Content uploaded by Juliet Cohen
All content in this area was uploaded by Juliet Cohen on Aug 16, 2017
Content may be subject to copyright.
Questions of Credibility:
Omissions, Discrepancies and Errors of
Recall in the Testimony of
The issue of credibility is frequently raised in refusal notices and appeal
determinations dealing with the thousands of asylum seekers arriving in the
United Kingdom. The diﬀerences between Immigration and Nationality
Department (IND) interviews, statements of claim and later statements
(if given) are commonly used as a basis for denial of credibility and
dismissal of claim. The evidence reviewed below challenges the validity
of using these diﬀerences as grounds for denying credibility.
In the publication ‘Still No Reason At All’, produced by Asylum Aid,
many examples of such denial are quoted.
‘You claimed that your husband was taken by soldiers but in his asylum claim
he claims to have been taken by police.’
A Kurdish man who was questioned as to why he had told the interviewing
oﬃcer that he left his country in June when he later said it was July — ‘I don’t
know, but I do know it was Summer.’
A Home Oﬃce refusal letter states: ‘in the event a well-prepared statement seven
months after the asylum interview has little weight on his claim. Had Mr Z a
genuine fear of persecution he would have said so in his (ﬁrst) interview.’
Oﬃcials also tend to be sceptical about incidents described in later
interviews of which no mention was made in the ﬁrst. In one such
∗MA, MB, BS, DipRACOG, MRCGP; Medical Foundation for the Care of Victims of Torture.
This article was ﬁrst published in the Medico-Legal Journal, volume 69, Part 1, 2001, and is republished
with kind permission of the Editors.
‘Still no reason at all — Home Oﬃce decisions on asylum claims,’ Asylum Aid. Asylum Aid,
International Journal of Refugee Law Vol. 13 No. 3
Oxford University Press 2002. All rights reserved
example where two rapes were later disclosed, the Home Oﬃce response
was: ‘the late inclusion of such information is entirely at variance to your
previous interview and thus as a result reﬂects very unfavourably on your
veracity and credibility of your later statements’.
However, paragraph 199 of the UNHCR Handbook on Criteria for the
Determination of Refugee Status reminds decision-makers that ‘it may be
necessary for an examiner to clarify any apparent inconsistencies and to
resolve any contradictions in a further interview and ﬁnd an explanation
for any misrepresentation or concealment of facts . . .’ Clearly this advice
is not always followed.
To test the assumption that memories are detailed, accurate and
consistent across successive reports this review examines the reliability of
ordinary people’s memory for autobiographical details. In addition, it
evaluates the particular medical and psychological conditions potentially
inﬂuencing memory from which asylum seekers may suﬀer. It will be
shown that various conditions aﬀect the accuracy of recall.
Recent research on memory, especially in the context of witness
statements and interview techniques, is highly relevant to this issue. The
present review also examines the evidence for the eﬀects on memory of
the following factors: weight loss/malnutrition, minor traumatic brain
injury, raised stress hormone levels, post traumatic stress disorder, sleep
deprivation, depression, and pain. In the light of these studies the
assumption that discrepancies and omissions undermine credibility cannot
be justiﬁed. It is argued that there are alternative explanations for these
diﬀerences that are at least equally likely and which must be ruled out
before testimony is disbelieved.
Current research on normal memory
We cannot observe the actual physical act of remembering but only
indirectly test its eﬃciency. Hypotheses about memory have been widely
tested and a number of diﬀerent models proposed to explain observed
phenomena of everyday memory. Certain core observations can be
explained reliably without exceptions. Around these are more contentious
areas, where speciﬁc variations and exceptions to particular models
can be demonstrated under experimental conditions. Much research is
concerned with identifying the factors that cause such variations. Both
the general principles of normal memory function, and the speciﬁc factors
that produce variability, are relevant when the accuracy of asylum seekers’
recall is being assessed.
Case 92, ‘Reviewing the asylum determination procedure — Part 1’, Refugee Legal Centre,
Questions of Credibility 295
Short term memory is thought to be able to store about seven items
for a few seconds only, until new incoming information displaces the old.
If the information does not then move into long term store it is lost.
Recall of long term memory depends on retention and retrieval. Memory
tends to deteriorate and so become less accurate with time. This is known
as the retention interval. As well as being retained, memories must be
able to be retrieved. Tulving (1972) suggested that at a given time only
a small proportion of all memories are available for retrieval.
For long term memory, visual, verbal and auditory information is
thought to be coded by meaning, and then linked to related information
and associations. Consequently what is recorded is not an accurate copy
of the data but an interpretation. What we remember is inﬂuenced by
what we already know. Details tend to be lost over time and become
generalised, sometimes merging with similar memories. Repeated
childhood holidays to the same beach will result in blurred and blended
memories. How do we then try to remember more about a particular
incident? A further level of processing is proposed in which longer
lasting memory is achieved by attaching meaning and signiﬁcance to the
information. If little is attached, recall will be less easy. In our holiday
example, we can recall the year in which the dog was lost on the beach
by attaching other memories to that year such as the age of the dog, the
people present at the incident, the emotions experienced, and so on.
Memory is also inevitably inﬂuenced by higher cognitive interactions
with personality, mood and the perceived intentions of the interviewer.
