Content uploaded by Richard Loewenstein
Author content
All content in this area was uploaded by Richard Loewenstein on Jan 31, 2017
Content may be subject to copyright.
Content uploaded by Bethany L Brand
Author content
All content in this area was uploaded by Bethany L Brand on Feb 24, 2014
Content may be subject to copyright.
Content uploaded by Bethany L Brand
Author content
All content in this area was uploaded by Bethany L Brand
Content may be subject to copyright.
Evaluation of the Evidence for the Trauma and Fantasy Models of Dissociation
Constance J. Dalenberg
California School of Professional Psychology at Alliant
International University, San Diego
Bethany L. Brand
Towson University
David H. Gleaves and Martin J. Dorahy
University of Canterbury Richard J. Loewenstein
Sheppard Pratt Health System, Baltimore, Maryland, and
University of Maryland School of Medicine, Baltimore
Etzel Carden˜a
Lund University Paul A. Frewen
University of Western Ontario
Eve B. Carlson
National Center for Posttraumatic Stress Disorder, Menlo Park,
and Veterans Administration Palo Alto Health Care System,
Palo Alto, California
David Spiegel
Stanford University School of Medicine
The relationship between a reported history of trauma and dissociative symptoms has been explained in 2
conflicting ways. Pathological dissociation has been conceptualized as a response to antecedent traumatic
stress and/or severe psychological adversity. Others have proposed that dissociation makes individuals prone
to fantasy, thereby engendering confabulated memories of trauma. We examine data related to a series of 8
contrasting predictions based on the trauma model and the fantasy model of dissociation. In keeping with the
trauma model, the relationship between trauma and dissociation was consistent and moderate in strength, and
remained significant when objective measures of trauma were used. Dissociation was temporally related to
trauma and trauma treatment, and was predictive of trauma history when fantasy proneness was controlled.
Dissociation was not reliably associated with suggestibility, nor was there evidence for the fantasy model
prediction of greater inaccuracy of recovered memory. Instead, dissociation was positively related to a history
of trauma memory recovery and negatively related to the more general measures of narrative cohesion.
Research also supports the trauma theory of dissociation as a regulatory response to fear or other extreme
emotion with measurable biological correlates. We conclude, on the basis of evidence related to these 8
predictions, that there is strong empirical support for the hypothesis that trauma causes dissociation, and that
dissociation remains related to trauma history when fantasy proneness is controlled. We find little support for
the hypothesis that the dissociation–trauma relationship is due to fantasy proneness or confabulated memories
of trauma.
Keywords: trauma, dissociative disorder, dissociation, suggestibility, fantasy
Scientific interest in the concept of dissociation and the etiology
of the dissociative disorders has increased markedly in recent
decades. Building on the foundational work of Janet (1889, 1919),
researchers have empirically identified and investigated various
types and categories of dissociation: the identity alterations and
amnesias prominent in the dissociative disorders (Putnam, 1991),
This article was published Online First March 12, 2012.
Constance J. Dalenberg, Trauma Research Institute, California School of
Professional Psychology at Alliant International University, San Diego;
Bethany L. Brand, Department of Psychology, Towson University; David
H. Gleaves and Martin J. Dorahy, Department of Psychology, University of
Canterbury, Christchurch, New Zealand; Richard J. Loewenstein, Shep-
pard Pratt Health System, Baltimore, Maryland, and Department of Psy-
chiatry and Behavioral Sciences, University of Maryland School of Med-
icine, Baltimore; Etzel Carden˜a, Department of Psychology, Lund
University, Lund, Sweden; Paul A. Frewen, Department of Psychology,
University of Western Ontario, London, Canada; Eve B. Carlson, National
Center for Posttraumatic Stress Disorder, Menlo Park, and Veterans Ad-
ministration Palo Alto Health Care System, Palo Alto, California; David
Spiegel, Department of Psychiatry and Behavioral Sciences, Stanford
University School of Medicine.
David H. Gleaves is now at School of Psychology, Social Work and
Social Policy, University of South Australia, Adelaide, Australia.
We thank Franziska Unholzer for assistance with constructing Table 5.
Correspondence concerning this article should be addressed to David
Spiegel, Department of Psychiatry and Behavioral Sciences, Stanford
University School of Medicine, 401 Quarry Road, Office 2325, Stanford,
CA 94305-5718. E-mail: dspiegel@stanford.edu
Psychological Bulletin © 2012 American Psychological Association
2012, Vol. 138, No. 3, 550–588 0033-2909/12/$12.00 DOI: 10.1037/a0027447
550
the depersonalization and derealization experiences related to both
posttraumatic stress disorder (PTSD) and the dissociative disorders
(E. B. Carlson & Dalenberg, 2011; Lanius et al., 2010; Simeon,
2009), and the extreme forms of absorption that appear to be a
diathesis for a variety of pathologies (Allen, Coyne, & Console,
1996). For the upcoming fifth edition of the Diagnostic and
Statistical Manual of Mental Disorders, the American Psychiatric
Association (2010) is proposing a definition of dissociation that
includes reference to the disruptive nature of the symptoms, which
involve “a subjective loss of integration of information or control
over mental processes that, under normal circumstances, are avail-
able to conscious awareness or control, including memory, iden-
tity, emotion, perception, body representation, motor control, and
behavior” (para. 1). Carden˜a and Carlson (2011, pp. 251–252)
have further specified that dissociative symptoms
are characterized by (a) a loss of continuity in subjective experience
with accompanying involuntary and unwanted intrusions into aware-
ness and behavior (so-called positive dissociation); and/or (b) an
inability to access information or control mental functions, manifested
as symptoms such as gaps in awareness, memory, or self-
identification, that are normally amenable to such access/control (so-
called negative dissociation); and/or (c) a sense of experiential dis-
connectedness that may include perceptual distortions about the self
or the environment.
These definitions include the forms of dissociation labeled by
Holmes et al. (2005) as psychological compartmentalization (lack
of continuity and integration between psychological processes)
and forms of dissociation labeled as detachment (altered self-
experience characterized by estrangement from self and/or others).
All these phenomena are included in the Dissociative Experiences
Scale (DES; Bernstein & Putnam, 1986), a measure of dissociation
that has been used in most studies of dissociative phenomena in
clinical and nonclinical samples. The Carden˜a and Carlson (2011)
definition is used in this review.
As described below, all these forms of dissociation have been
studied in a variety of populations, particularly those groups ex-
posed to negative events. Each has also been theoretically and
empirically linked to antecedent experiences of traumatic stress
and/or severe psychological adversity. In general, the trauma
model of dissociation (hereafter the TM) holds that dissociation is
a phylogenetically important aspect of the psychobiological re-
sponse to threat and danger that allows for automatization of
behavior, analgesia, depersonalization, and isolation of cata-
strophic experiences to enhance survival during and in the after-
math of these events (Bremner & Marmar, 1998; Putnam, 1991;
Spiegel, 1984).
In the early 1990s, the clinical syndrome of dissociative amnesia
became temporarily conflated with the more general concept of
recovered memory, creating heated rhetoric about the ubiquity of
trauma as the cause of psychopathology on one the hand and
accusations about false memories of trauma on the other (Dalen-
berg et al., 2007). As the polemics subsided, empirical interest in
dissociation itself increased. The PILOTS database, a centralized
source for research on trauma and PTSD, contains 78 peer-
reviewed citations including the word dissociation or dissociative
from 1986 to 1990. By 2006–2010, this number had increased by
nearly a factor of 5.80 (441 citations). Consensus statements
crossing theoretical divides (e.g., Lindsay & Briere, 1997) suc-
cessfully spurred dissociation theorists and their critics to tighten
their methodology and to test contrasting hypotheses. Inclusion of
dissociation in a wide range of research led to repeated findings
that dissociation was related to more severe forms of trauma-
related syndromes (Alexander & Schaeffer, 1994; Allen, Huntoon,
& Evans, 1999), and that dissociative symptoms were present in a
variety of syndromes not generally thought to be trauma-related,
such as schizophrenia (Yu et al., 2010), attention-deficit disorder
(Endo, Sugiyama, & Someya, 2006), obsessive-compulsive disor-
der (Watson, Wu, & Cutshall, 2004), and bipolar disorder (Oede-
gaard et al., 2008), likely due, at least in part, to inclusion of
individuals with undiagnosed dissociative disorders (Yu et al.,
2010). Dissociation was also implicated in treatment nonresponse
or relapse among the following groups: heroin users (Somer,
2003), agoraphobia patients (Michelson, June, Vives, Testa, &
Marchione, 1998), and those suffering from affective, anxiety, and
somatoform disorders (Spitzer, Barnow, Freyberger, & Grabe,
2007). Furthermore, dissociation was positively associated with
attrition from treatment for drug use among those with dissociative
disorders (Tamar-Gurol, S¸ ar, Karadag, Evren, & Karagoz, 2008)
and among children with higher levels of parent-reported dissoci-
ation (He´bert & Tourigny, 2010). Among individuals who were
treated with exposure therapy for PTSD, 69% of the individuals
exhibiting high dissociation still met criteria for PTSD at
follow-up compared with only 10% of individuals who did not
present with significant dissociative symptoms (Hagenaars, Van
Minnen, & Hoogduin, 2010). This pattern of poor response to, and
lower engagement in, treatment highlights the importance of clin-
ical assessment of dissociation.
In recent years, a number of authors (e.g., Giesbrecht, Lynn,
Lilienfeld, & Merckelbach, 2008; McNally, 2003; Merckelbach,
Horselenberg, & Schmidt, 2002; Merckelbach & Muris, 2001)
have proposed and attempted to test an alternative to the TM.
Proponents of the alternative model (hereafter the fantasy model
[FM]) argue that dissociation is a psychological process causally
unrelated to antecedent traumatic or stressful events. The FM
posits that the trauma histories reported by individuals with dis-
sociative experiences and/or dissociative disorders are largely con-
fabulations or exaggerations resulting from fantasy proneness,
suggestion, and cognitive distortions. As such, FM theorists ac-
knowledge the relationship between reported trauma and dissoci-
ation, but stand the TM on its head by suggesting that dissociation
overlaps with or gives rise to fantasy proneness, suggestibility, and
cognitive distortion, which in turn heighten trauma reporting.
Figure 1 illustrates this basic difference between how the trauma–
dissociation relationship is conceptualized by the TM and the FM.
In essence, for the TM, trauma leads to dissociation via various
biopsychosocial mediator and moderator variables, whereas for the
FM, dissociation leads to reports of trauma via various biopsycho-
social mediator and moderator variables. Figure 1 does not sketch
out the full causal explanation of all endogenous variables, but
does list the typical constructs used by each group of theorists.
Although FM theorists at times present the two views as “con-
trasting” models (e.g., Giesbrecht et al., 2008, p. 622), virtually all
theorists studying one or both models, including Giesbrecht et al.
(2008), would agree that the propositions are not mutually exclu-
sive: that (a) trauma may lead to dissociation and (b) fantasy
proneness—among other factors—may lead to inaccurate trauma
reports. Both also accept a nonrandom relationship between re-
551
TRAUMA AND FANTASY MODEL OF DISSOCIATION
ported trauma and dissociation, but ascribe different reasons for
the relationship. It is therefore important to clarify the true points
of distinction in the two models. These appear to fall into eight
categories, each of which we explore in detail below.
Differential Predictions of the Trauma and Fantasy
Models of Dissociation
Predictions Regarding the Role of Trauma
Most broadly, proponents of the TM posit that the relationship
between trauma and dissociation appears within multiple clinical
and nonclinical groups, and varies in strength depending on a
variety of trauma-specific features (Allen, Fultz, Huntoon, & Bre-
thour, 2002; E. B. Carlson et al., 2001). Proponents of the FM, on
the other hand, propose that the relationship between trauma and
dissociation is weak and inconsistent and/or may be restricted to
cases of profound dissociative psychopathology (Giesbrecht et al.,
2008; Giesbrecht, Lynn, Lilienfeld, & Merckelbach, 2010). In fact,
several FM theorists make the more extreme statement that “dis-
sociation is related to self-reported but not objective trauma”
(Giesbrecht, Lynn, et al., 2010, p. 10), positing instead that the
apparent relationship is an artifact of positive reporting bias (Mer-
ckelbach, Muris, Horselenberg, & Stougie, 2000) and/or the co-
variation of both dissociation and trauma report with suggestibility
and fantasy proneness (Merckelbach et al., 2002).
Because the TM includes a causative role for trauma in the
etiology of dissociation, dissociation is generally predicted to be
higher in recently traumatized populations (Carden˜a & Spiegel,
1993) and chronically or severely traumatized groups (Putnam,
1997). Dissociative symptoms would be predicted to decrease over
time for most individuals, and would be predicted to decrease with
trauma-related treatment (E. B. Carlson & Dalenberg, 2011). In
contrast, given that the FM regards dissociation as a sign of mild
neurological impairment (Giesbrecht, Merckelbach, Kater, &
Sluis, 2007) or fantasy proneness (Giesbrecht et al., 2008), the FM
makes no strong predictions regarding a decline in dissociative
symptoms over time. In fact, a number of FM theorists argue that
trauma treatment is likely to enhance or increase dissociative
symptoms through suggestion (Loftus & Ketcham, 1994).
Predictions Regarding the Role of Suggestibility and
False Memory
Inherent to the FM is the hypothesis that the dissociative indi-
vidual is prone to the construction of fantasies of abuse that are
mistaken for memories (Loftus & Ketcham, 1994; McNally,
2003). The TM differs here by predicting that dissociation is in fact
related to objective trauma and that self-reports of trauma are
generally accurate. In fact, although TM theorists warn against the
use of suggestive language with traumatized clients, particularly in
the context of a potential history of exploitation by authority
(Courtois, 1999; Dalenberg, 2000), there is no strong prediction
made by the TM that dissociative individuals will be prone to false
memories.
Given that fantasy proneness might be an alternative psycho-
logical escape from a traumatic childhood (Barrett, 1992), and
given the shared relationship of both concepts with absorption
(Platt, Lacey, Iobst, & Finkelman, 1998), the models do agree that
the two concepts of fantasy proneness and dissociation would
correlate. Importantly, however, the FM predicts that dissociation
relates to trauma report through the mediators of fantasy proneness
Figure 1. The trauma model and fantasy model of the trauma–dissociation relationship.
552 DALENBERG ET AL.
and suggestibility, and therefore predicts little to no relationship
between dissociation and trauma if fantasy proneness and suggest-
ibility are controlled. Alternatively, the TM clearly predicts an
increment for trauma over fantasy proneness or suggestibility in
the prediction of dissociation, and an increment for dissociation
over fantasy proneness and suggestibility in the prediction of
trauma history.
