One self - help book, for
example, paradoxically claimed that the absence of a memory of abuse was
evidence that a person had been abused (Blume,
1990 ). As psychological
research has shown, this runs the risk of creating false beliefs or memories of
events that never occurred. How and why such false memories and beliefs arise
is the focus of the next section of this chapter.
Four m ain ‘ f alse m emory ’ m ethods
There are four main methods that psychologists have used to examine the
circumstances under which individuals might come to report events, or details
of events, that they did not experience:
• the DRM method;
• the misinformation method;
• the ‘ crashing memories ’ method;
• the parental misinformation method.
As will be seen, there are strengths and weaknesses with each method. This
means that these methods do not speak equally to the question of whether an
individual can come to report that they remember an entirely false, autobio-
graphical, emotionally
- charged childhood event (see Smeets, Merckelbach,
Horselenberg & Jelicic, 2005 ; Pezdek & Lam, 2007 ; Wade et al. , 2007 ).
There is also a further problematic distinction concerning whether these
methods are tapping into, or changing, memories of past events, beliefs about
past events, confi dence about whether a past event occurred or not or simply
reports about past events (Smeets et al. , 2005 ; see also Ost, 2003 ; Loftus &
Bernstein, 2005 ). Furthermore, whilst it has been argued that a false belief
is an important and necessary precursor to developing a false
memory
(Gudjonsson, 2003 ), it does not follow that a false belief will always lead
188
Handbook of Psychology of Investigative Interviewing
to a false memory (Ost,
2003 ). Therefore, some caution is warranted in
interpreting these fi ndings as a whole. Nevertheless, all four methods are cited
in the literature as providing evidence that ‘ false memories ’ can occur. With
these important caveats in mind, each method will now be critically
examined.
The DRM m ethod
Roediger
& McDermott
(1995) adapted a method previously developed
by Deese (1959) , referred to as the DRM method. In a typical study, partici-
pants are asked to remember a list of words, such as mad, fear, hate,
rage, temper, fury, ire, wrath, happy, fi ght, hatred, mean, calm, emotion,
enrage. Some time later, participants are recalled for a ‘ surprise ’ memory test
and are asked to indicate whether the word
‘ anger ’ was contained in the
original list. Many of them frequently report remembering the critical non -
presented word (e.g.
‘ anger ’ ) as having been present in the original list.
Several studies have successfully replicated this effect using various different
experimental manipulations (e.g., Gallo, Roberts & Seamon, 1997 ; Brainerd
& Reyna, 1998 ; Smith & Hunt, 1998 ). Roediger & McDermott ( 1995 : 803)
claim that the results of these studies ‘ reveal a powerful illusion of memory:
people remember events that never happened ’ , although there appears to be
some confusion about what the term ‘ event ’ actually refers to in psychological
research (see Freyd & Gleaves, 1996 , and the reply by Roediger & McDermott,
1996 ).
There also appear to be individual differences that leave certain participants
more likely to succumb to the DRM effect. For example, Winograd, Peluso
& Glover (1998) found that participants who scored higher on measures of
dissociation (high scorers have a tendency to experience problems in the inte-
gration of thoughts and feelings) and of vividness of mental imagery (high
scorers report having more vivid imaginative abilities) were more likely to
claim to remember the critical non - presented words. However, as Freyd &
Gleaves (1996) note, there are important differences between misremember-
ing words that have not been presented in a list, and misremembering an
otherwise happy childhood as being abusive. Similarly, Wilkinson & Hyman
(1998) demonstrated that in laboratory experiments there are important dif-
ferences between participants ’ performance on word list tasks and their per-
formance on autobiographical memory tasks. They found that self - reported
dissociative tendencies were related to errors on both the word list and auto-
biographical memory tasks, but that self - reported vividness of mental imagery
was only related to errors on the word list (DRM) task. Wilkinson & Hyman
(1998) argue that this is because remembering words and remembering auto-
biographical events rely on different underlying psychological processes. They
argue that it is therefore unwise to assume that participants who are suscep-
tible to the DRM are also more vulnerable to developing false autobiographi-
cal memories.
Recovered Memories
189
The m isinformation m ethod
The second method that psychologists have used examines whether subtle
changes in the way in which questions are asked about an event can change
what participants subsequently claim to remember about that event. The classic
studies of the misinformation effect, conducted by Loftus and colleagues,
examined whether misleading post - event information could alter eyewitnesses ’
memories of events they had recently witnessed (Loftus, 1979 ). In two studies
by Loftus & Palmer (1974) participants were shown a fi lm of an accident
involving two cars. Participants were then questioned to fi nd out how much
they could remember about the event. The question ‘ About how fast were the
cars going when they smashed i
nto each other ’ elicited higher estimates of
speed than questions in which the verbs collided , bumped , contacted or hit were
used (see also Loftus, Miller & Burns, 1978 ). This is a robust and easily rep-
licated effect, although there is still disagreement as to the mechanisms that
cause the misinformation effect (e.g., Bekerian & Bowers, 1983 ; McCloskey
& Zaragoza, 1985 ; Zaragoza, McCloskey & Jamis, 1987 ; Weingardt, Loftus
& Lindsay, 1995 ). Nevertheless it seems that subtle changes in wording can
lead participants to report non - existent details of events that they have expe-
rienced (see also Nourkova, Bernstein & Loftus, 2004 ). However, it is also
possible that such subtle changes in wording can lead individuals to report
that they remember entire events that they did not witness.
The ‘ c rashing m emories ’ m ethod
Crombag, Wagenaar & van Koppen (1996) asked participants whether they
remembered seeing a fi lm of a plane crashing into a block of fl ats in Amsterdam.
No fi lm of the crash existed so we can be sure that any participants who claimed
to have seen it must be mistaken. Nevertheless, 55 – 66% of the respondents to
Crombag et al. ’ s questionnaire claimed to have seen such a fi lm. Participants
were so convinced that they had seen it that they were willing to give details,
such as the angle at which the plane hit the building, how long before fi re
broke out and how long it was before the emergency services arrived. Crombag
et al. (1996) suggest that perhaps the vivid and emotionally - charged nature
of the event led people to think about the event, picture it in their heads and
then subsequently come to confuse the resulting imaginings as being real
memories. This is referred to as a ‘ source monitoring ’ error and occurs when
we misremember the source of information (Johnson, Hashtroudi & Lindsay,
1993 ).
Ost, Vrij, Costall & Bull (2002) replicated this effect using a different event
– the car crash in Paris in which Diana, Princess of Wales, Dodi Fayed and the
driver, Henri Paul, were killed. Whilst no fi lm of the crash has ever been
broadcast, there had been reports in the press that Diana ’ s car was being
pursued by paparazzi on motorbikes who were allegedly fi lming the chase. Ost
et al. (2002) therefore asked participants whether they had seen the paparazzis ’
Handbook of Psychology of Investigative Interviewing: Current Developments and Future Directions Page 36