Bartlett in 1932 introduced the idea of ‘schemata’ to explain the
observation that when people remember stories the recall is not accurate,
but people typically omit some details, and reconstruct the story in the
light of their own experience and knowledge.
He proposed that the story
is stored in memory in a pre-formed schema based on prior knowledge.
Recent research endorses this observation. In one study by List in 1986,
subjects were asked to view a video with eight diﬀerent acts of shoplifting.
The acts that were rated as highly probable were remembered better
than those rated as less probable. Subjects also falsely ‘remembered’
some events that were highly probable but had not actually occurred.
Particularly with repeated experiences, information speciﬁc to one episode
tends to drop out while information common to other similar episodes is
incorporated into the general schema and retained. A kind of blended
memory is formed. It is thought that information is not stored in distinct
compartments and does not remain inert but is dynamic. However, if
Tulving, E., (1972), Episodic and Semantic Memory, New York Academic Press, 381–403.
Bartlett, F. C., (1932), Remembering, Cambridge: Cambridge University Press.
List, J. A., (1986), ‘Age and schematic diﬀerences in reliability of eyewitness testimony’, Development
the information is particularly unusual, distinctive or emotional in relation
to the general experience, it may be retained.
McIntyre and Craik (1987) showed memory for facts is better than
that for the source of those facts, so people retain the information but
are unable to say how they know it or where it came from.
showed that memory for dates and times is notoriously unreliable, probably
because there are fewer links for this kind of information to other
knowledge. Yet date errors have been used to undermine credibility of
asylum seekers’ testimony, as in the example cited in the introduction.
Remembering and forgetting
Storage failure describes the case where the memory cannot be retrieved
and it is lost. Retrieval failure suggests that ﬁnding the right cues and
hints can result in successful recall. This is also known as cue-dependent
forgetting. ‘Blocks’ may persist for long periods to even trivial information.
‘Pop-up’ recall may occur later, spontaneously or in response to a diﬀerent
Free recall is where open questions are asked and no cues given. In
cued recall closed questions containing suggestions as to the target
information are used. This may cause problems in that it may aﬀect the
accuracy of the recall, provoking falsely ‘remembered’ details. On the
other hand it may also trigger far better and more detailed recall than
by open questions.
It has been shown that closed questions may cause shifting responses
under repeated questioning of child witnesses, while open-ended questions
do not impair accuracy. Gisli Gudjonnsen (1992) suggested that although
cued recall after free recall can elicit more full testimony, cues may
inﬂuence the recall and be misleading, amounting to post-event
To distinguish between such real/perceived memory and
suggested/confounding memory, Gudjonnsen recommends asking further
questions. ‘Real’ memories contain more sensory information such as
colours, size, shape and sound. ‘Suggested’ memories tended to be long
winded but lacking in vividness. These observations are further explored
in the paper by Schooler, Gerhard and Loftus (1986).
that ‘real’ memories contain more sensory and geographical detail and
are expressed with greater conﬁdence. ‘Suggested’ memories are described
with more words, verbal ‘hedges’, justiﬁcations, rationalisations and
descriptions of function rather than actuality.
McIntyre, J. S. & F. I. M. Craik, (1987), ‘Adult age diﬀerences for item and source information’,
Canadian Journal of Psychology, 41: 175–92.
Gudjonnsen, Gisli, (1992), The Psychology of Interrogations, Confessions and Testimony, Chichester: John
Wiley and Son.
Schooler, J. W., D. Gerhard & E. F. Loftus, (1986) ‘Qualities of the Unreal’, Journal of Experimental
Psychology, Learning, Memory and Cognition, Apr, 12 (2): 171–81.
Questions of Credibility 297
The eﬀectiveness of cues in aiding recall has been used by the police
in the cognitive interview technique in which witnesses are encouraged
to remember as much detail as they can about an event, no matter how
irrelevant, as any detail may trigger further recall of more relevant
information. One of the obvious diﬀerences between IND interview
technique and that of immigration law solicitors, as can be seen by
transcripts of the interviews, is just such a diﬀerence in the relative use
of free and cued recall. In initial immigration department interviews
asylum seekers are invited to answer mainly closed questions with brief
details. In later statements to their solicitor, questions are more often
open and asylum seekers are encouraged to give as much detail as
Hypermnesia — Remembering more
Hypermnesia describes the observation that people remember more
details with repeated recalls. In 1987 Payne showed that this is a reliable
phenomenon even when the time between recalls of word lists is varied
or the nature of the material to be remembered is varied.
He also showed
that it is more common when subjects are asked to recall high imagery
material than low imagery material. Pictorial material produced
hypermnesia in 95 per cent of cases compared to verbal material in 50
per cent. This is thought to be because the more elaborate or complex
material can give rise to greater numbers of recall cues which then
increase the chances of recall over time. Black, Levine and Laulhere
(1999) demonstrated the phenomenon of hypermnesia in autobiographical
This occurs when individuals are seen to recall more
information over repeated sessions even after they thought they could recall
nothing further. Other workers have shown that personal autobiographical
memories are highly imaginatively recorded. Over time and repeated
recall, there may be a tendency to confabulate and produce more false
responses. Even when the material to be recalled is a videotape, for
example, of a crime,
there is an increase in error rate with repetition.