Predictions Regarding Omission and Fragmentation of
Memory
The TM posits that the dissociative individual is largely attempt-
ing to avoid recall of trauma by conscious and unconscious dis-
avowal of the importance, implications, and/or accuracy or reality
of the memory. According to the TM, the dissociative individual
attempts to avoid thinking about the memory, disconnects from the
emotional content of the memory, and ultimately may fail to recall
some or all of the memory (e.g., DePrince & Freyd, 2004; Dorahy,
2006). The avoidance associated with dissociation may be both
conscious and unconscious, or may be an initially conscious pro-
cess that becomes unconscious over time (see Erdelyi, 1990).
Automatic withdrawal of attention upon exposure to trauma or
reminders of trauma, potentially resulting from dissociative epi-
sodes during encoding, may inhibit associative processing (Lyttle,
Dorahy, Hanna, & Huntjens, 2010), and may result in a lack of the
rich associative network typical of important emotional memories
(cf. Spiegel & Carden˜a, 1991; Stern, 1997). The result is a set of
nonintegrated and fragmented memories (data driven/perceptual
rather than autobiographical/conceptual; Brewin, Dalgleish, & Jo-
seph, 1996; Ehlers & Clark, 2000; Holmes, Brewin, & Hennessey,
2004). This type of processing might account for omissions and
poor agreement in detail across narrative recountings. Over the
course of time, fragmented memories lacking associative networks
may be more easily forgotten. This reasoning supports TM hy-
potheses regarding relationships between dissociation and frag-
mentation of memory and between dissociation and lost or recov-
ered memory.
FM theorists make no claim for the relationship of fragmenta-
tion and dissociation. Omission, however, is thought by FM the-
orists to be negatively related to dissociation (cf. Giesbrecht et al.,
2008). The FM argument here is that any elevation in trauma
report by dissociative individuals is due to exaggeration and fan-
tasy. Therefore, omission of data and loss of detail in severe
trauma is less likely for dissociative individuals than is addition of
detail and enhancement of the trauma description.
Predictions Regarding the Biology and Neurobiology
of Trauma
Both the TM and the FM are consistent with a biological or
sociobiological foundation for dissociation. The TM predicts that
the experience of trauma and high levels of stress are related to
cognitive deficits (Vasterling et al., 2002). The effects will appear
in individuals with clinical dissociative disorders, as well as in
traumatized nondissociative individuals, and will include the errors
of omission, commission, and narrative fragmentation mentioned
earlier (Harvey & Bryant, 1999; Kleim, Wallott, & Ehlers, 2008).
Further, TM theorists expect differences between dissociative and
nondissociative individuals in neurobiological studies, such as in
psychophysiological and functional neuroimaging of trauma sur-
vivors (Lanius et al., 2010), and expect these differences to reflect
or broadly relate to known biologically based responses seen in
animals.
In contrast, the causal path for the FM does not posit a role (or
at least a significant role) for trauma in the neuropsychological or
cognitive deficits seen in dissociative individuals. In Merckelbach
et al.’s (2002) model, for instance, the relationship between dis-
sociation and trauma self-report was hypothesized to be fully
mediated by absent-mindedness and fantasy proneness, with no
role for actual trauma. Cognitive deficiencies inherent to dissoci-
ation were thought to be a primary source of the trauma report.
Mild executive functioning disorder in dissociative individuals is
thought to be present in the presence and absence of trauma
stimuli. An additional point of departure between the models
relates to genetic factors; whereas the TM clearly posits a role for
trauma exposure in the development of dissociation, such that
trauma exposure should increment over genotypes in the predic-
tion of dissociation, the FM makes no such prediction.
Summary of Predictions
1. The TM predicts a consistent positive relationship across
studies between trauma and dissociation, whereas the FM does not.
2. The TM predicts that the relationship between trauma and
dissociation will appear in populations with proven or well-
supported assessment procedures for trauma, whereas the FM
theorists argue that the dissociation–trauma relationship is largely
spurious, and therefore will greatly diminish or disappear if ob-
jective (rather than self-report) measurements of trauma are used
(Merckelbach et al., 2002).
3. The TM posits that the relationship between traumatic expe-
rience and dissociation is, at least in part, causal. Accordingly, the
TM predicts that dissociation will increase after known trauma.
For most affected individuals, the dissociative symptoms will also
wane spontaneously over time. A negative-slope dissociative
symptom frequency and strength should also be seen in pre- and
posttrauma treatment designs. The FM predicts that dissociation is
largely a mental state characterized by high fantasy proneness and
weak executive functioning, and makes no prediction of relation-
ship to time or trauma-based treatment (McNally, 2003; Merck-
elbach et al., 2002), other than proposing that treatment might
increase dissociative symptoms (Loftus & Ketcham, 1994).
4. The TM predicts that although fantasy proneness and disso-
ciation are likely to correlate, trauma will have an increment over
fantasy proneness in the prediction of dissociation. The FM makes
the opposite prediction. Fantasy proneness should predict trauma
report with both variables in the model, whereas dissociation
should not.
5. The FM makes a strong prediction that dissociative individ-
uals, relative to nondissociative individuals, are at higher risk for
false memories of personal trauma. The TM would predict that
fantasy proneness characterizes only a portion of dissociative
individuals, and that it is fantasy proneness, rather than dissocia-
tion, that will control the relationship of the variables to false
memory. The relationship between dissociation and false memory
therefore should be weak and inconsistent.
6. The TM makes a prediction that dissociation is related to the
character of trauma memory, including decreased narrative cohe-
553
TRAUMA AND FANTASY MODEL OF DISSOCIATION
sion and increased fragmentation. The FM, which presents disso-
ciation as related to exaggeration and false generation of trauma,
predicts no relationship or a negative relationship between disso-
ciation and fragmentation or omission.
7. The TM predicts that, over time, dissociative individuals will
be more likely to “forget” or have difficulty accessing important
facets of the memory. The FM states that those who claim recovery
of a memory are unlikely to be recalling an actual trauma.
8. Both models predict some relationship between dissociation
and neuropsychological measures such as working memory (sim-
ilar to those seen in work with PTSD; Vasterling & Brewin, 2005).
The TM holds that the biology of dissociation will ultimately fit
with a theory of a brain-based regulatory response to fear or other
extreme emotion (Lanius et al., 2010). Thus, the psychophysiology
of the dissociative individual should be differentiable from the
nondissociative individual in fear-relevant situations. The FM
makes no prediction in this area.
Measurement of Dissociation and Fantasy Proneness
Prior to the analysis of the evidence for the TM and FM of
dissociation, attention should be given to the measurement of this
construct. The DES (Bernstein & Putnam, 1986) has been used in
over 2,000 studies of dissociation to date, as both the focus for
reviews of positive findings and the central instrument cited by
critics of dissociation and its measurement. The DES also has an
adolescent variant (the Adolescent Dissociative Experiences Scale
[ADES]; Armstrong, Putnam, Carlson, Libero, & Smith, 1997) and
a checklist form for use by parents or other adults assessing young
children (Child Dissociative Checklist [CDC]; Putnam, Helmers,
& Trickett, 1993).
In addition to the DES and its variants, a number of alternative
instruments have appeared, such as the Questionnaire of Experi-
ences of Dissociation (Riley, 1988) and the Dissociation Question-
naire (Vanderlinden, Van Dyck, Vandereycken, Vertommen, &
Verkes, 1993), but these alternatives have not received substantial
research attention. Briere’s (2002) Multiscale Dissociation Inven-
tory (MDI) is a promising new addition to the library of dissoci-
ation measures, particularly given the availability of clinical
norms, but again little is yet available to establish the ability of the
measure to tap important dissociation-related phenomena.
Wright and Loftus (1999) have developed a creative alternative
to the DES. Using the same items as the DES, Wright and Loftus’s
DES-C asks participants not to rate their dissociative symptoms,
but instead to rate whether they are dissociating less or more than
others whom they know. The contention that this capacity is within
the skill set of the dissociative patient (or even the normal control)
has yet to be demonstrated. Further, the DES-C correlates only .25
with the DES (Wright & Loftus, 1999), clearly raising questions
about the similarity of the two measures. We could find no
published evidence showing that the DES-C is in fact a measure of
dissociation. In the review below, research focuses on the original
DES and its child and adolescent variants.
In addition, several diagnostic inventories and interviews have
been developed for the diagnosis of clinical dissociative disorders.
They are not discussed in detail here. However, they include two
diagnostic interviews, the Structured Clinical Interview for DSM–
IV–TR Dissociative Disorders (SCID-D; Steinberg, 1994) and the
Dissociative Disorders Interview Schedule (Ross et al., 1989), as
well as a self-report measure, the Multidimensional Inventory of
Dissociation (Dell, 2006; see Carden˜a, 2008, for a review of the
main measures).
The Dissociative Experiences Scale
The DES is a 28-item self-report measure. In the original Bern-
stein and Putnam (1986) measure, the frequency of each item was
rated along an 11-point visual analog scale. In a revision by E. B.
Carlson and Putnam (1993), the scale was changed to a Likert
model with choices ranging from 0% (never) to 100% (always)at
10 percentage point increments. A sample item is “Some people
have the experience of finding themselves in a place and having no
idea how they got there” (Item 3). The DES has also been shown
to measure both a taxon, often described as “pathological” disso-
ciation, typically measured by the eight-item dissociative taxon, or
DES-T (Waller, Putnam, & Carlson, 1996), and a continuum,
measured by the total scale or by the “nonpathological” absorption
subscale (Waller et al., 1996). The DES-T consists of lower base
rate items targeting measurement of depersonalization and dereal-
ization, identity fragmentation, and amnesia. The absorption sub-
scale is a subset of higher base rate DES items assessing normal
experiences of deep focal attention as well as lapses in attention.
Critics of the current measurement of dissociation and, in par-
ticular, of the DES tend to focus on three issues: the inclusion of
absorption in the domain of dissociation, the reliability and mean-
ing of the taxon, and the more general issue of giving a unitary
label (dissociation) to a wide range of topics, often symbolized by
the argument of whether the DES is unifactorial or multifactorial
(Bernstein, Ellason, Ross, & Vanderlinden, 2001; Giesbrecht et al.,
2008; Watson, 2003). The argument against the inclusion of ab-
sorption in the measurement of dissociation can be made in two
ways: (a) that high absorption is not a symptom of dissociative
disorders, because it is more common in the general population
than DES taxon items, and (b) that absorption is normal and
nonpathological at all levels. The first assumption is not supported
by the empirical evidence. For example, approximately 75% of
patients with diagnosed dissociative disorders in Leavitt’s (2001)
sample had high scores on absorption scales. Dalenberg and
Paulson (2009), using a version of the DES corrected for skew-
ness, found that over 95% of taxon-positive individuals were also
above the cutoff for high absorption. Further, the correlation
between the taxon and absorption factors is very high (e.g., r⫽.80
in Levin & Spei, 2004; r⫽.36–.72 in six psychiatric groups in
Leavitt, 1999). These findings call into question the contention that
“cleaner” measures of dissociation should exclude absorption.
Instead, the data support the inclusion of items that measure
capacities that may be facilitators, precursors, or lower level symp-
toms of dissociation.
With reference to the second assumption, high absorption has
been shown repeatedly to be a marker for severe psychopathology.
Indeed, Allen, Coyne, and Console (1997) reported surprise that
the nonpathological absorption facets of dissociation were more
related to psychosis than were the taxonic items. Absorption cor-
related more highly with severe psychopathology on the Minne-
sota Multiphasic Personality Inventory and the Millon Multiaxial
Inventory than did the amnesia and depersonalization factors (Al-
len et al., 2002).
554 DALENBERG ET AL.
TM theorists and FM theorists both share the concern that the
DES-T yields unacceptably high false-positive rates if used as a
sole diagnostic instrument (cf. Carden˜a, 2008), and that it has
modest reliability in nonclinical samples when dissociative disor-
der should be rare or nonexistent (r⫽.62 over 2 months; Watson,
2003). However, the DES was designed as a screening, not a
diagnostic, instrument (Bernstein & Putnam, 1986). Given the
higher likelihood of false positives in screening for low base rate
diagnoses, we agree that the DES-T should signal the likely
presence of dissociative symptoms and the need for further eval-
uation for dissociative disorder, rather than the definitive presence
of such a disorder. We also agree that dissociative symptoms do
not always (or even typically) lead to dissociative disorder. We
focus on the research using the DES, and the adolescent and child
versions (the ADES and CDC), as they are the dissociative mea-
sures most frequently chosen in trauma studies, most inclusive of
the full range of dissociative symptoms, and most psychometri-
cally acceptable (given lower base rates for DES-T and lower
reliability relative to the DES in nonclinical samples).
The proper use of the full scale and subscales of the DES has
generated considerable discussion within and across the TM and
FM literature (cf. Giesbrecht et al., 2008; Waller & Ross, 1997;
Watson, 2003). Although we focus on the full DES in keeping with
our overall definition of dissociation, we do not view the issue of
the potentially multidimensional nature of the DES as an inherent
problem. Most screening scales meant to measure complex, di-
mensional, diagnostically relevant concepts are multifactorial. For
example, Shafer (2006) conducted a meta-analysis of four com-
monly used depression scales (the Beck Depression Inventory,
Center for Epidemiologic Studies Depression Scale, Hamilton
Rating Scale for Depression, and Zung Self-Rating Depression
Scale), finding all four to be reliably multidimensional.
On a more technical note, the multidimensionality of the DES
may be an artifact of “difficulty factors” (Carroll, 1945), that is,
sets of items with differing base rates of agreement. Bernstein et al.
(2001), who examined the skewness and difficulty of each DES
item across clinical and nonclinical samples, concluded that the
absorption items might also have been called “commonly endorsed
dissociative symptoms,” the amnesia factor might have been la-
beled “infrequently endorsed dissociative symptoms,” and the
depersonalization–derealization factor could have been called
“dissociative symptoms endorsed at an intermediate level” (p.
107). Further, although many researchers find that the DES is
multifactorial using the Kaiser criterion, the first factor of the
analysis often has a much greater eigenvalue than the second (e.g.,
11.61 vs. 1.79 for Amdur & Liberzon, 1996; 11.14 vs. 1.82 for
Dunn, Ryan, & Paolo, 1994; 12.65 vs. 1.83 for Ruiz, Poythress,
Lilienfeld, & Douglas, 2008). With the scree criterion, these anal-
yses would be redefined as unifactorial. Finally, researchers who
use the factors rather than the total scale have repeatedly noted
high correlations between the factors (Allen et al., 2002; Gies-
brecht, Merckelbach, et al., 2007; Pekala et al., 1999–2000).