In Black, Levine and Laulhere’s study, the memory tested was for the
verdict of the O.J. Simpson trial.
The numbers of errors increased in
successive recalls cumulatively, although the ratio of errors to accurate
information did not change over time. This means that the increased
information recalled in subsequent interviews was not due to an increased
Payne, D. G., (1987), ‘Hypermnesia and Reminiscence in Recall: A Historical and Empirical
Review’, Psychological Bulletin, 1015–27.
Black, S., L. J. Levine & T. M. Laulhere, (1999), ‘Autobiographical Remembering and
Hypermnesia: A Comparison of Older and Younger Adults’, Psychology and Ageing, 14 (4): 671–82.
Scrivner, E. & M. A. Safer, (1988), ‘Eyewitnesses show hypermnesia for details about a violent
event’, Journal of Applied Psychology, 73: 1–77.
Above, n. 10.
error rate, and confabulation was not the reason for the hypermnesia. In
three interviews conducted within one hour the information recalled
increased between the ﬁrst and second interviews. Between the second
and the third, although no new information was recalled, previously-
recalled information was ‘forgotten’ or omitted, so no overall increase
was shown. Their interpretation of this result is that autobiographical
memories are not traces that are retrieved and described, but are
reconstructed from event-speciﬁc knowledge. The exact form is guided
by the social and situational context in which they are recalled. Thus no
two reformulations can be identical.
Reproductive versus reconstructive memory
Memories that remain exactly the same each time they are recalled
appear to be reﬂecting a reproductive mechanism but memories that
vary are more likely to be generated by a reconstructive process. As long
ago as 1932 Bartlett observed that retold stories change with each
and more recently, a study by Anderson, Cohen and Taylor
(2000) conﬁrmed the variability of autobiographical memory.
examined successive recalls of personal memories by older and young
adults and found that older adults’ memories had greater stability. The
memories of younger adults varied more in both content and output
order. It was also found that recent memories varied more than older
ones. This suggests a shift over time from dynamic reconstruction toward
a reproductive mechanism, whereby a memory becomes more ﬁxed after
a long time has elapsed. In both age groups the second recall of a memory
produced an elaboration of the original version with less than 50 per
cent of the facts being identical and much new detail being added. There
were few verbatim repetitions, and diﬀerences in phrasing suggested that
the recall is reconstructed from a non-verbal store. The fact that such
marked variability occurs in the recall of everyday experiences that are
not traumatic indicates that it is misguided to expect the successive recalls
of asylum seekers to be perfectly consistent. Although it was formerly
considered that so-called ‘ﬂashbulb’ memories for dramatic events that
are highly important and emotionally charged remain ﬁxed, this view
has been challenged by recent research which has shown that these
memories also show variability.
Above, n. 4.
Anderson, S. J, G. Cohen & S. Taylor, (2000), ‘Rewriting the past: some factors aﬀecting the
variability of personal memories’, Applied Cognitive Psychology, 14: 435–54.
Christianson, S. A., (1989), ‘Flashbulb memories: special but not so special’, Memory and Cognition,
17 (4): 435–43; Neisser, U. & N. Harsch, (1992), ‘Phantom ﬂashbulbs: false recollections of hearing
the news about Challenger’, in Winograd, E. & U. Neisser, eds., Aﬀect and Accuracy in recall: Studies of
ﬂashbulb memories, Cambridge: Cambridge University Press.
Questions of Credibility 299
Anderson, Cohen and Taylor comment that there is a possible eﬀect
of ‘demand characteristics’ of the task.
When people are asked to repeat
information they have already given they usually assume that the ﬁrst
account is unsatisfactory in some way and may try to rectify this by
supplying more and diﬀerent details.
Tversky and Marsh (2000) showed
that when people retell events they take diﬀerent perspectives for diﬀerent
audiences and purposes.
These observations are directly relevant to the
diﬀerent settings in which asylum seekers give successive statements.
Eﬀects of the experience being recalled
Memory stability is known to be aﬀected by the nature of the event being
recalled and the level of associated emotion with it.
The accuracy of
recall of torture victims can be shown to be further inﬂuenced by a
number of special factors related to torture and its consequences. Studies
of victims of torture have established the most common symptoms suﬀered
to be: depression, anxiety, emotional lability, disturbed sleep, nightmares,
impaired memory and concentration, headache, cardiovascular
symptoms, dyspepsia, joint and muscle pain.
These are described by
clinicians as diagnoses of post traumatic stress disorder, sleep disorder,
depression, anxiety state, post-concussion syndrome, chronic pain state
and others. Other symptoms and conditions may be directly related to
an individual’s particular history such as of signiﬁcant weight loss —
discussed below — or a speciﬁc torture. For example repeated submersion
and other forms of suﬀocation may cause cerebral hypoxia leading to
loss of consciousness, confusion, disorientation and memory impairment.
These eﬀects may be transient or persistent depending on the extent of
hypoxic damage to the brain.
Emotional arousal and coping mechanisms
As Schactel (1947) deﬁned it: ‘Memory as a function of the living
personality can be understood as a capacity for the organisation and
reconstruction of past experiences and impressions in the service of present
Above, n. 14.
Edwards, D. & J. Potter, (1992), ‘The Chancellor’s Memory: Rhetoric and truth in discursive
remembering’, Applied Cognitive Psychology 6: 187–215.
Tversky, B. & E. J. Marsh, (2000), ‘Biased retellings of events yield biased memories’, Cognitive
Psychology, 40(1): 1–38.
Wynn, V. E. & R. H. Logie, (1998), ‘The Veracity of Long-term Memories — Did Bartlett
get it right?’, Applied Cognitive Psychology, 12: 1–20.
See, for example, Petersen, H. D. & P. Jacobsen, (1985), ‘Psychical and physical symptoms
after torture: A prospective, controlled study’, Forensic Science International, 29 (3–4): 179–89; Hougen,
H. P., J. Kelstrup, H. D. Petersen & O. V. Rasmussen, (1988), ‘Sequelae to torture: A controlled
study of torture victims living in exile’, Forensic Science International, 36 (1–2): 153–60.
Norfolk, G. A., (1999), ‘Physical illnesses and their potential inﬂuence’, in Heaton-Armstrong,
A., E. Shepherd & D. Wolchover, eds. Analysing Witness Testimony, Blackstone Press Ltd., London,
fears, needs and interests.’
There is ample evidence that memory is
aﬀected by the need to cope with emotional and traumatic experiences.
Allodi in 1991
and several others have demonstrated the upsetting
nature of torture recall and the eﬀect of this on testimony. Christianson
and Loftus (1991) showed that increased arousal during an event led to
a concentration on certain detail with reduced recall of peripheral detail.
Open-ended questions and free recall led to the greatest distress and
limited reporting while neutral cues including reading from a list of
possible events produced better recall. The diﬀerent eﬀects of open versus
closed questioning on recall in the general population have already
been noted. Obviously, the particular eﬀects are very dependent on the
circumstances of the interview, the time elapsed since the torture, and
the relationship with the interviewer. Mollica (1988) showed that the
interviewer’s own mental protective devices will be employed to resist the
negative eﬀects of hearing about upsetting events. Often there is a fear
held by the interviewer that recall will trigger uncontrollable mental
distress for the interviewee. This leads on to feelings of inadequacy in
comforting the person and voyeurism in ‘forcing’ them to relive traumatic
There may be further complications due to the interpreter if
one is present. If interpreters are also torture victims, or closely involved
with such victims, they may close oﬀcertain questions and answers or
give non-verbal cues discouraging elaboration of detail. The use of
checklists as an aid to free recall can be helpful in overcoming the barriers
of awkwardness and emotional distress, especially for experiences diﬃcult
to accept or verbalise, such as sexual assault.
Events encoded during high levels of arousal have been shown to be
more diﬃcult to retrieve, although they can be retrieved in time.
Weight loss and malnutrition
Sutker et al. (1986 and 1991) demonstrated that prolonged malnutrition
and weight loss can aﬀect memory function.
This work was mainly on
Schactel, E. G., (1947), ‘On memory and childhood amnesia’, Psychiatry, 10: 1–26.
Allodi, F., (1991), ‘Assessment and treatment of torture victims: a critical review’, Journal of
Nervous and Mental Disease, 170 (1): 4–11.
Christianson, S. A, E. F. Loftus, H. Hoﬀman & G. R. Loftus, (1991), ‘Eye ﬁxations and memory
for emotional events’, Journal of Experimental Psychology, Learning, Memory and Cognition,17(4): 693–70.
Mollica, R. F., (1988), ‘The trauma story: the psychiatric care of refugee survivors of violence
and torture’, in Ochberg, F. M., ed., Post traumatic therapy and victims of violence, Brunner/Mazel, New
Bradley, B. P. & A. D. Baddeley (1990), ‘Emotional factors in forgetting’, Psychological Medicine,
Sutker, P. B., D. K. Winstead, K. C. Goist, R. M. Malow, & A. N. Allain, (1986),
‘Psychopathology subtypes and symptom correlates among former prisoners of war’, Journal of
Psychopathology and Behavioural Assessment, 8:89–101; Sutker, P. B., D. K. Winstead, Z. H. Galina, &
A. N. Allain, (1991), ‘Cognitive deﬁcits and psychopathology among former prisoners of war and
combat veterans of the Korean conﬂict’, American Journal of Psychiatry, 148(1): 67–72.
Questions of Credibility 301
prisoners of war and holocaust survivors from the Second World War,
but in medical terms is congruent with established knowledge on vitamin
deﬁciency disorders, especially the B vitamins. In patients on total
intravenous nutrition multivitamins must be included or deﬁciency
syndromes may rapidly ensue.
A condition known as Wernicke’s
encephalopathy can follow severe thiamine deﬁciency and memory and
cognitive deﬁcits have been demonstrated in this condition, reversible
after treatment with thiamine. Elderly patients with low folic acid levels
had poor episodic recall.
At the other end of the life cycle, a randomised
controlled trial of treatment with micro-nutrient fortiﬁed biscuits carried
out with children from a poor rural area in South Africa,
distinct improvement after treatment, in both cognitive function and short
term memory. In torture victims subjected to prolonged detention a
history of available diet and estimations of weight loss would indicate the
possible presence of this eﬀect.