Peritraumatic and State Dissociation
Although inclusion of a range of types of dissociation seems
appropriate as long as adequate intercorrelations can be shown, the
equation of trait dissociation with peritraumatic dissociation is
problematic. Peritraumatic dissociation (i.e., dissociation at or
around the time of trauma) is often measured by the Peritraumatic
Dissociative Experiences Questionnaire (PDEQ; Marmar, Weiss,
& Metzler, 1997) or the Stanford Acute Stress Reaction Question-
naire (SASRQ; Carden˜a, Koopman, Classen, Waelde, & Spiegel,
2000). The PDEQ contains 10 self-report items (also available in
clinician administration form) measured along a 5-point scale
ranging from 1 (not at all true)to5(extremely true). Dissociative
experiences assessed during or immediately after a traumatic event
include “I felt disoriented; that is, there were moments when I felt
uncertain about where I was or what time it was” (Item 10). The
SASRQ is a 30-item self-report measure of acute reactions to
stress, including a subscale for dissociative reaction, with separate
items for severity of disturbance and duration of worst symptoms.
The 30 experiential items are assessed along a rating scale of 0 (not
experienced)to5(very often experienced). A sample item is “I felt
a sense of timelessness” (Item 4).
Dissociative symptoms may appear during trauma consequent to
hyperventilation, panic, or arousal (Bryant et al., 2011; Nixon &
Bryant, 2006). Such symptoms are not necessarily signs of a
general tendency to dissociate or risk for a trauma disorder (Briere,
Scott, & Weathers, 2005; Tichenor, Marmar, Weiss, Metzler, &
Ronfeldt, 1996). In Tichenor et al.’s (1996) veteran sample, the
correlation of the DES with the PDEQ was .26. An additional
problem with measurement of peritraumatic dissociation is that
reports are likely to be accurate only if they are collected soon after
an event. Reports of peritraumatic dissociation associated with
events that occurred months or years earlier will likely be influ-
enced by emotional states at the time of recall.
A commonly used state dissociation measure is Bremner et al.’s
(1998) Clinician-Administered Dissociative States Scale, contain-
ing 27 items rated along a scale from 0 (not at all)to4(extremely),
with 19 items self-reported and eight observer reported. Equating
measures of trait dissociation such as the DES with measures of
state dissociation is somewhat more defensible (rwith DES ⫽.48
in the initial study) than the equating of the DES with peritrau-
matic dissociation, but should also be considered in the context of
the trigger event. Although dissociation at the time of the trauma
is arguably a state, the term state dissociation (in contrast to
peritraumatic dissociation) refers to dissociative symptoms at a
particular point in time subsequent to the trauma. Additionally, it
should be noted that a marked increase in state dissociation during
experimental procedures designed to induce it, such as ketamine
infusion (an anesthesia with dissociative side effects; Rowland et
al., 2005) or a dot-staring task (used to induce hyperfocus and
absorption; Leonard, Telch, & Harrington, 1999), may have a
different meaning than state dissociation elevation reported during
admission to a psychiatric center or after trauma memory exposure
(Bremner et al., 1998). Thus, trauma-related state dissociation may
have more in common with trait dissociation than does pharma-
cologically induced state dissociation.
Measurement of Fantasy Proneness
The concept of fantasy proneness, introduced by S. C. Wilson
and Barber (1983), is most commonly measured by their Inventory
of Childhood Memories and Imaginings (ICMI) or by Merck-
elbach, Horselenberg, and Muris’s (2001) Creative Experiences
Questionnaire (CEQ). Merckelbach et al.’s factor analysis of the
CEQ found nine factors with eigenvalues greater than 1.0. Klinger,
555
TRAUMA AND FANTASY MODEL OF DISSOCIATION
Henning, and Janssen (2009) found 18 factors in the ICMI, with
only a two-component promax-rotated solution proving stable
across two approximate halves of their college sample (n⫽232).
The components of fantasy proneness proposed by S. C. Wilson
and Barber based on in-depth interviews, and since supported by
further empirical work (cf. Merckelbach et al., 2001), include large
amounts of time spent fantasizing, vivid childhood memories, the
experience of bodily component of fantasies, and intense religious
and paranormal experiences.
The factors of the ICMI and the CEQ appear to measure very
different constructs and correlate differently with dissociation. For
example, Klinger et al.’s (2009) factor analysis of the ICMI found
that Component 1 of the ICMI correlated .66 with an estimate of
the DES taxon (using the Curious Experiences Survey; Goldberg,
1999) and Component 2 correlated .22 with this estimate. Com-
ponent 1 was also related to depression, anxiety, and somatization,
whereas Component 2 was unrelated to these variables. It remains
unclear whether the two components actually measure proneness
to fantasy in equal measure, or whether they instead reflect dif-
ferent goals to which fantasy may be harnessed (e.g., fearful
avoidance, which correlates more with Component 1, and positive
constructive daydreaming, which correlates with both factors). In
either event, Klinger et al. concluded that the full-scale score
“cannot yield general statements regarding dispositions to fanta-
size, and fantasy-proneness is accordingly a misleading summary
label for what the full-scale ICMI measures” (p. 510). Similarly,
after finding the CEQ to be multidimensional, Sa´nchez-Bernardos
and Avia (2004) concluded that vividness–intensity of fantasies,
make-believe or suggestibility, and fantasy to escape may be
separate components of fantasy proneness with differential rela-
tionships to psychological risk. Recent evidence further suggests
that Component 1 is prevalent among high hypnotizables who are
also strong dissociators, but not among those who do not have such
propensity (Terhune, Carden˜a, & Lindgren, 2011). In summary,
deeper analyses of the concept and construct of fantasy proneness,
and separate analysis of the factors of fantasy proneness scales, are
necessary to further the understanding of the fantasy proneness–
dissociation relationship. Nonetheless, in keeping with the models
used by FM theorists, herein the CEQ and the ICMI are used as the
best available proxies for the trait of fantasy proneness.
Research Inclusion Criteria
The proposition that good data are lacking to support the link
between trauma and dissociation is a cornerstone of the FM
(Kihlstrom, 2005; Lilienfeld et al., 1999; Merckelbach et al., 2002;
Merckelbach & Muris, 2001) and is one of the primary tenets
distinguishing this theory from the more widely accepted TM. For
example, McNally (2003) wrote that trauma theorists (DePrince &
Freyd, 2001, in this case) “appear to believe that a high DES
[dissociation] score is related to trauma,” noting simply that this is
“incorrect” (p. 176). We would argue that such a statement is
questionable, but the point that TM theorists should not assume a
causal relationship between trauma and dissociation is well taken.
On the other hand, gathering the full literature on dissociation to
test these conflicting assumptions is a daunting task. Entering the
words dissociation and dissociative into PsycINFO yielded 16,237
references in February 2011, given that the word dissociative is
used in a number of nonpsychological contexts (for instance, as an
opposite of associative). To gather the research for this review, we
therefore took the following steps:
1. To overinclude those articles that might support the FM
prediction, we crossed the words dissociative and dissociation with
the terms commission
ⴱ
,false mem
ⴱ
, and fantasy. Only peer-
reviewed articles available in English were included. This yielded
273 references.
2. The first 250 randomly selected empirical articles on psycho-
logical dissociation located among the original 16,237 using the
term dissociation or dissociative showed that 96% of the articles
with a dissociation measure chose the DES, the ADES, or the CDC
as the measure of dissociation. Therefore, the citation index for
PsycINFO was used to collect all articles that cited the DES,
ADES, or CDC. The articles used as foundational citations were
Bernstein and Putnam (1986), E. B. Carlson and Putnam (1993),
Armstrong et al. (1997), and Putnam et al. (1993). All articles
listed in PsycINFO as citing one or more of the foundational
articles were included.
3. To maximize the quality of the studies, we also crossed the
search term dissociative or dissociation with prospective or lon-
gitudinal. The results were reviewed, and articles that referred to
psychological dissociation were added to the database.
The total reference base consisted of 1,492 articles meeting one
or more of these criteria. This number was further supplemented
by inclusion of a search of the ProQuest database for dissertations
and theses to partially capture the gray (unpublished) literature.
This last task is recommended in order to avoid the bias created by
lack of publication of statistically nonsignificant findings. This
yielded 40 dissertations for the fantasy search and 73 dissertations
(from 2000 to 2010) for the DES search.
Findings
Evidence for Prediction 1: Is There Consistent
Evidence for the Trauma–Dissociation Connection?
Relationship between trauma and dissociation. The rela-
tionship between trauma and dissociation has been found in a
large array of specific populations, including patients with
schizophrenia (Holowka, King, Saheb, Pukall, & Brunet, 2003),
obsessive-compulsive disorder (Lochner et al., 2004), trichotil-
lomania (Lochner et al., 2004), and psychosomatic disorders
(Besiroglu et al., 2009), as well as those with alcohol depen-
dency (Evren, S¸ ar, & Dalbudak, 2008). The number of such
studies is too large to be reviewed in detail here. With such a
variety of studies to discuss, there is a danger of overemphasis
on one of the few studies that found no correlation between
trauma and dissociation in a small and nongeneralizable sample
(e.g., Cima, Merckelbach, Klein, Shellbach-Matties, & Krem-
er’s, 2001, study of 30 male forensic psychiatric patients; r⫽
⫺.07, p⬎.05) or those that find unusually high correlations in
such samples (e.g., Lochner et al.’s, 2004, study of 31 tricho-
tillomania patients; r⫽.61, p⬍.01). Focusing on evidence
from the most methodologically rigorous studies, Table 1 pres-
ents all studies from the database that met the following criteria:
(a) effect size of the trauma–dissociation relationship was re-
ported or data could be transformed into an effect size (e.g.,
from means and standard deviations), with statistically nonsig-
nificant studies with no effect size data set to 0; (b) participants
556 DALENBERG ET AL.
Table 1
Relationship of Trauma and Dissociation
Study Participants
Trauma
type Trauma measure Dissociation measure
a
r
Nonclinical samples
Akyu¨z et al., 2005 251 adult women PA CANQ DES (Turkey) .06
See above SA CANQ DES .18
ⴱⴱ
See above TOT CANQ DES .22
ⴱⴱⴱ
Chu & DePrince, 2006 72 adult mothers BT UCLA–PTSD Index DES .34
ⴱⴱ
Collin-Ve´zina & Hebert, 2005 67 children evaluated for abuse
and matched controls
SA Hospital evaluation and interview CDC .38
ⴱⴱⴱ
DePrince et al., 2008 97 children TOT Guardian report on UCLA–PTSD Index ADES .21
ⴱ
97 children TOT Guardian report on UCLA–PTSD Index CDC .25
ⴱ
Dorahy et al., 2007 66 adults DV TEC DES .40
ⴱⴱⴱ
Dutra et al., 2009 56 young adults DMC Behavioral codes on AMBIAC DES .38
ⴱⴱ
56 young adults TOT CTES–R DES .16
Geraerts et al., 2005 98 adults
b
SA Self-report DES (Netherlands) .31
ⴱⴱⴱ
Kisiel & Lyons, 2001 114 wards of DCFS PA Reported by DCFS caretaker ADES .20
ⴱ
114 wards of DCFS SA Reported by DCFS caretaker ADES .24
ⴱⴱ
114 wards of DCFS PA Reported by DCFS caretaker CDC .32
ⴱⴱⴱ
114 wards of DCFS SA Reported by DCFS caretaker CDC .30
ⴱⴱ
Macfie et al., 2001a 198 children SA CPS records CDC .11
198 children PA CPS records CDC .39
ⴱⴱⴱ
Macfie et al., 2001b 88 children TR Coded from CPS records CDC .42
ⴱⴱⴱ
McNally et al., 2000 68 adults
b
SA Self-report DES .32
ⴱⴱ
McNally et al., 2006 166 adults
b
SA Self-report DES .42
ⴱⴱⴱ
Narang & Contreras, 2005 76 mothers PA CHQ DES .37
ⴱⴱⴱ
Na¨ring & Nijenhuis, 2005 147 adults TOT TEC DES (Netherlands) .27
ⴱⴱⴱ
Nilsson & Svedin, 2006 391 adolescents TOT DIS-Q ADES (Sweden) .28
ⴱⴱⴱ
Noll et al., 2003 166 children SA Substantiated by CPS CDC .36
ⴱⴱⴱ
Ogawa et al., 1997 168 young adults TR
c
Home observation, CPS records, parent
interview
DES .26
ⴱⴱⴱ
Sayar et al., 2005 173 adolescents PA Self-report ADES (Turkey) .31
ⴱⴱⴱ
Smith et al., 2010 50 adults TOT THS DES .44
ⴱⴱⴱ
Somer, 2002 90 adults TOT TEQ DES (Israel) .39
ⴱⴱⴱ
Trickett et al., 2001 166 children SA Verified through DCFS CDC at 6 months (Time 1) .34
ⴱⴱⴱ
158 young adults SA Verified through DCFS ADES at 7 years after Time 1 .16
ⴱ
Twaite & Rodriguez-Srednicki, 2004 284 adults PA Self-report DES .24
ⴱⴱⴱ
284 adults SA Self-report DES .18
ⴱⴱ
Zorog˘lu et al., 2003 839 adolescents TR CANQ ADES (Turkey) .33
ⴱⴱⴱ
Clinical samples
Brunner et al., 2000 198 adolescent inpatients SA Therapist reports based on guardian report,
DCFS records, and self-report
ADES (Germany) .36
ⴱⴱⴱ
See above PA See above ADES .22
ⴱⴱⴱ
E. B. Carlson et al., 2001 178 adult inpatients Violent SA Structured interview DES .52
ⴱⴱⴱ
See above Other SA See above DES .49
ⴱⴱⴱ
See above Violent PA See above DES .35
ⴱⴱⴱ
See above Other PA See above DES .28
ⴱⴱⴱ
Dell, 2006 204 clinical and nonclinical PA TEQ DES .34
ⴱⴱⴱ
See above SA TEQ DES .44
ⴱⴱⴱ
See above TOT TEQ DES .47
ⴱⴱⴱ
(table continues)
557
TRAUMA AND FANTASY MODEL OF DISSOCIATION
Table 1 (continued)
Study Participants
Trauma
type Trauma measure Dissociation measure
a
r
El-Hage et al., 2002 140 adult outpatients TR CAPS DES (France) .49
ⴱⴱⴱ
Francia-Martı´nez et al., 2003 100 adult inpatients SA BSAE DES (Puerto Rico) .14
Freyd et al., 2005 99 adults with chronic illness
or pain
BT BBTS DES .43
ⴱⴱⴱ
Gast et al., 2001 102 adult inpatients PA CTQ DES (Germany) .47
ⴱⴱⴱ
See above SA CTQ DES .32
ⴱⴱⴱ
See above TOT CTQ DES .46
ⴱⴱⴱ
Peleikis et al., 2005 112 women SA Confirmation by therapists DES (Norway) .21
ⴱ
Reyes-Pe´rez et al. 2005 64 children SA Confirmation by therapists CDC (Puerto Rico) .63
ⴱⴱⴱ
Salmon et al., 2003 123 gastroenterology patients SA MHQ DES .27
ⴱⴱⴱ
See above PA MHQ DES .23
ⴱⴱ
See above ASA MHQ DES .27
ⴱⴱⴱ
Skarbø et al., 2004 100 outpatients TOT KLEHS DES (Norway) .30
ⴱⴱ
Somer et al., 2001 70 outpatients TOT TEQ DES (Israel) .62
ⴱⴱⴱ
Spitzer, Vogel, et al., 2007 122 inpatients with severe
mental illness
TOT PDS DES (Germany) .27
ⴱⴱ
Sullivan, 2003 192 inpatients PA Self-report DES .25
ⴱⴱ
See above SA Self-report DES .40
ⴱⴱ
Note. Values in italics derived from studies using objective documentation of trauma. Trauma type: PA ⫽physical abuse; SA ⫽sexual abuse; TOT ⫽total on trauma scale; TR ⫽any trauma; BT ⫽
betrayal trauma; DV ⫽domestic violence; DMC ⫽dysfunctional maternal communication; ASA ⫽adult sexual assault. Trauma measure; CANQ ⫽Child Abuse and Neglect Questionnaire;
UCLA–PTSD ⫽University of California, Los Angeles–Posttraumatic Stress Disorder; TEC ⫽Traumatic Experiences Checklist; AMBIAC ⫽Atypical Maternal Behavior Instrument for Assessment
and Classification; CTES–R ⫽Childhood Traumatic Experiences Scales–Revised; self-report ⫽investigator determined questions; DCFS ⫽Department of Child and Family Services; CHQ ⫽
Childhood History Questionnaire; DIS-Q ⫽Dissociation Questionnaire; CPS ⫽Child Protective Services; THS ⫽Traumatic History Screen; TEQ ⫽Traumatic Experiences Questionnaire; CAPS ⫽
Clinician Administered PTSD Scale for DSM–IV; BSAE ⫽Brief Scale of Abusive Experiences; BBTS ⫽Brief Betrayal Trauma Survey; CTQ ⫽Childhood Trauma Questionnaire; MHQ ⫽Medical
History Questionnaire; KLEHS ⫽Kerkhof Life Events and History Scale; PDS ⫽Posttraumatic Diagnostic Scale. Dissociation measure: DES ⫽Dissociative Experiences Scale; CDC ⫽Child
Dissociation Checklist; ADES ⫽Adolescent Dissociative Experiences Scale.