Minor traumatic brain injury
‘Minor traumatic brain injury’ describes head injury of the kind that does
not involve prolonged loss of consciousness but may nevertheless have
signiﬁcant eﬀects on health and in particular on cognitive function and
memory. It is generally established that more major head injury has
similar, though more serious eﬀects, but it is the consequences of minor
head injuries that have tended to be overlooked. Many victims of torture,
not surprisingly, are unable to clearly estimate periods of unconsciousness,
or to distinguish the cause between such other possibilities as vaso-vagal
inhibition (fainting) or suﬀocation. However, a detailed history should be
able to elicit rough estimates of severity and frequency of head injury,
prolonged or brief loss of consciousness and symptoms noted afterwards
attributable to head injury (post-concussion syndrome). Such symptoms
include dizziness, drowsiness, double vision, headache and nausea in the
short term and persisting headache, dizziness, poor concentration, poor
memory, fatigue, irritability, anxiety, noise sensitivity and insomnia in
the longer term.
Clear-cut examples of retrograde and post-traumatic amnesia have
been accepted as inﬂuencing ability to give testimony.
Where the history
Hahn, J. S., W. Berquist, D. M. Alcorn, L. Chamberlain & D. Bass, (1998), ‘Wernicke
encephalopathy and beriberi during total parental nutrition attributable to multivitamin infusion
shortage’, Paediatrics, 10(1): E10.
Hassing, L., A. Wahlin, B. Winblad, & L. Beckman, (1999), ‘Further evidence on the eﬀect of
vitamin B 12 and folate levels on episodic memory functioning: A population-based study of healthy
very old adults’, Biol. Psychiatry, 45(11): 1472–80.
Van Stuijvenberg, M. E, J. D. Kvalsvig, M. Faber, M. Kruger, D. G. Kenoyer & A. J. Benade,
(1999), ‘Eﬀect of iron-, iodine-, and beta-carotene fortiﬁed biscuits on the micronutrient status of
primary school children: a randomised controlled trial’, American Journal of Clinical Nutrition, 69 (3):
Above, n. 21.
is less clear cut, there may still be eﬀects on the brain from minor injury.
In 1999 Voller et al. published ﬁndings in a study of very minor traumatic
brain injury, deﬁned as loss of consciousness less than 20 minutes with a
normal score on neurological examination.
They found signiﬁcant
impairment of verbal memory persisted even after six weeks, together
with attention deﬁcits and poor information processing. On MRI scan
25 per cent had positive ﬁndings of traumatic lesions to the brain. In a
review of post-concussion syndrome, Evans (1992) described the organic
nature of the syndrome as being well-documented in ﬁndings in neuro-
pathology, neuro-physiology, neuro-imaging and neuro-psychological
The principal sequelae are headache, psychological and somatic
complaints and cognitive impairment. Most resolve within three months
of injury, but a minority persist for months or even years. Risk factors
identiﬁed for such persistence include age over 40, lower socio-economic
level, female sex, alcohol abuse, prior head injury and multiple trauma.
The latter two categories at least would therefore potentially include
victims of torture.
Stress, arousal and cortisol
In both human and animal experiments, glucocorticoids such as cortisol
have been shown to regulate hippocampal mechanism in the brain and
so aﬀect memory. This has also been observed in patients with Cushing’s
disease, in which excessive amounts of these hormones are produced
from the adrenal gland, and in patients requiring treatment with steroids
for conditions such as arthritis or asthma. Both are at risk of impaired
memory. Impaired memory and raised cortisol levels have also been
found in the elderly and in patients with depression. In an experiment
by Newcomer et al. (1999),
subjects were given four days treatment
with low dose cortisol, a glucocorticoid known to be produced when
under stress. Other subjects were given higher doses to simulate major
stress. The trial was conducted as a double blind, randomised, placebo
controlled study. Cognitive testing was done at intervals of day 0, day 1,
day 4 and day 10. Cortisol treatment at higher dose produced reversible
reduction in verbal declarative memory without eﬀects on non-verbal
memory or attention. The levels of cortisol given were based on those
detected in the blood of those undergoing surgery, which provokes a
physiological stress response. The authors conclude that these results are
Voller, B., T. Benke, K. Benedetto, P. Schnider, E. Auﬀ& F. Aichner, (1999),
‘Neuropsychological, MRI and EEG ﬁndings after very mild traumatic brain injury’, Brain Injury
Evans, R. W., (1992), ‘The post concussion syndrome and the sequelae of mild injury’, Neurol.
Clin. 10(4): 815–47; also, Gfeller, J. D., J. T. Chibnall & P. N. Duckro, (1994), ‘Post concussion
symptoms and cognitive functioning in post traumatic headache patients’, Headache, 34(9): 503–7.
Newcomer, J. W., G. Selke, A. K. Melson, et al., (1999), ‘Decreased memory performance in
healthy humans induced by stress-level cortisol treatment’, Arch. General Psychiatry 56: 527–33.
Questions of Credibility 303
directly relevant to the interpretation of decreased memory performance
in humans under periods of extended stress due to the eﬀect of raised
plasma cortisol on the memory encoding and retrieval processes. These
results were conﬁrmed in a similar experiment by de Quervain et al.