a
If a translated version of measure was used, the country or territory in which the study was conducted is listed.
b
Recovered and continuous memory survivors combined and compared with
control.
c
Traumatic events in first 2 years of life.
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
558 DALENBERG ET AL.
with no trauma of the type studied were included; (c) sample
size was 50 or greater; and (d) the study used a community
sample or a clinical sample including a range of psychiatric
diagnoses. Thus, samples consisting entirely of dissociative
disordered patients or those with PTSD, which may have re-
stricted values on trauma likelihood or dissociation, were not
included, but consecutive psychiatric admissions samples or
groups of children in therapy are represented. College samples,
which are likely to be biased in favor of low impairment, were
not included. Lev-Wiesel, Daphna-Tekoah, and Hallak’s (2009)
large sample of pregnant women was not included given the
complex literature on the relationship of pain, stress, and dis-
sociation (cf. Luda¨scher et al., 2007). Studies using only certain
subscales of the DES also were not included. Studies that
appeared to test the same sample in different publications and
studies that limited trauma effects to emotional abuse were
excluded. The effect size rwas chosen, since the majority of
studies reported this figure.
Table 1 presents the results of 38 studies that met our criteria.
The average weighted effect size was .31 for the 19 sexual
abuse samples, .27 for the 12 physical abuse studies, and .34 for
the 16 total trauma score studies (for the E. B. Carlson et al.,
2001, study, the two relevant values were averaged). The over-
all weighted restimate was .32. Fixed-point estimates were
made via Comprehensive Meta-Analysis software. All values
are moderate effect sizes. Qvalues were between 24.59 (for
sexual abuse) and 63.71 (for all studies), indicating heteroge-
neity of effect sizes.
The table also illustrates that large population studies and well-
controlled comparison studies do exist that test the relationship of
trauma and dissociation. Collin-Ve´zina and He´bert’s (2005) study
of 134 children (abused children, evaluated in a hospital, and their
matched controls) found a statistically significant relationship be-
tween sexual abuse and dissociation with a large effect size.
Zorog˘lu et al. (2003), who examined the relationship between
trauma and dissociation in 839 Turkish high school students, found
that trauma and dissociation were strongly related, with stepwise
increments in dissociation based on the number of types of trauma
experienced (i.e., one trauma vs. no trauma Hedges’s g⫽0.56,
two traumas vs. no trauma Hedges’s g⫽0.84, and three traumas
vs. no trauma Hedges’s g⫽1.12). E. B. Carlson et al. (2001)
found large magnitude correlations for both sexual abuse and
physical abuse in their inpatient sample, with violent sexual abuse
showing an increment over family environment variables in pre-
dicting dissociation and PTSD. Thus, in summary, the TM Pre-
diction 1 is supported; the relationship between trauma and disso-
ciation is not “weak or nonexistent,” as suggested in the FM, but
instead is consistent and moderate in size, as suggested by the TM.
Relationship between trauma and dissociative disorder di-
agnosis. Critics of the TM also question the relationship of
trauma to the dissociative disorder diagnoses. To specifically ex-
amine this question, we collected the 481 citations on dissociative
disorders and abuse or trauma history (child abuse or sexual abuse
or physical abuse or trauma hist
ⴱ
) to locate empirical evaluations
of the prevalence of trauma in the background of the most severe
dissociative cases. Table 2 shows the results of the four studies
with n⬎50 that examined this question. Many more studies exist
(e.g., Gast, Rodewald, Nickel, & Emrich, 2001; Middleton &
Butler, 1998; Ross, Duffy, & Ellason, 2002; Ross et al., 1991);
however, most such studies did not include a comparison group
Table 2
Abuse and Trauma Rates in Dissociative Disorder Samples Versus Comparison Groups
Study Comparison Abuse type Trauma rate (%) Effect size ()
Duffy, 2000 82 DD PA 57.3 .36
119 Inpatients 21.8
82 DD SA 51.2 .24
119 Inpatients 27.7
82 DD TR 75.6 .36
119 Inpatients 39.5
Foote et al., 2006 24 DD PA 70.8 .38
58 Outpatients 29.3
24 DD SA 75.0 .44
58 Outpatients 27.5
Ross & Ness, 2010 266 female DID PA 83.8 .74
318 controls 9.7
37 male DID PA 83.8 .78
184 controls 4.3
266 female DID SA 90.2 .78
318 controls 11.6
37 male DID SA 73.7 .75
184 controls 2.7
S¸ar et al., 2007 115 DD PA 18.3 .13
513 non-DD 8.4
115 DD SA 9.6 .20
513 non-DD 1.2
115 DD MT 60.9 .22
513 non-DD 32.7
Note. All comparison groups were from the general population unless otherwise noted. DD ⫽dissociative
disorder; DID ⫽dissociative identity disorder; PA ⫽physical abuse; SA ⫽sexual abuse; TR ⫽any trauma;
MT ⫽any maltreatment (includes neglect).
559
TRAUMA AND FANTASY MODEL OF DISSOCIATION
that would allow an effect size computation, presenting instead
large samples of dissociative disorder or dissociative identity dis-
order (DID) patients. Trauma history was found in 50%–100% of
such individuals in all studies (with the exception of the Turkish
study by S¸ ar, Akyu¨z, & Dog˘an, 2007). These results also support
the TM, but differ from the data in Table 1. For Table 1, all clinical
samples showed a general relationship between trauma and level
of dissociativity on the DES. However, the base rate of DID in
most clinical samples is low (1.3% in Ross’s, 1991, nonclinical
sample); thus, the correlation coefficient can be misleading. For
instance, in a large sample (N⫽618), Briere’s (2006) correlation
of .11 between trauma and clinical elevation on the MDI accounts
for less than 2% of the variance. If the same results are translated
into the language of binomial probability to make base rates more
visible, as Briere made possible through cross-tabulation charts,
the probability of clinical elevation in the MDI is 4 times greater
in the trauma-exposed compared with a nonexposed sample (8%
vs. 2%). Further, the probability of a trauma history given an
elevated MDI in this sample was 90%. Similarly, in the study with
the weakest effect size in Table 2 (S¸ ar et al., 2007), the probability
of abuse within the dissociative disorder samples were still 2–4
times higher than the rates within psychiatric controls. The heter-
ogeneity of effect sizes are reflected in the very high Qvalues of
263.63 (p⬍.001) for physical abuse and 270.40 (p⬍.001) for
sexual abuse. The mean-weighted rwas .54 for the five sexual
abuse samples and .52 for the five physical abuse samples. Again,
with diagnosis rather than dissociation as a continuum, the hypoth-
esis of the consistent relationship between trauma and dissociation
in Prediction 1 is supported.
Evidence for Prediction 2: Does the Trauma–
Dissociation Relationship Disappear in Studies With
“Objective” Measures of Trauma?
In Table 1, 10 studies included external criteria for determina-
tion of maltreatment status. Ten graduate student raters—blind to
the hypotheses of this review and blind to the results of each
study—made this judgment with 100% agreement. The “objec-
tive” data included confirmation by therapists (with access to
guardians and Child Protective Services [CPS] reports), protective
agency report determined by researchers, or, in the case of Dutra,
Bureau, Holmes, Lyubchik, and Lyons-Ruth (2009), observer be-
havioral codes of mothers’ treatment of their infants. In Dutra et
al., disrupted maternal communication included ratings of sexual-
ized behavior, hostile and intrusive behavior, contradictory cues,
withdrawal, and fearful-disoriented behavior on the part of the
mother in the Ainsworth Strange Situation task. Nine of these 10
studies tested the correlation between dissociation and sexual
abuse, whereas three also tested the correlation between dissocia-
tion and physical abuse. The FM prediction that objectively deter-
mined trauma would show lower correlations with dissociation
than self-reported trauma thus could be tested by comparing the
effect size of the objective studies with the studies using a stan-
dardized self-report measure or a single-item self-label of sexual
abuse. Using a weighted mean effect size, the objective studies on
sexual abuse had a weighted average rof .30, whereas the self-
report, standardized measure, or structured interview studies had a
weighted average effect size of 32. The three objective measure
studies on physical abuse had a weighted average rof .30, com-
pared with the average weighted rfor the remaining physical abuse
studies of .26. The objective physical abuse analysis yielded a
nonsignificant Qvalue of 3.67 (p⬎.05), with the remaining
analyses showing Qvalues at or greater than 23.32 (p⬍.01).
These results contradict the FM prediction, and go to the heart of
the FM argument. If the trauma–dissociation relationship were
largely due to fantasy proneness and subsequent exaggeration of
trauma, clearly the relationship should be weaker when trauma is
measured with greater objectivity. This argument has been made
explicitly in Giesbrecht et al.’s (2008) recent review. They were
able to locate two studies with objective criteria, both with small
and nongeneralizable samples, noting that neither reached statis-
tical significance. The 10 studies with larger and more generaliz-
able samples, all of which did support the TM hypothesis, were not
discussed by those authors. In this full review comparing studies
with self-report to those using objective measures, studies with
self-report measures of trauma did not show a greater relationship
to dissociation than those with objective measures. Again, these
findings support the TM position, not the FM view.
There have been no large-scale studies of the objective evidence
for trauma reported by dissociative disordered patients including
control groups. Longitudinal studies are less realistic here, given
the base rate of dissociative disorders. However, the smaller stud-
ies that have followed up on the evidence for child trauma history
in DID patients have confirmed the existence of such trauma.
Coons (1994) found documented corroboration (e.g., CPS and
police records) for 20 of his 21 child and adolescent DID and
dissociative-disorder-not-otherwise-specified patients. Similarly,
Coons and Milstein (1986) found documentation through medical
records or family testimony in 17 of 20 adult DID patients (see
also Lewis, Yeager, Swica, Pincus, & Lewis, 1997). Further,
Hornstein and Putnam (1992) described two samples of children
and adolescents with dissociative disorders totaling 74 partici-
pants, all of whom had reported histories of a wide variety of types
of maltreatment, including physical abuse, sexual abuse, witness-
ing parental death, and/or neglect. Social service investigation
substantiated 95% of these histories.
In support of the call for further research with more sophisti-
cated models, it should be emphasized here that prospective lon-
gitudinal studies have found that objective trauma leads to height-
ened dissociation in children who have disorganized attachment
(e.g., Lyons-Ruth, Dutra, Schuder, & Bianchi, 2006; Ogawa et al.,
1997), been victims of corroborated sexual abuse (Noll, Trickett,
& Putnam, 2003; Trickett, Noll, Reiffman, & Putnam, 2001), or
experienced verified painful medical procedures (Diseth, 2006).
For example, in an ongoing, case-controlled longitudinal study of
girls with a substantiated history of child sexual abuse (CSA;
Trickett et al., 2001), participants were assessed with a variety of
biological, psychometric, and educational measures, as well as
measures of social functioning. They were assessed within 6
months of the initial report of CSA to protective services and again
7 years later. The abused girls had higher levels of caregiver-rated
dissociation at intake than nonabused controls. Furthermore, the
abused girls who had experienced more severe forms of abuse (i.e.,
earlier onset, father figure abuse) had higher levels of self-reported
dissociation at a 7-year follow-up. Similarly, in the longitudinal
studies of E. A. Carlson (1998) and Ogawa et al. (1997), disorga-
nized attachment in infancy predicted observed dissociative be-
havior reported by elementary and high school teachers, and cor-
560 DALENBERG ET AL.
related with self-reported dissociation at age 19. Within the
disorganized group, higher dissociation scores were found for the
group that had experienced documented traumas in childhood and
adolescence. Ogawa et al. also reported a statistically significant
correlation between trauma (with both objective and parent self-
report documentation) and dissociation at Time 1 (infancy), Time
4 (age 16–17), and Time 5 (age 19) with a sample of 168. In
Diseth’s (2006) smaller study of children who had experienced
repeated and painful medical procedures (N⫽42), an objective
trauma, dissociation in both adolescence and young adulthood
correlated with number of hospitalizations (r⫽.59 with the ADES
and r⫽.79 with the DES, 10 years later).
Many prospective studies follow at-risk samples in order to have
realistic probability of finding traumatized individuals with vary-
ing symptom levels. Barring the random (and unethical) assign-
ment of individuals to traumatizing conditions, the optimal deter-
mination about whether dissociation is causally related to trauma
would be to prospectively study dissociative symptoms and PTSD
symptoms in a sample that is representative of the general popu-
lation that has been exposed to verified trauma. It would further
clarify the relationship between dissociation and trauma if symp-
toms were assessed in real time rather than retrospectively and if
the symptoms were assessed longitudinally at multiple time points.
A study following those criteria has recently been conducted (E. B.