Post traumatic stress disorder
It has been known since at least as far back as the First World War that
battle experiences can cause episodes of memory loss. Pelmanism, a
system of memory training exercises, was used with shell shocked patients
to improve their memory and concentration. In 1889, a French doctor,
Janet, was writing about amnesia for part or all of traumatic experiences.
Post traumatic stress disorder was deﬁned after the Vietnam War, but it
essentially describes the symptoms that may develop in any victim or
witness of violent and terrifying traumatic experience. These symptoms
are characterised by distressing recall, nightmares, ﬂashbacks, avoidance
behaviour, sleep disorder, irritability, hyper-arousal and social withdrawal.
According to the criteria for the diagnosis of post traumatic stress disorder
they must persist for more than one month. Disturbances of memory and
concentration have been found in studies on prisoners of war from the
Second World War and the Korean War. Torrie in 1944 found that
immediately after a major campaign about 5 per cent of soldiers had no
memory at all of the events.
Other studies have shown dissociative
amnesia, which includes the inability to remember some aspects of the
trauma, occurs in large numbers of disaster victims: 29 per cent of
earthquake survivors, 57 per cent of ambush victims and 61 per cent of
Such dissociative processing complicates the capacity to communicate
the trauma. The memory may be wholly or partly organised on an
implicit or perceptual level, with no accompanying narrative about
what occurred. During provocation of traumatic memories under neuro-
imaging, an experiment showed decreased activation of Broca’s area, the
speech area of the brain. At the same time there was enhanced imaging
of the right hemisphere areas most associated with intense emotion and
In 1992, Bremner et al. reported lower hippocampal volume in patients
with combat-related post traumatic stress disorder than in matched
De Quervain, D., B. Roozendaal, R. Nitsch, J. McGaugh & C. Hock, (2000), ‘Acute cortisone
administration impairs retrieval of long-term declarative memory in humans’, Nature Neuroscience 3(4):
Torrie, A., (1944), ‘Psychosomatic casualties in the Middle East’, Lancet, 29: 139–43.
Rauch, S., B. A. van der Kolk, R. Fisher et al., (1994), ‘PET imagery-positron emission scans
of traumatic imagery in PTSD patients’; paper presented at the annual meeting of the International
Society for Traumatic Stress Studies, Chicago, IL.
Interestingly, recent research on London cab drivers who are
required to memorise all the streets of the city shows that they have
increased hippocampal volume. In 1993, Bremner et al. showed that
Vietnam veterans with post traumatic stress disorder had lower scores on
both immediate and delayed recall on memory testing.
This links with
the evidence cited above that severe stress induces a cortisol release that
has a neurotoxic eﬀect on the hippocampus, an important part of the
brain in memory storage mechanisms.
Numerous other studies illustrate the eﬀects of post traumatic stress
disorder on memory. For example, Yehuda et al. (1995) found that
veterans with post traumatic stress disorder had a quite circumscribed
cognitive deﬁcit aﬀecting memory retention.
Jenkins et al. (1998) studied
rape victims with post traumatic stress disorder and found they had recall
There is relatively less published work speciﬁcally on victims of torture,
but in two reviews of a series of patients’ symptoms, impaired memory
and poor concentration are speciﬁcally cited as amongst the most common
In this case no formal diagnosis of post traumatic stress
disorder was made, although the other psychological symptoms listed in
these studies are essentially those of post traumatic stress disorder:
disturbed sleep, nightmares, emotional lability, anxiety and depression.
In one study almost 50 per cent of London asylum seekers presenting to
the Medical Foundation for the Care of Victims of Torture were found
to have post traumatic stress disorder compared to an expected incidence
of 5–7 per cent in the normal population, using the internationally agreed
DSM-IV diagnostic criteria.
With speciﬁc reference to autobiographical memory, Harvey et al.
(1998) have studied both acute stress disorder, which can develop within
the ﬁrst month after a traumatic experience, and post traumatic stress
disorder, in which symptoms persist for longer than one month.
found that patients with acute stress disorder reported fewer speciﬁc
memories of the trauma than did non-acute stress disorder patients.
Bremner, J. D., J. P. Seibyl, T. M. Scott et al., (1992), ‘Decreased hippocampal volume in
PTSD’ in New Research Program and Abstracts, 145
Annual meeting of the American Psychiatric
Association, Washington DC.
Bremner, J. D., T. M. Scott, R. C. Delaney et al., (1995), ‘Deﬁcits in short term memory in
post traumatic stress disorder’, American Journal of Psychiatry, 150: 1015–19.
Yehuda, R., R. S. Keefe, P. D. Harvey, et al., (1995), ‘Learning and memory in combat
veterans with post traumatic stress disorder’, American Journal of Psychiatry, 152: 137–39.
Jenkins, M. A., P. J. Langlais, D. Delis & R. Cohen, (1998), ‘Learning and memory in rape
victims with post traumatic stress disorder’, American Journal of Psychiatry, 155: 278–79.
Above, n. 20.
Ramsay, R., C. Gorst-Unsworth & S. Turner, (1993), ‘Psychiatric morbidity in survivors or
organised state violence including torture’, British Journal of Psychiatry, 162: 55–59.
Harvey, A. G., R. A. Bryant & S. T. Deng, (1998), ‘Autobiographical memory in acute stress
disorder’, Journal of Consulting and Clinical Psychology, 66(3): 500–506.