Carlson et al., 2011). Dissociation and PTSD symptoms were
assessed in real time (at 4-hr intervals over 7 days) in 62 adults
who were exposed to traumatic stress (either severe injury or
severe injury to a loved one). Participants experienced an initial
elevation in dissociative symptoms that dissipated over time. In
addition, the relationship between PTSD and dissociation symp-
toms assessed in real time was extremely high at r⫽.83 and
homoscedastic, indicating that dissociation symptoms were
strongly related to the expected responses to trauma exposure. In
summary, across methodologies, dissociation is related to objec-
tive trauma.
Evidence for Prediction 3: Does Level of Dissociation
Change With Time and After Trauma Treatment?
The effect of treatment on dissociative symptoms. Table 3
shows the results of seven studies showing a pre- and postchange
in dissociation after short-term trauma treatment. Given that power
for repeated measures is greater than power for between-group
designs, studies with samples greater than 25 are presented. To be
consistent and conservative, we calculated effect sizes (dvalues)
using the standard deviation of baseline scores as the denominator.
Thus, we did not use the standard deviation of change scores, even
if the original researchers did, because doing so can lead to an
overestimate of effect sizes (Dunlap, Cortina, Vaslow, & Burke,
1996). All studies reported reductions in dissociation after treat-
ment for at least one treatment approach. Controls did not show a
statistically significant decrease. In the case of Bohus et al. (2004);
Chard (2005); Rothbaum, Astin, and Marsteller (2005); and Van
Emmerik, Kamphuis, and Emmelkamp (2008), nontreatment con-
trols did not show a statistically significant decrease in dissociation
in the comparable period. The same pattern appeared in the only
pharmacology study that met our criteria. Chronic PTSD patients
who were treated with paroxetine (in a randomized double-blind
study) reliably dropped in dissociation symptoms, with no statis-
tically significant change in the placebo condition, but this study
had a large dropout rate and should be replicated (Marshall et al.,
2007). Because the studies varied in treatment type and treatment
length, we do not present one overall estimate of effect size.
Rather, we note that most or all studies of the effect of trauma-
relevant treatment on dissociative symptoms found results support-
ing the TM. We were unable to locate any FM explanations of
decrease in dissociation after trauma treatment.
Change in dissociative symptoms across time. Researchers
asking questions about short-term changes in dissociation after
stressors have typically used state dissociation measures. A num-
ber of pre- and postdesign studies have been conducted. Lanius et
al. (2005) observed an increase in state dissociation subsequent to
exposure of combat veterans to their own scripted trauma memo-
ries. Morgan et al. (2001) also reported increases in dissociation in
a resilient group of Special Forces soldiers after highly stressful
survival training. Although most soldiers had prior trauma, 42% of
the sample reported dissociative symptoms pretraining, whereas
96% reported symptoms after training. In perhaps the most theo-
retically interesting of the state dissociation studies, Zoellner,
Sacks, and Foa (2007) used the Velten mood induction procedures.
In this procedure, phrases associated with state dissociation, such
as “I feel detached and distant today,” are read to the participant to
induce a dissociative experience. Dissociation induction was most
easily accomplished by those with PTSD as compared with
nontrauma-exposed participants, suggesting that trauma-exposed
individuals have been sensitized to experience dissociation.
Table 3
Pre- and Posttrauma Effect Sizes of Decreases in Dissociation Measures With Treatment
Study Type of treatment and description of sample NEffect size (d)
Abramowitz & Lichtenberg, 2010 Hypnotic treatment for combat-related PTSD 36 men .68
Bohus et al., 2004 DBT for BPD patients 31 women .53
Chard, 2005 CPT for sexual abuse 28 women .92
Ross & Haley, 2004 Stabilization and trauma treatment in hospital program 46 adults .30
Rothbaum et al., 2005 PE or EMDR for recent rape victims 40 women .83 and .97
Sachsse et al., 2006 Dynamic trauma-focused treatment for BPD and CPTSD 75 women .28
Van Emmerik et al., 2008 CBT or writing task groups for ASD/PTSD patients 41 and 44 adults .18 and .39
Note. PTSD ⫽posttraumatic stress disorder; DBT ⫽dialectical behavior therapy; BPD ⫽borderline personality disorder; CPT ⫽cognitive processing
therapy; PE ⫽prolonged exposure; EMDR ⫽eye movement desensitization and reprocessing; CPTSD ⫽complex posttraumatic stress disorder; CBT ⫽
cognitive behavioral therapy; ASD ⫽acute stress disorder.
561
TRAUMA AND FANTASY MODEL OF DISSOCIATION
Over longer time spans, the TM prediction would be that trau-
matized individuals would be temporarily elevated in dissociative
symptoms as a group, and that these symptoms would diminish for
most individuals over time as the trauma becomes more integrated
into cognitive systems and trauma-related emotions (e.g., fear and
anxiety) dissipate. In studies in which participants were followed
after trauma—as in Carden˜a and Spiegel (1993); E. B. Carlson et
al. (2011); Feeny, Zoellner, Fitzgibbons, and Foa (2000); and
Feeny, Zoellner, and Foa (2000)—large and statistically signifi-
cant drops in dissociative symptom severity occur over time with-
out intervention in most individuals. Two to 10 days after trauma
exposure in E. B. Carlson et al., 40% of the sample showed
elevated levels of dissociation when compared with a normative
sample of adults with no prior trauma exposure. One week later,
39% still reported dissociation at elevated levels. At 2 months
postevent, only 27% of participants reported dissociative symptom
elevation. This pattern also fits the TM and not the FM prediction.
In summary, the increase in state dissociation after exposure to
high stressors or traumatic events and trauma reminders is consis-
tent with TM Prediction 3. Similarly, findings support the TM
prediction of the short-term decrease in dissociation (relative to
comparison groups) with trauma-relevant psychological or phar-
macological treatment and the long-term decrease in dissociation
over time. If dissociation were a stable outgrowth of fantasy
proneness and mild neurocognitive disturbance (cf. Giesbrecht et
al., 2008), such patterns would be much harder to explain. These
findings clearly support TM Prediction 3, that dissociation is
temporally related to trauma and trauma treatment.
Evidence for Prediction 4: Does Dissociation Show an
Increment Over Fantasy Proneness in the Prediction
of Trauma?
Both the TM and the FM predict a relationship between the
measures typically used for the dissociation and fantasy proneness
concepts, because both types of scales were developed from a
theoretical base that included an etiological role for psychological
absorption and trauma. Fantasy proneness is acknowledged to be a
“close cousin” of absorption by Geraerts, Merckelbach, Jelicic,
Smeets, and Van Heerden (2006, p. 1143). The authors of both of
the most commonly used fantasy proneness scales report that they
developed their measures from a theoretical framework that in-
cludes absorption (Merckelbach et al., 2001; S. C. Wilson &
Barber, 1983). Similarly, absorption items were purposely in-
cluded in the DES, the most commonly used dissociation scale
(Bernstein & Putnam, 1986). It is easy for theoreticians from all
perspectives to lose track of this history, reifying the scale totals,
and reporting as an independent and surprising finding that ab-
sorption correlates strongly with each measure.
Merckelbach et al.’s (2001) CEQ and S. C. Wilson and Barber’s
(1983) ICMI do correlate with dissociation (Merckelbach et al.,
2002; Pekala et al., 1999–2000; Rauschenberger & Lynn, 1995;
Waldo & Merritt, 2000), but the reason for the correlation is
unclear. Highly fantasy-prone individuals have been reported to be
diagnosed with dissociative disorders more often than low- or
medium-level fantasy-prone individuals (Rauschenberger & Lynn,
2002–2003). The inclusion of absorption within each scale type is
the most obvious explanation. Merckelbach et al., the developers
of the CEQ, also noted that there is overlap between the item
content of the CEQ and the DES. They suggested:
Two CEQ items (i.e., “I often confuse fantasies with real memories”
and “I sometimes feel that I have an out of body experience”) clearly
overlap with some DES items (e.g., “not sure whether one has done
something or only thought about it” and “feeling as though one’s body
is not one’s own,” respectively). (p. 989)
Such similar items would contribute to correlations between mea-
sures of fantasy proneness and dissociation.
Further, it is consistent with prior theory and research on fantasy
proneness scales that trauma is one cause, although not the sole
cause, of fantasy proneness. In early articles on the CEQ, Merck-
elbach et al. (2001) conceded there are different paths to fantasy
proneness, including coping with childhood adversity: “Other fan-
tasy prones,” they wrote, “reported a heightened frequency of
aversive childhood events. In these cases, a profound fantasy life
may have become a means to cope with or escape from negative
experiences” (p. 988). Rhue and Lynn (1987, p. 121), for instance,
noted that fantasy-prone participants reported “greater frequency
and severity of physical punishment, greater use of fantasy to
block the pain of punishment, more thoughts of revenge toward the
person who punished them, greater loneliness, and a preference for
punishing their own children less severely” than those lower in
fantasy proneness. Lynn and Rhue (1986) and S. C. Wilson and
Barber (1983) also reported that fantasizers acknowledged more
severe and more frequent childhood punishment. In keeping with
the TM hypothesis of use of fantasy as escape, fantasy proneness
is related to the five scales of the Childhood Trauma Questionnaire
(Pekala et al., 1999–2000).
Therefore, dissociation and fantasy proneness may correlate
spuriously in part through their common connection to trauma
history. Again, from the TM perspective, those who voluntarily
and (over time) involuntarily shift attention from stimuli that
trigger unwanted memories (dissociate) will also use other tech-
niques to escape from unwanted environments (such as voluntary
shifts of attention to internally generated images in the form of
fantasizing or daydreaming). A definitive answer to the question of
the etiology of this relationship awaits more sophisticated studies
that include all relevant variables. Particularly helpful would be
studies that track these relationships over time.
Although the relationship of fantasy proneness and dissociation
is not incompatible with either model, the FM does make a
prediction of the relative relationship of these variables to trauma
self-report. In the FM given by Merckelbach et al. (2002), and
replicated in Figure 1, a statistical prediction can be made that
fantasy proneness will produce an increment over dissociation in
the prediction of trauma self-reports, whereas dissociation will
produce no significant increment over fantasy proneness. Because
the TM posits a causal role for trauma in producing dissociation,
an increment for dissociation is predicted.
We were able to locate four studies with samples greater than 50
(to allow sufficient power) that included the three relevant corre-
lations allowing partial correlation to be computed. Support for the
TM contention (statistically significant partial correlation of
trauma and dissociation controlling for fantasy proneness) oc-
curred in all four studies: research by Merckelbach et al. (2002);
Pekala, Angelini, and Kumar (2001); Pekala et al. (1999–2000);
and Thomson, Keehn, and Gumpel (2009). Specifically, in each
562 DALENBERG ET AL.
case, fantasy proneness did relate to trauma history and dissocia-
tion, but trauma history did have an increment over fantasy prone-
ness in the equation predicting the DES. Dissociation does relate to
report of trauma history controlling for fantasy proneness.
Furthermore, the few studies on fantasy proneness in dissocia-
tive disordered samples do not indicate the strong elevations in
fantasy proneness that would be expected if their trauma histories
were entirely fantasized. Huntjens et al. (2006) found that DID
patients scored higher on fantasy proneness than controls and
nonclinical DID simulators. However, the DID mean score on the
CEQ (9.92) was very similar to means of male and female college
students reported in Merckelbach et al.’s (2001) psychometric
article on the CEQ (M⫽9.2, SD ⫽4.4, and M⫽8.7, SD ⫽4.0,
respectively). Using the ICMI, Levin, Sirof, Simeon, and Gural-
nick (2004) also found elevated levels of fantasy proneness in
patients with depersonalization disorder (DPD) compared with
nonsymptomatic controls. However, as Levin et al. wrote, the total
scores for the DPD group were well below typically used thresh-
olds for high fantasy proneness. The DPD mean was 14.7 (SD ⫽
7.3), which falls at the low end of the range for medium fantasy
proneness on this instrument (14–36 in Levin et al., 2004). These
findings in general support the TM prediction (Prediction 4) re-
garding the independent contribution of dissociation over fantasy
proneness in the prediction of trauma history.
Evidence for Prediction 5: Are Dissociative Research
Participants at High Risk for Suggestibility and False
Memory?
Research on suggestibility is also central to the FM contentions
about the dissociation–trauma connection. The controversial con-
tentions of the FM are not only that the dissociation and trauma
report connection is mediated by fantasy proneness, which appears
unfounded as discussed earlier, but also that dissociation produces
enhanced probability of confabulation of trauma memory itself.
Giesbrecht et al. (2008) repeatedly cited their concern that disso-
ciative individuals will overreport trauma on standardized ques-
tionnaires unless provided with a context that “discourages report-
ing of traumatic experiences” (p. 622). It seems ill-advised and
potentially harmful to discourage patients from reporting trauma
exposure due to fears of high rates of false report without strong
support for this hypothesis.
Suggestibility paradigms. In the standard FM argument of
dissociation as a risk factor for suggestibility, many nonequivalent
forms of suggestibility are mentioned and tested (Giesbrecht et al.,
2008; Merckelbach & Muris, 2001). To extend the range of studies
reported, all research with samples greater than 25 are presented in
Table 4. The best known are clustered under event suggestibility
studies, and represent forms of suggestion that include acceptance
of the false suggestion that one has seen or experienced an event.
In the nonautobiographical studies of this type, participants are
typically shown slides or read paragraphs, and pressed at a later
point to agree to a false statement about a slide seen or fact heard.
The Gudjonsson (1997) suggestibility paradigm is a standardized
form of this type of suggestibility. In this paradigm, participants
are read paragraphs and then (through social pressure or mislead-
ing questions) pushed toward acceptance of false statements about
the information heard. An overall suggestibility score, a yield
score (degree of acquiescence to leading questions), and a shift
score (the number of times the individual changed an answer in
response to interpersonal pressure) are then calculated.
The methodology in autobiographical event suggestibility stud-
ies is more varied. In studies typically referred to as “false mem-
ory” studies (e.g., Hyman & Billings, 1998), participants are told
that a knowledgeable person (typically the individual’s mother)
recalls an event in the person’s life. The dependent variable is the
degree to which the research participant appears to accept the truth
of this false memory. In misinformation studies, the dependent
variable is the same, but the procedures typically involve less
powerful suggestion (misleading questions, varying in terms of
source, number, and strength).
In source monitoring or source confusion studies, the task of the
participant is typically to discriminate between competing sources
for an alleged memory (e.g., whether information came from a
picture seen, a paragraph read, or a new story heard). Alterna-
tively, in the Deese–Roediger–McDermott (DRM) paradigm
(Deese, 1959; Roediger & McDermott, 1995), the participants read
a series of words that relate to an overarching nonpresented word
(e.g., read the words nap,doze, and dream—all words related to
the concept “sleep”). The dependent variable is whether the indi-
vidual recalls or falsely assents to seeing the nonpresented concept
word.
Finally, in the imagination inflation studies, participants imag-
ine a series of incidents and are asked about their feeling of
remembering the event, as opposed to merely knowing or believ-
ing that the event might have happened. The events are typically
plausible or known events from childhood.
The degree to which each of these paradigms is linked to a
general “suggestibility” trait is unknown, but sets of studies are
reviewed in turn as examples of suggestibility as defined within
the FM. Historically, false memory has been fairly loosely defined
in such paradigms (cf. DePrince, Allard, Oh, & Freyd, 2004).
Research testing general memory skills of dissociative individuals,
or errors on event memory tasks in the absence of suggestion, are
not considered as examples of suggestibility paradigms.
Nonautobiographical event suggestibility. Table 4 contains
data from eight studies with 10 samples investigating suggestibil-
ity for nonautobiographical events, all using the Gudjonsson meth-
odology, and the examination of suggestibility relationship with
dissociative experiences. The clinical samples—a small group of
anxious patients reported by Wolfradt and Meyer (1998) and the
larger mixed sample by Little (1996)—and the only abuse sample
(Schultz, Passmore, & Yoder, 2003) produced nonsignificant re-
sults. The weighted estimate for the correlation between dissoci-
ation and suggestibility in this category is .12. Further, the pattern
of correlations on the Gudjonsson subscales varied across the few
studies reporting statistically significant results. Wolfradt and
Meyer in their nonclinical sample found DES correlations with
both Shift and Yield scales; Merckelbach, Muris, Rassin, and
Horselenberg (2000) reported DES correlations with the Shift (but
not Yield) score; and Merckelbach, Muris, Wessel, and Van Ko-
ppen (1998) found correlations with the Yield (but not Shift) score.
Horselenberg et al. (2000) came to the conclusion that “the relation
between dissociative tendencies and memory distortions is not as
impressive as some authors have suggested” (p. 136), noting that
the few previous studies that had found positive associations had
significant methodological limitations. Gudjonsson (2003) himself
specifically noted with surprise the lack of consistent relationship
563
TRAUMA AND FANTASY MODEL OF DISSOCIATION
Table 4
Relationship of Dissociation and Suggestibility
Study Sample
Dissociation
measure Suggestibility task r
Nonautobiographical event suggestibility
Haraldsson, 2003 30 children, half with reincarnation beliefs CDC E; GSS .11
Hekkanen & McEvoy, 2002 111 UG DES E; acceptance of false suggestions regarding slides .00
Horselenberg et al., 2004 38 UG DES E; GSS .08
Merckelbach, Murrin, Rassin, &
Horselenberg, 2000 56 female UG DES S; GSS .37
ⴱⴱ
Merckelbach et al., 1998 40 UG in Study 1 DES S; GSS .32
ⴱ
Schultz et al., 2003 51 CM and 31 RM sexual abuse survivors DES E; GSS ⫺.06 and .06
Torrens, 2005 146 UG DES E; GSS .06
Wolfradt & Meyer, 1998 45 controls and 37 anxious patients DES E; GSS .34
ⴱ
and ⫺.04
Autobiographical events
Drivdahl & Zaragoza, 2001 149 UG DES E: errors after suggestion regarding staged event .07
Eisen & Carlson, 1998 130 UG DES E; agreement with misleading information after staged event .21
ⴱⴱ
Eisen, Morgan, & Mickes, 2002 111 UG DES E; agreement with misleading information after staged event .13
Eisen, Qin, et al., 2002 49 children CDC E; agreement with misleading information after abuse assessment .01
Horselenberg et al., 2004 38 UG DES E; false recognition of foils regarding autobiographical events .19
Horselenberg et al., 2003 34 UG DES E; internalization of false suggestion of performed action ⫺.30
Hyland, 2000 100 UG and community DES E: acceptance of false autobiographical event .16
Hyman & Billings, 1998 48 UG DES E: acceptance of false autobiographical event .48
ⴱⴱⴱ
Ost et al., 1997 35 UG DES E; confidence in suggested false or distorted, autobiographical events .45
ⴱⴱ
and .00
Porter et al., 2000 47 UG DES E; agreement to suggested false autobiographical memory .45
Porter et al., 2008 60 UG DES E; acceptance of false media event .27
ⴱ
Qin et al., 2008 119 UG and community adults DES E; acceptance of false autobiographical event .16
ⴱ
Spanos et al., 1999 117 UG DES E; acceptance of false memories of infancy .06
Source monitoring
Clancy et al., 2000 30 abused, 15 no memory, 15 RM, and 15
controls
DES E; false recognition in written DRM .31
ⴱⴱ
Geraerts et al., 2005 94 abuse survivors, 20 controls
a
DES E; false trauma word recall and recognition in written DRM .10 and .16
b
Jelinek et al., 2009 76 mixed clinical and control DES E; false recall and recognition in visual DRM .00 and .00
Little, 1996 160 mixed control and psychiatric DES E; false recognition of texts and slides seen .08 and .00
McNally et al., 2005 138 abused and 26 controls
a
DES E; nondiscrimination between words seen versus imagined .02 first test and ⫺.13
second test
Platt et al. 1998 73 UG DES E; false recall and recognition in written DRM ⫺.03 and .08
b
Qin et al., 2008 86 UG and community adults DES E; false recall in written DRM .00
Torrens, 2005 146 UG DES E; false recall and false recognition in written DRM .06 and ⫺.04
b
Wilkinson & Hyman, 1998 92 UG DES E; false recall and false recognition in written DRM ⫺.09 and .04
b
K. Wilson & French, 2006 100 UG DES E; acceptance of suggestion of seeing video of known event .26
ⴱⴱ
Winograd et al., 1998 42 UG DES E; false recognition of recalled and nonrecalled lists in written DRM .18 and .32
ⴱ
Zermatten et al., 2006 75 UG DES E; nondiscrimination of acts performed and imagined .25
ⴱ
Imagination inflation
Paddock et al., 1998 98 UG in Study 1 DES E; increased sense of remembering after guided imagination .32
ⴱⴱ
Paddock et al., 2000 143 UG DES E; increased sense of remembering after guided imagination .04
Wilkinson & Hyman, 1998 94 UG in Study 1, 92 UG in Study 2 DES E; increased sense of remembering after guided imagination .26
ⴱ
in Study 1 and
.11 in Study 2
Note. Study 1 and Study 2 are treated as separate in effect size calculation. Recognition used in recognition–recall paradigms. Values averaged if two figures are given. Sample: UG ⫽undergraduate;
CM ⫽continuous memory; RM ⫽recovered memory. Dissociation measure: CDC ⫽Child Dissociative Checklist; DES ⫽Dissociative Experiences Scale. Suggestibility task: E ⫽experiment or
group comparison design; GSS ⫽Gudjonsson Suggestibility Scale; S ⫽survey and correlational design; DRM ⫽Deese–Roediger–McDermott paradigm.
a
Includes “repressed memory” group (n⫽38).
b
This figure was used in average effect size calculations.
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
564 DALENBERG ET AL.
between dissociation and his measure of suggestibility. Thus, our
best estimate is that dissociation accounts for an unimpressive 1%
of the variance in nonautobiographical event suggestibility. Q, the
statistic used to determine heterogeneity of effect size, was non-
significant at 9.89. Thus, the average weighted effect size was a
good estimate of the full range of studies.
In the only published study examining interrogatory suggest-
ibility using the Gudjonsson procedure in a group of clinical
subjects with reported delayed recall of trauma memories (disso-
ciative amnesia), Leavitt (1997) found that “recovered memory”
patients scored lower on interrogatory suggestibility than the psy-
chiatric comparison group. Thus, the research to date indicates that
patients with delayed recall of trauma are actually less suggestible
in this paradigm than other psychiatric patients. The question of
suggestibility and its relationship to overreporting or underreport-
ing of trauma memories requires far more research with ecologi-
cally sound studies before broad conclusions can be made.
Autobiographical event suggestibility. Eleven studies fell in
the category of autobiographical event suggestibility. The mislead-
ing questions in the study varied in power, from the strong insis-
tence that the false event was true in Hyman and Billings (1998) to
the misleading questions or foils used in the majority of studies.
The overall weighted rin this case was .16, or 2% of the variance,
with a Qof 25.86 (p⬍.05), representing heterogeneity of effect
size.
The elevated correlation found by Hyman and Billings (1998)
and Ost, Fellows, and Bull (1997) might be explained by some of
the unusual features of the studies. In many senses, Hyman and
Billings’s work was an advance over other false memory studies in
that the false memory “event” chosen was unique and unlikely to
have occurred to any participant. Participants were told that they
had spilled a bowl of punch on a bride during a wedding reception
they attended when very young. Most importantly, the measure-
ment of false memory in this case was a continuous one, with
participants receiving credit (in the sense of a higher false memory
score) if they indicated that they were trying to remember the event
but failed to do so. Further credit was given if the participants
recalled remembering attending a wedding (which could well have
been true) but not the event of spilling the punch. Thus, the
distinction between dissociation as a predictor of recovery of
accurate memory and dissociation as a predictor of false memory
cannot be made with certainty. Similarly, Ost et al. reported only
that the dissociative individuals were more confident in false
events presented as true than nondissociative individuals, but did
not report the level of confidence achieved.
Several false memory studies used the DES-C (Ost, Foster,
Costall, & Bull, 2005; Ost, Granhag, Udell, & Roos af Hjelmsa¨ter,
2008), the revision of the DES by Wright and Loftus (1999). These
studies are not cited in Table 4, given the lack of evidence for the
DES-C as a valid measure of dissociation. Although the DES-C
literature was not fully reviewed, it does contain both nonsignifi-
cant correlations between dissociation and suggestibility
(Horselenberg et al., 2006) and positive correlations in a similar
range to those in Table 4 (e.g., Ost et al., 2005).
Source monitoring studies. The source monitoring section
contains 12 studies, most of which use the DRM paradigm. The
overall weighted correlation here is .08, accounting for less than
1% of the variance. The Qvalue of 18.64 is nonsignificant,
showing homogeneity of effect size. Only three of the 12 studies
reached statistical significance. The statistically significant studies
do not show any particular shared similarities, and accounted for
6%–10% of the variance in their samples. It seems quite plausible
here to conclude that source monitoring is a complex and multi-
determined concept, and not strongly or uniquely related to disso-
ciation. Note that the methodology here, unlike the event suggest-
ibility paradigms, requires the individual to internally generate the
false idea or memory. Thus, arguably this methodology is appli-
cable to the situation of internally manufactured memories of
abuse.
Further research might include more trauma-relevant variables,
such as degree of trauma exposure or level of PTSD symptom-
atology. Zoellner, Foa, Brigidi, and Przeworski (2000), using the
DRM paradigm, found that traumatized participants with and
without PTSD generated more false recalls of critical nonpresented
words than nontraumatized participants. False recall was related to
PTSD severity. More recently, Brennen, Dybdahl, and Kapidzˇic´
(2007) found that war-traumatized participants with PTSD were
even more susceptible to false intrusions of war-related material
than traumatized individuals who did not have PTSD. Again, it
should be emphasized that (a) these effects occurred with war
veterans, acknowledged to be genuinely traumatized, and (b) in-
creases were found in the recognition of trauma-related words, not
in the fabrication of battles that did not occur. One could not
logically argue that such data cause one to doubt the soldiers’
report that they had been exposed to grave danger or that they had
been traumatized, as the FM interpretation of the data would
suggest. Rather, the data are more compatible with the hypothesis
that the intrusions represent an aspect of overinclusive responding
and fear stimulus generalization, other potential costs of trauma
exposure.
Imagination inflation studies. The four imagination infla-
tion results have the best base rate of significant correlations with
dissociation (two of four samples). The weighted ris low (.17),
accounting for 2.89% of the variance (with a nonsignificant Q
value, 6.03), but methodological differences in the samples point
to possible further directions for research. Paddock et al. (1998),
for instance, found a statistically significant correlation between
dissociation and the robust imagination inflation effect in their
young adult college sample. However, in their factory worker
sample, arguably more likely than college samples to include
individuals whose cognitive capacities are more impaired by
trauma, the imagination inflation effect itself disappeared. Pad-
dock et al. therefore did not attempt further correlational analyses.
Overall findings from suggestibility studies. Given the
strong statements made within the FM regarding suggestibility and
dissociation, including the more extreme statements making the
claim that all trauma-related dissociative disorders may be artifacts
of suggestibility (Piper & Merskey, 2004), or warnings that asking
a routine screening question might produce false memory (Gies-
brecht et al., 2008), it is surprising to see average weighted effect
sizes between dissociation and suggestibility that are so small
(1%–3% of variance accounted for) across categories. Looking
through the whole of Table 4 at the most statistically significant
studies, with the caveat that much more work needs to be done,
one would see that the dissociative individual may be at risk for
false memory in situations that directly or indirectly encourage
imagination and rumination over a believed-in but nonremembered
event. This is the methodology of the imagination inflation studies,
565
TRAUMA AND FANTASY MODEL OF DISSOCIATION
but also may occur under event suggestibility paradigms that
convince the participant of the truth of a nonremembered event and
encourage general attempts to remember (Hyman & Billings,
1998; Ost et al., 1997).
A challenge for suggestibility theorists is the differentiation of
acquiescence and false memory. If the task is to remember an
allegedly true event when asked by an authority figure, the
achievement-oriented undergraduate might claim memory to show
intelligence, just as the traumatized dissociating individual might
claim memory out of fear-based deference, or resist due to fear or
anger-based distrust of authority. One study provides some indi-
cation that those with higher levels of dissociative symptoms may
be motivated by fear or distrust of authority to acquiesce when
prompted. A study of psychiatric inpatients found that dissociative
symptoms were moderately related to measures of dependent
personality and antisocial personality, among others (Modestin,
Ebner, Junghan, & Erni, 1996).
Dissociation thus may interact with other variables to form
groups at high or low risk for suggestibility, false memory forma-
tion, or acquiescence to false statements. Eisen, Goodman, Qin,
Davis, and Crayton (2007), for instance, showed differing predic-
tors of memory errors for their high- and low-dissociating groups
within a child maltreatment sample. For low-dissociating children,
increase in stress, posttraumatic symptoms, and cortisol level
predicted better memory scores in general. For high-dissociating
children, greater increase in cortisol from baseline to after the
anogenital exam and higher trauma symptom report predicted
memory error commissions (r⫽.28, p⬍.05), and greater increase
in cortisol and lower trauma symptom report predicted omissions
(r⫽.25, p⬍.05). Malek (2010), in a recently completed study,
found that dissociation related to the report of a significant false
memory experience (defined as a personally significant memory
later proven to be inaccurate) only in the context of elevations in
executive dysfunction scores. Future research might attempt to
differentiate dissociative from fantasy-prone groups.