Questions of Credibility 305
Depression was found to play a signiﬁcant role in the memory deﬁcits of
acute stress disorder patients, but when this was controlled for, some
eﬀect of acute stress disorder alone was still evident. Harvey also showed
that the presence of acute stress disorder was highly predictive of those
who would go on to develop post traumatic stress disorder. Seventy-eight
per cent of acute stress disorder patients had post traumatic stress disorder
at six months where the average expected number in the non-acute stress
disorder population is less than 30 per cent. The model postulated is thus
that high cortisol levels released at the time of maximum stress aﬀect the
organisation of memories leading to disrupted retrieval processes, reduced
optimal recall of the traumatic memories and possible unwanted excessive
recall in the form of ﬂashbacks, nightmares and persistent thoughts of
the trauma. Some therapeutic approaches to post traumatic stress disorder
reﬂect this model by working on the ‘processing’ of traumatic memories
in a way that reduces the associated distress and aims for integration of
the memories into ‘normal’ long-term storage.
Sleep deprivation is a common form of torture. In addition, as quoted
above, torture survivors often suﬀer from ongoing sleep disorder with
diﬃculty falling and staying asleep and frequent nightmares. This may
be part of a diagnosis of post traumatic stress disorder or may be present
without the rest of the syndrome. In a retrospective analysis of ﬁfty
patients seen for any clinical problem at the Medical Foundation for the
Care of Victims of Torture, the author found 75 per cent to be gaining
4 or less hours sleep in 24.
Studies in sleep-deprived subjects have
shown impaired cognition and recall. Harrison and Home (1992) showed
impaired facial recognition in sleep deprived subjects even when given
Idzidowski in 1984 showed sleep deprivation impairs long-term
memory and other workers have shown a link speciﬁcally with the lack
of REM phase sleep.
Depression may be part of the post traumatic stress disorder spectrum
or made as a separate diagnosis in its own right. In the words of
Dietrich (2000), ‘one of the most frequent and neuro-psychologically well
investigated symptoms in depression is reduced memory capacity’.
Recent conﬁrmation can be found in the work of Pelosi et al. (2000),
Cohen, J., unpublished report.
Harrison, Y. & J. A. Horne, (2000), ‘Sleep loss and temporal memory’, The Quarterly Journal of
Experimental Psychology, A 53(1): 271–279.
Idzidowski, C., (1984) ‘Sleep and memory’, British Journal of Psychology, 75(4): 439–449.
Dietrich, D.E., A. Kleinschmidt, U. Hauser et al., (2000), ‘Word recognition memory before
and after successful treatment of depression’, Pharmacopsychiatry, 33(3): 221–228.
who demonstrate that depressed patients had poor recall compared to
controls, and this became worse as the memory load increased.They
concluded that major depression signiﬁcantly aﬀects working memory.
Depressed patients with minor traumatic brain injury reported more
severe cognitive symptoms.
Autobiographical memory is known to be
aﬀected by depression.
In the study on London asylum seekers, 30 per cent were found to
have depression, compared to 5–10 per cent of the normal population.
This study used the internationally agreed diagnostic criteria of DSM-IV.
It can be very diﬃcult to separate the eﬀects of chronic pain from
depression as pain itself is such a potent trigger of depression. In addition,
pain patients often have very poor sleep. Iezzi et al. in 1999 studied the
neuro-cognitive performance of pain patients related to their emotional
They found that those highest in emotional distress experienced
most diﬃculty in intellectual function, delayed recall and problem solving.
Schnurr and MacDonald (1995) tried to exclude the eﬀects of depression
in their study of pain patients and found that diﬀerences in memory
complaint were still greater than in controls.
Although pain patients
often themselves attribute their memory problems to their use of codeine
and other strong analgesics, there was no evidence of this in their study.
In victims of torture there is a combined incidence of chronic pain from
musculo-skeletal injury, sleep disorder, depression and emotional distress
which would be very diﬃcult to study separately but clearly all of these
conditions can combine to produce similar eﬀects on memory.
Assessment and quantiﬁcation
How can all of the above conditions be quantiﬁed and documented? A
detailed history and examination by independent medical experts is the
simplest, and arguably the most important element. CT scans, bone
scintigraphy and other medical tests can provide further evidence for a
history of torture but are expensive and not necessarily conclusive. A
Pelosi. L., T. Slade, L. D. Blumhardt & V. K. Sharma, (2000), ‘Working memory dysfunction
in major depression; an event-related potential study’, Clinical Neurophysiology, 111(9); 1531–43.
Gfeller et al., above, n. 33.
Brittlebank, A. D., J. Scott, J. M. G. Williams & I. N. Ferrier, (1993), ‘Autobiographical memory
in depression: state or trait marker’, British Journal of Psychiatry, 162: 118–121; Kuyken, W. & C. R.
Brewin, (1995), ‘Autobiographical memory functioning in depression: reports of early abuse’, Journal
of Abnormal Psychology, 104: 585–591.
Ramsay, Gorst-Unsworth and Turner, above, n. 43.