The failure of TM theorists, for the most part, to include sug-
gestibility among their panoply of routinely studied variables, and
the parallel failure of FM theorists, again for the most part, to
consider trauma-relevant covariates and to study traumatized
groups, leaves much territory for collaborative endeavors. Given
the low effect size averages found by considering dissociation as a
main effect across multiple domains in predicting suggestibility,
interaction studies (together with longitudinal studies) appear to be
the appropriate next step. In the meantime, the current evidence
supports the TM prediction regarding suggestibility. The literature
on dissociation and suggestibility supports the existence of a weak
and inconsistent relationship between the two concepts.
Evidence for Prediction 6: Is Dissociation Related to
Fragmentation, Omission, and Narrative Cohesion?
The TM and FM differ in their position on the relationship of
fragmentation of memory and memory loss as it relates to disso-
ciation. Writing from the FM perspective, Giesbrecht et al. (2008)
argued that there are no studies linking dissociation to objective
fragmentation, but two student sample studies (Kindt & Van den
Hout, 2003; Kindt, Van den Hout, & Buck, 2005) link dissociation
to subjective fragmentation (the belief that one’s memory is frag-
mented). The TM posits a relationship between fragmentation of
memory and the experience of trauma, a relationship that would be
stronger for dissociative than for nondissociative individuals. First,
fragmentation does appear to relate to traumatic symptoms in
general, as shown by Foa, Molnar, and Cashman (1995) in work
with rape survivors and replicated by Van Minnen, Wessel, Dijk-
stra, and Roelofs (2002) with a group varying in trauma type. Both
showed covariation in improvement in trauma symptoms and
reduction in fragmentation over time (r⫽.73 for Foa et al., N⫽
14; effect size for Van Minnen et al., N⫽20, not reported). Both
used an objective measure of narrative fragmentation. In a larger,
well-controlled study by Silva (2006), the free narratives of 98
children were examined as they discussed their general likes and
dislikes during a rapport phase of a child abuse interview and as
they answered questions about their alleged abuse in the abuse
phase. For the children alleging severe abuse, narrative fragmen-
tation rose significantly (effect size for rapport vs. abuse phase:
r⫽.37, p⬍.01), with no such relationship occurring in children
alleging single exposure abuse by a nonattachment figure (r⫽.14,
p⬎.05).
Kindt et al. (2005) used peritraumatic rather than state or trait
dissociation measures in their research finding of statistically
significant relationships between subjective fragmentation and dis-
sociation. The trait measure used (the DES) reportedly did not
correlate with either subjective or objective measures of fragmen-
tation, although effect sizes were not given. Using a DPD group
rather than a nonclinical college population, Giesbrecht, Merck-
elbach, Van Oorsouw, and Simeon (2010) found differences in
both objective (r⫽.39, p⬍.05) and subjective (r⫽.51, p⬍.01)
fragmentation measures, despite using an emotional film clip as
the provocation.
A more direct test was provided by Halligan, Michael, Clark,
and Ehlers (2003), who assessed self-reported disorganization,
experimenter-rated disorganization, and an objective measure of
narrative disorganization in both retrospective (n⫽81) and pro-
spective samples (n⫽73) of assault victims. All three measures
related to dissociation in both samples (r⫽.24–.48, all ps⬍.05).
Thus, the only study of dissociative disordered clients, the only
retrospective study of trauma victims, and the only prospective
study of trauma victims that were located all support Prediction 6,
the TM hypothesis of the correlation of dissociation and objective
fragmentation of memory (see Huntjens, Dorahy, & Van Wees, in
press).
Evidence for Prediction 7: Is the Recovered Memory
Phenomena a Product of Dissociation or Fantasy?
Report of recovered memory or traumatic amnesia as fan-
tasy. The hypothesis of confabulation as a primary source of
recovered memory of trauma after dissociative amnesia (see Lof-
tus & Ketcham, 1994) must rely on evidence that recovered
memory victims are less likely (relative to those with continuous
memories) to be authentic abuse victims. Methodological chal-
lenges to such research are plentiful, but it appears that the current
evidence supports the TM rather than the FM view. Equivalent
accuracy of recovered and continuous memories of child trauma
was reported by Williams (1995), using hospital records (e.g., of
genital or anal injury) as the criterion, and by Dalenberg (1996),
using records combined with perpetrator confessions, both objec-
tive measures of accuracy. In a volunteer sample, Geraerts et al.
566 DALENBERG ET AL.
(2007) found spontaneously recovered memories to be similarly
likely to have corroboration (37%) when compared with continu-
ous memories (corroborated in 45%). However, memories recov-
ered in therapy, which represent a small proportion of the total
recovered memory reports (Eliott, 1997; Wilsnack, Wonderlich,
Kristjanson, Vogeltanz-Holm, & Wilsnack, 2002), were never
corroborated in Geraerts small sample (N⫽16).
Longitudinal studies also support the TM. Mechanic, Resick,
and Griffin’s (1998) study of amnesia postrape found that 37% of
assaulted women reported some degree of amnesia at the 2-week
point. At the 3-month marker, this number had dropped to 16%.
This is the pattern to be expected for a traumatic reaction that is
resolving for a portion of the participants. The opposite pattern
would occur if involved therapists were creating the illusion of
amnesia over time.
Experimental studies have achieved similar results, although
they have also been criticized on the basis of ecological validity
(Freyd & Gleaves, 1996). If recovered memory were reliably equal
to confabulated memory, and particularly if dissociative amnesia
were a myth (Loftus & Ketcham, 1994), then one would expect
that the negative-event memories recovered in experimental para-
digms would be more often false than true. This is not the case. In
Hyman and Billings’s (1998) study, for instance, participants were
asked to attempt to recall actual childhood events (contributed by
their parents) or false events constructed by the researcher (such as
spilling a punch bowl on the wedding party at a reception).
Participants initially denied 26% (of 218) true memories and 97%
(of 66) false memories. Given that it is unlikely that the partici-
pants confabulated a complex memory on the spot, the small
number of initial agreements in such studies are likely to be
responses to social demand rather than memories. Thus, the typical
dependent variable in false memory research is the number of false
“recoveries” after the research participants are sent home to at-
tempt retrieval of their lost memories. In Hyman and Billings, the
participants subsequently recalled 25% of the false events and 44%
of the true events. The comparable percentages in Hyman and
Pentland (1996) are 65% of the initially denied true memories and
26% of the initially denied false memories. Thus, in experimental
false memory studies, recovered memories are predominantly ac-
curate, but should be questioned (just as perceived continuous
memory should be questioned) in situations of strong suggestion
(Geraerts et al., 2007). Across all samples—abused or nonabused,
clinical, nonclinical, and experimental—it has been found that (a)
recovered memories and continuous memories were equally accu-
rate and (b) both recovered and continuous memories of trauma are
more likely to be true than false (cf. Dalenberg, 2006).
Such results should not be taken to contradict the theory that
susceptible individuals could be pushed into acceptance of a false
statement of trauma (or even create a false memory of trauma)
after repeated exposure to suggestion. The “retractor” group—
individuals who now state that they were led by suggestion to a
false memory—is understudied, but there is no current evidence
that this group is more likely to be dissociative, or that the
individuals’ retractions are more likely true or less subject to
suggestive influence. As one FM review conceded, current fantasy
proneness studies should not lead to an implication that patients
high on these constructs would create “wildly inaccurate” re-
sponses to self-report measures of trauma (Merckelbach et al.,
2002, p. 703).
The corroboration issue could also be applied to the verification
of dissociative phenomenology itself. Is there evidence that disso-
ciative symptoms are actually occurring, rather than being a man-
ifestation of malingering or misunderstanding of more common
cognitive states? Gleaves, Hernandez, and Warner (1999) found in
a survey of clinicians that therapists for 446 DID patients reported
evidence for corroboration of dissociative symptoms from family
or psychiatric records in 67% of patients prior to diagnosis. Be-
cause this type of evidence is likely to be discounted by FM
theorists as due to the credulity of therapists, it should be noted that
observers using other documentation have reached similar conclu-
sions. Somer and Weiner (1996) used just such a methodology,
comparing the adolescent diaries of two DID patients and four
controls. The diaries were rated for overt mention of dissociative
symptoms (e.g., derealization, depersonalization, amnesia). The
DID patient diaries contained over 6 times the frequency of dis-
sociative contents compared with the control diaries. In a similar
study with a large sample (N⫽126), Bagley, Rodberg, Wellings,
Moosa-Mitha, and Young (1995) identified dissociative traits that
could be behaviorally rated as present by reviewers of child
welfare department records. The number of traits rated as present
correlated .67 with the ADES in the 55 adolescents who had been
given this measure.
Dissociation and traumatic amnesia. Dissociation has been
correlated with the presence of traumatic amnesia in a number of
settings and methodologies. The DES correlates with the presence
of DID, a disorder that requires amnesia for Diagnostic and
Statistical Manual of Mental Disorders (fourth edition, text re-
vised) diagnosis, in mixed psychiatric samples with high effect
sizes (r⫽.69–.86; Dale, Berg, Eldin, Odegard, & Holte, 2009;
Dorahy, Irwin, & Middleton, 2002; Gleaves, Eberenz, Warner, &
Fine, 1995; Martı´nez-Taboas, 1995; Scroppo, Drob, Weinberger,
& Eagle, 1998). In research samples of abused adults who have not
been diagnosed with psychiatric disorders, dissociation still typi-
cally differentiates between those with continuous and those with
recovered memories (Geraerts, Smeets, Jelicic, Van Heerden, &
Merckelbach, 2005; McNally, Clancy, Schachter, & Pitman, 2000;
Melchert, 1999; Palesh, 2001). McNally, Clancy, Barrett, and
Parker (2005) presented the only null finding that we could locate.
A few laboratory paradigms exist to support these cross-
sectional findings. Geraerts, McNally, Jelicic, Merckelbach, and
Raymaekers (2008) found that those who had spontaneously re-
ported recovered memories of abuse were more successful in a
thought suppression paradigm. Recovered memory survivors were
better able to suppress thoughts than continuous memory survi-
vors, and experienced less rebound of the targeted thoughts when
the suppression period was over. Directed forgetting paradigms,
which use single words as the target to be recalled or forgotten,
produce more variable results (DePrince & Freyd, 2001, 2004;
Devilly et al., 2007). The distinction between these paradigms
most likely rests on the issues raised by Geraerts and McNally
(2008), discussing the directed forgetting results. First, a complex
thought, or a complex experience such as abuse, should not be
equated with the ability to remember or forget a word such as
incest. Thus, the ecological validity and generalizability of this
paradigm to actual traumatic memories is unclear. Second, forget-
ting abuse memories may occur over a period, sometimes follow-
ing an initial conscious strategy of thought suppression (Koutstaal
567
TRAUMA AND FANTASY MODEL OF DISSOCIATION
& Schacter, 1997), rather than immediately following the presen-
tation of trauma-related words in the directed forgetting paradigm.
Interidentity amnesia studies. Interidentity amnesia in DID
is a separate issue from that of dissociative amnesia in general.
Authors from both TM and FM positions, including several of the
authors of this review, have contributed to the general finding that
implicit memories often cross dissociative identity barriers.
Interidentity amnesia has been studied as a paradigm for mem-
ory in DID since the late 19th and early 20th centuries (Prince &
Peterson, 1908; see Dorahy, 2001). With renewed interest in
multiple personality disorder and DID, this phenomenon has been
examined to attempt to understand the nature of memory and
amnesia in DID, often with contradictory findings (Eich, Macau-
ley, Loewenstein, & Dihle, 1997). In a series of studies designed
to overcome these contradictions, Huntjens and others (Huntjens,
2003; Huntjens, Peters, Woertman, Van der Hart, & Postma, 2007)
compared DID patients reporting mutually amnestic identities with
simulator and normal controls. Studies included tests of neutral
episodic information, perceptual and conceptual priming, proce-
dural memory, transfer of trauma-related words, and stimulus
valence as shown by affective priming. These researchers reported
no objective evidence of interidentity amnesia in any of these
studies. Huntjens (2003) concluded that dissociative amnesia in
DID may have more to do with subjective appraisal and
“metamemory” than actual lack of accessibility of memory be-
tween alternate identities.
Despite the amount of effort put into these studies, they have
limitations. First, the notion of relatively stable, fixed “two-way”
amnestic identities is based in the classical notion of DID as a
small set of relatively unchanging, structured “personalities” with
separate memory subsystems. This review is not the place to detail
the TM-based view of the phenomenology of DID. Suffice it to say
that the TM views DID as a posttraumatic developmental disorder
with a relatively dynamic self-state system derived from a variety
of developing intrapsychic, interpersonal, and psychosocial needs
over time, and a phenomenology usually based in, overlap, inter-
ference, intrusion, and shifting (not simply switching) among
personality states (Dell, 2006; Putnam, 1997). Further, this phe-
nomenological model contrasts with the classical notion of well-
defined identities with characteristics that can be reliably repro-
duced across clinical interviews and research trials (Dell, 2006;
Putnam, 1997; Putnam, Zahn, & Post, 1990). Proponents of the
TM—and, for that matter, proponents of the FM—do not take at
face value DID identities’ prevalent beliefs that they actually are
“real people” with varying demographic and psychological char-
acteristics, including differing ages, genders, etc. Nor would pro-
ponents of either model take at face value other common beliefs
that alternate identities are animals, mythical beings, internalized
“outside” people, demons, or omniscient beings. Therefore, it is
unclear why claims of two-way amnesia between identities should
also be accepted at face value preferentially by either set of model
theorists.
Thus, Forrest (1999, 2001), in a study of explicit memory in
identities claiming coconsciousness, or shared memory, found
evidence of interidentity amnesia, compared with normal and
simulating controls, despite the identities’ beliefs in their cocon-
sciousness. In additional support of the notion that alternate iden-
tities may not accurately assess their own subjective psychological
characteristics, Loewenstein, Hamilton, Alagna, Reid, and deVries
(1987) found, in a case study of DID using experiential sampling
techniques, that alternate identity self-reports of personality char-
acteristics were discrepant with objective data provided by rating
scales filled out in real time.
Elzinga, Phaf, Ardon, and Van Dyck (2003), in a directed
forgetting paradigm between subjectively amnestic DID identities,
found evidence supporting explicit memory disruption in a state-
dependent fashion between alternate identities, indicating not only
a possible encoding problem but also a retrieval inhibition between
identities. In addition, this disruption was partial, not complete.
The latter finding was thought to indicate that amnesia between
states is not rigidly compartmentalized, whatever their subjective
experience, consistent with the TM-based understanding of the
phenomenology of DID, where access to memory between iden-
tities may vary in complex ways, depending on a variety of factors.