Iezzi, T., Y. Archibald, P. Barnett, A. Klinck & M. Duckworth, (1999), ‘Neurocognitive
performance and emotional status in chronic pain patients’, Journal of Behavioural Medicine, 22(3):
Schnurr, R. F. & M. R. Macdonald, (1995), ‘Memory complaints in chronic pain’, Clinical
Journal of Pain, 11(2): 103–111.
Questions of Credibility 307
negative bone scintigraphy test does not mean no bone injury occurred.
Psychological assessments with batteries of questionnaires can give scores
for depression, anxiety, post traumatic stress disorder, short and long
term memory, trauma experience and chronic pain, but these tests are
by no means all internationally validated. There are also considerable
diﬃculties in performing these tests through an interpreter if all questions
are not available in translation, as is generally the case. There is a
subjective element to many such questionnaires and day-to-day variation
in responses can be signiﬁcant. It is diﬃcult with the current state of
knowledge to determine if such assessments are in eﬀect going to be more
useful than a general medical examination.
In assessing the credibility of asylum seekers what should we regard as
reasonable degrees of error or omission? In what numbers? Classes of
error may be categorised as: calendar errors, detail diﬀerences from one
period of detention to another similar one, errors of deﬁnition or
translation, for example, soldiers/police/men and numbers of men present
during torture, telescoping and expansion of time-frames, omissions of
rape and other deeply traumatic incidents. It is possible some of these
can be explained by the potential for variability of true memories.
The observation of a lack of supporting detail, especially sensory and
geographical, for example, describing cell, food, and hygiene
arrangements, may indicate unprepared answers to an unforeseen
question. However, it may also simply indicate limitations in the interview
technique. An important element often neglected in written evidence is
the presence of visual cues for the interviewer including changes of
expression, gesture, body language indicating emotion and re-enactment
of posture during torture. Documentation of secondary symptoms, for
example, post traumatic stress disorder, sleep disorder, gastritis, shortness
of breath, palpitations, headaches, chronic back and joint pain and skin
irritation which are all well-recognised in victims of torture is also too
The earlier sections have demonstrated just how unusual it is if recall
is accurately reproduced and how common diﬀerences in detail can be.
So is there any way in which variable statements can be said to aﬀect
credibility or should the legal system be altering its approach?
Let us look at the treatment of witnesses in court. Acceptance of their
credibility can be crucial in establishing guilt or innocence and yet these
judgements may be made on what is all too probably an unproven and
First comes the observation: ‘liars change their story’. This is supposedly
because a made-up story is harder to remember consistently than an
autobiographical event, or because when challenged, liars change details
to cover inconsistencies.
This leads to the hypothesis: ‘changes in a story indicate falsehood’, but this
is the converse of the observation and has never been conclusively proven
to be so. Just because cats like milk does not mean any creature drinking
milk must be a cat. In logic this is known as the ‘fallacy of converting
Current research on memory shows that stories can change for many
reasons and such changes do not necessarily indicate that the narrator is
lying. In the real world, we know that the most rigidly reproduced
accounts may be so because they have been memorised from a script.
Conversely, those with certain discrepancies may be so because they have
been genuinely reconstructed from autobiographical memories. Yet we
encourage consistency in all testimony because it ‘keeps it simple’.
Motivation to be consistent is only present if the subject ﬁrst knows that
consistency is valued above everything. If not, it is ‘accidental’ rather
than intended. In Britain we give witnesses their statements to read before
going into court, to ensure they are happy to swear to them on oath and
to make sure they do not then depart from the ‘established’ story.
Presumably this is based on the assumption that they are likely to do so.
This does not mean we are suggesting they lie, just that experience in the
courts has shown it is almost impossible to maintain absolute consistency,
especially if it is a long time since the events to be recalled. Yet this
latitude is not given to asylum seekers who are repeatedly judged and
found not credible on this very issue. This application of dual standards
There are strong grounds for arguing that lack of consistency per se
cannot be used to give any negative weight to the assessment of credibility.
In addition, it needs to be acknowledged that judgements about credibility
are extremely fallible. Schooler, Gerhard and Loftus (1986) tried to give
‘judges’ cues on diﬀerentiating ‘real’ from ‘suggested’ memories.
were able to improve their scores from 50 per cent to a mere 60 per cent
success rate. This clearly still leaves enormous scope for error in such
judgements. The ﬁndings of this review have wider implications for any
witness evidence presented in court. In the case of asylum seekers,
especially, it is clear that great caution needs to be exercised in denying
credibility. The normal variability of memory is likely to be exacerbated
by the medical factors reviewed above and a general impairment of recall
is to be expected as a result of their traumatic experiences and physical
and mental state.
Further research would be invaluable in quantifying the degree of
memory impairment suﬀered by asylum seekers with some of the medical
Above, n. 8.
Questions of Credibility 309
conditions reviewed above, and assessing the possible use of trauma scales
and other measurements.
On a practical level, standardising questions and formats of all interviews
would go some way to improving consistency. Increasing the detail of
medical histories with particular reference to the conditions discussed:
weight loss/malnutrition, head injury, post traumatic stress disorder, sleep
disorder, depression and chronic pain, would also aid in this diﬃcult task
of assessing credibility.
In eﬀect, however, this review concludes that credibility assessment by
the determination of accuracy and reproducibility of an asylum seekers’
recall is not a valid component of asylum decision making.