Finally, the interidentity amnesia studies did not test autobio-
graphical memory in DID, presumably the type of memory most
importantly affected in these patients. Notwithstanding method-
ological limitations (e.g., low power, lack of controls, confirmation
biases), individual case studies using autobiographical memory
paradigms have found marked alterations in autobiographical
memory in DID, for both global memory of life history and
memory between identities (Bryant, 1995; Schacter, Kihlstrom,
Kihlstrom, & Berren, 1989). In addition, it is no longer controver-
sial that material can influence memory and behavior without
conscious processing or conscious awareness, and that conscious
and unconscious awareness have measurably different neural cor-
relates (e.g., Morris, Öhman, & Dolan, 1998).
Neurobiological studies are beginning to elucidate aspects of
autobiographical memory functioning in DID. For example,
Reinders et al. (2003, 2006), using a positron emission tomography
(PET) scan regional cerebral blood flow (rCBF) paradigm, as well
as measurement of autonomic functioning, assessed the reactions
of different DID alternate identities to personal trauma scripts.
Alternate identity pairs were said to experience trauma memory as
either part of personal (i.e., autobiographical) memory (traumatic
identity state) or not part of personal memory (neutral identity
state; Van der Hart, Nijenhuis, & Steele, 2006). In brief, analysis
of rCBF regional activation patterns and autonomic function
showed that the identity having self-referential understanding of
the trauma experienced the trauma script as an emotionally un-
pleasant autobiographical memory. This included activation of
areas such as the amygdala and insula with associated autonomic
increases. The neutral identity state appeared not to subjectively
experience the trauma scripts as personal autobiographical mem-
ory. These identity states showed a different rCBF and autonomic
activation pattern similar to DPD and dissociative PTSD patients
with medial prefrontal cortex inhibition of emotional and brain
association areas, and little or no autonomic activation (Lanius et
al., 2010).
We do not disagree that DID is in part a disorder of self-
understanding. Clearly those with DID have the inaccurate idea
that they are more than one person. However, this inaccurate belief
or perception is not evidence for the inherent invalidity of the
patients’ psychopathology, just as delusions of those with psy-
chotic disorders are not indicators that they do not have a psychi-
atric disorder. The psychotic mind may develop delusional beliefs
more easily, require less evidence for belief generation or main-
tenance. Similarly, some fundamental mechanism of dissociative
568 DALENBERG ET AL.
disorders may cause and maintain dissociative self-schemata (e.g.,
the idea that one is not a coherent self). High dissociativity on
standardized measures is a characteristic of DID patients that is
found across virtually all studies of these patients. Inconsistent
access to autobiographical information is another. Neither of these
features is identical to interidentity amnesia.
In summary, the findings from studies of traumatic amnesia,
dissociative amnesia, and recovered memory are not consistent
with the FM. However, amnestic phenomena are related to
dissociation across a variety of clinical manifestations (e.g.,
DID, dissociative amnesia, and recovered memory), supporting
Prediction 7.
Evidence for Prediction 8: Can Biological Studies
Inform the Debate?
Both FM and TM theorists agree that biological research might
be informative for the understanding of dissociation. However,
biological studies published to date have not been designed ex-
plicitly with the goal of differentiating predictions emanating from
the FM and TM. Accordingly, the relevance of biological studies
to these models can only be determined post hoc. This stated, most
biological studies, particularly those examining dissociative symp-
toms in individuals with PTSD, have assumed that trauma plays an
etiological role and have used fear-relevant paradigms and path-
ways as a foundation. In contrast, we are not aware of research
examining biological underpinnings for fantasy proneness in re-
portedly traumatized persons. Table 5 includes a summary of
relevant psychobiological studies of dissociation.
Genetic Studies
With the exception of Waller and Ross (1997), who did not
identify a genetic contribution to dissociative symptoms, twin
studies suggest that heritability estimates for dissociative symp-
toms approximate 50%–60% (Becker-Blease et al., 2004; Jang,
Paris, Zweig-Frank, & Livesley, 1998; Pieper, Out, Bakermans-
Kranenburg, & Van IJzendoorn, 2011). By contrast, research sug-
gests that shared environmental factors explain a negligible
amount of variance in dissociative symptoms as compared with
nonshared environmental factors that do contribute to a dissocia-
tive diathesis (Becker-Blease et al., 2004). Accepting that disso-
ciative symptoms have a sizable genetic loading, Geraerts et al.
(2006) argued that it is difficult to conceptualize dissociation as a
defensive reaction to traumatic experience. Thus, they stated, the
high absorption scores of CSA survivors provide evidence for the
presence of pseudomemories, given that absorption also carries
some genetic loading (roughly 50%; Tellegen et al., 1988). Nev-
ertheless, the TM does not posit negligible effects for genetic
factors, but instead hypothesizes a significant role for trauma
exposure.
Genetic factors may act as vulnerabilities for pathological dis-
sociation specifically in the context of trauma exposure. Pieper et
al. (2011) found that individuals homozygous for the short (SS)
5-HTTLPR allele evidenced greater pathological dissociative
symptoms particularly when reporting a trauma history in the
presence of depressive symptoms; trauma exposure did not interact
with genotypic variants of 5-HTTLPR in the prediction of non-
pathological dissociation. One interpretation of such findings is
that SS carriers of the 5-HTTLPR gene are at increased risk of
developing a combined depressive–dissociative syndrome only in
the context of trauma exposure. In comparison, exposure to trau-
matic events in SS carriers is unlikely to be associated with
pathological dissociation if depression is not present as well. In a
related finding, Lochner et al. (2007) concluded that childhood
trauma histories were predictive of increased pathological disso-
ciation in individuals with obsessive-compulsive disorder only in
SS carriers of the 5-HTTLPR gene. Furthermore, dissociative
symptoms were predicted by the interaction of genes related to
dopaminergic function and childhood trauma history; dissociation
was highest in individuals with the Val/Val genotype of the COMT
gene (functional catechol-O-methyltransferase Val158Met poly-
morphism) who were also exposed to childhood trauma (Savitz et
al., 2008). Although this research is only in its infancy, preliminary
evidence thus far suggests that genetic factors alone are insuffi-
cient to account for variability in dissociative symptoms. Instead,
gene by environment (i.e., trauma exposure) interactions may
better explain dissociative phenotypes.
Psychophysiology and Neuroendocrine Response to
Stress
Studies have examined the psychophysiological and neuroen-
docrinological correlates of dissociation. However, most studies
fail to differentiate between peritraumatic dissociation, state dis-
sociation, and trait dissociation, or differentiate state dissociation
from measures of general distress or negative affect (cf. Morgan et
al., 2002). Higher (e.g., Hetzel-Riggin, 2010; Hetzel-Riggin &
Wilber, 2010; Ladwig et al., 2002; Nixon, Bryant, Moulds, Felm-
ingham, & Mastrodomenico, 2005), lower (e.g., Griffin, Resick, &
Mechanic, 2007; Pole et al., 2005), and null (e.g., Kaufman et al.,
2002) associations with cardiovascular (heart rate) and autonomic
(e.g., galvanic skin conductance response [SCR]) response to
trauma reminders have been observed in individuals differing in
peritraumatic dissociative experiences. In comparison, the major-
ity of studies of state (Lanius et al., 2005, 2002) and trait (Bo-
nanno, Noll, Putnam, O’Neill, & Trickett, 2003; Koopman et al.,
2004; Sierra et al., 2002; Sierra, Senior, Phillips, & David, 2006;
Simeon, Yehuda, Knutelska, & Schmeidler, 2008) dissociation
have shown decreased heart rate and SCR reactivity during stress,
emotional processing, and/or symptom challenge, although in-
creased physiological reactivity has also been observed (Gies-
brecht, Geraerts, & Merckelbach, 2007).
Few studies of trait dissociation have examined the degree of
state dissociation experienced by participants during psychophys-
iological measurements. However, Hauschildt, Peters, Moritz, and
Jelinek (2011) found that both trait and state dissociation were
associated with reduced heart rate variability when participants
viewed trauma-relevant videos. Surprisingly few clinical studies
have examined dissociative disorders other than DPD.
However, an elegant study by Reinders et al. (2003, 2006)
showed that in ego states defined by a personal autobiographical
experience of traumatic memories, in contrast to ego states that fail
to label such memories as autobiographical, trauma script-driven
imagery was associated with greater heart rate, higher systolic
blood pressure, and lower heart rate variability in individuals with
DID. Future studies of psychophysiological response to symptom
provocation in dissociative disorders should examine degree of
569
TRAUMA AND FANTASY MODEL OF DISSOCIATION
Table 5
Review of Psychobiological Studies of Dissociation
Study Sample description Measures and method
Measures of dissociation
and diagnosis Results of interest in brief
Heritability and genetics
Becker-Blease et al., 2004 75 unrelated adoptive siblings,
91 related siblings, 218 MZ
twins, and 173 DZ twins
Twin study of h
2
of trait dissociation;
parent and teacher rated trait
dissociation
6 trait dissociative items
from CBCL
h
2
of dissociation ⫽.60, c
2
⫽.00
Jang et al., 1998 General population sample of
177 MZ twins and 152 DZ
twins
Twin study of h
2
of trait dissociation DES-T, DES h
2
of DES-T and DES ⫽.48 and .55, respectively; c
2
⫽
.00 for both DES-T and DES
Lochner et al., 2007 83 OCD participants Genetic study of 5-HTTLPR, childhood
trauma history and trait dissociation
DES-T, DES, CTQ Childhood trauma and 5-HTT genotype predicted 22% of
the variance in DES-T scores. Moderate correlations
between CTQ and DES-T scores with SS genotype;
association nonsignificant with LL genotype.
Pieper et al., 2011 184 twin pairs Twin study of h
2
of trait dissociation,
5-HTTLPR, trauma history and trait
dissociation
DES-T, DES h
2
of DES-T and DES ⫽.43 and .44, respectively; c
2
⫽
.00 for both DES-T and DES. Participants with the SS
genotype who also had high depressive symptoms and
trauma had highest DES-T scores.
Savitz et al., 2008 178 individuals from 35
families bipolar proband
and one additional first-
degree relative with bipolar
disorder
Study of genes related to COMT,
polymorphism, trauma and trait
dissociation
DES DES scores predicted by the interaction of COMT genotype
with childhood trauma; DES scores highest in individuals
with the Val/Val genotype with childhood trauma
Tellegen et al., 1988 217 MZ and 114 DZ adult
twins reared together and
44 MZ and 27 DZ adult
twins reared apart
Twin study of h
2
of trait absorption MPQ h
2
of absorption ⫽.50, c
2
⫽.03
Waller & Ross, 1997 280 MZ and 148 DZ twins Twin study of h
2
of trait dissociation DES-T h
2
of DES-T scores ⫽.00; c
2
⫽.45
Psychophysiology (trait dissociation)
Bonanno et al., 2003 103 women, 48 with
documented CSA
HR while participants spoke of the
“most distressing event” was
contrasted with baseline HR
ADES-T ADES-T scores correlated negatively with increases in HR
(r⫽⫺.24) and facial expressions (r⫽⫺.21) during
discussion of distressing events (relative to baseline).
Giesbrecht et al., 2007 62 undergraduates Viewed a provocative video while SCR
was measured
DES DES correlated with SCRs to the video (r⫽.34); fantasy
proneness showed null effects (r⫽.18, ns).
Hauschildt et al., 2011 26 trauma exposed with
PTSD, 26 trauma exposed
without PTSD, 18
nontrauma controls
HRV recorded during videos of varying
emotional valence
DES, DSS, PDEQ Within trauma groups, higher DES (rⱕ⫺.24 ) and DSS (r
ⱕ⫺.20) related with lower HRV, whereas PDEQ was
not correlated with either.
Koopman et al., 2004 41 delinquent adolescents Randomly assigned to either talk about
their most stressful life experience or
talk freely while HR was measured
SCID-D Lower HR was associated with higher derealization (r⫽
⫺.29) and higher identity alteration (r⫽⫺.33).
Sierra et al., 2002 15 DD patients, 15 HC, 11
anxiety controls
SCRs to pleasant versus unpleasant
versus neutral pictures
CDS DD patients showed reduced and prolonged SCR to
unpleasant pictures.
Sierra et al., 2006 16 patients with DD, 15 HC,
15 anxiety disorder controls
SCRs to viewing facial expressions of
happiness and disgust
PSE, CDS DD group had attenuated SCR to disgust stimuli in
comparison with HC (d⫽.98). CDS scores in the DD
group correlated with SCR to disgust (r⫽⫺.40) and
happy faces (r⫽.20).
570 DALENBERG ET AL.
Table 5 (continued)
Study Sample description Measures and method
Measures of dissociation
and diagnosis Results of interest in brief
Neuroendocrinology
State dissociation
Morgan et al., 2001 44 healthy male soldiers NE, EPI, NPY, and plasma/salivary
cortisol assessed before, during, and
after exposure to physical and mental
stress
CADSS Increased CADSS associated with decreased cortisol during
stress (r⫽⫺.49) and increased cortisol 24 hr
subsequently (r⫽⫺.46)
Trait dissociation
Koopman et al., 2003 49 women with PTSD related
to CSA
Five salivary cortisol samples collected
at the beginning, end, and 1, 24, and
48 hr after an interview about stress
and childhood trauma
SASRQ High dissociators had elevated salivary cortisol 24 hr after
the interview.
Schechter et al., 2004 41 mothers of young children,
the mothers of whom had
PTSD from interpersonal
trauma
Maternal perception of children and
maternal behavior assessed during
separation–reunion sequences.
Salivary cortisol collected from
mothers before and 30 min after
reunions
HADSI Baseline salivary cortisol negatively correlated with severity
of trait dissociation (r⫽⫺.31), but cortisol reactivity to
separation–reunion was nonsignificant correlated with
severity of dissociation (r⫽.15, ns)
Simeon et al., 2007 46 DD without PTSD, 35
PTSD, 58 HC
24-hr urine and serial blood samples
collected before and after DST and
TSST
DES DD had higher basal cortisol in urine (but not plasma)
compared with HC. DD group had greater resistance to
and faster escape from DST. No differences in cortisol
reactivity. DES correlated negatively with peak cortisol
reactivity to the TSST (r⫽⫺.43).
Simeon et al., 2008 21 high exposure and 10
nontrauma HC without
major exposure to the
World Trade Center attack
24-hr urine cortisol after DST. During
TSST, plasma cortisol changes, HR,
and BP assessed during rest and at
peak response
DES DES negatively correlated with plasma cortisol levels at
08.00 h post-DST (r⫽⫺.56), but not with baseline
urinary cortisol (r⫽⫺.29, ns), DST suppression (r⫽
.12), or cortisol reactivity to the TSST (r⫽⫺.18). DES
negatively correlated with resting systolic BP (r⫽⫺.54)
and peak HR during the TSST (r⫽⫺.48) but unrelated
to other BP and HR measures during rest and TSST.
Neuroimaging
Structural imaging
Irle et al., 2009 10 PTSD with either DA or
DID, 25 HC
MRI of total brain volume, bilateral
amygdala, and bilateral hippocampus
SCID-D diagnosed DID
and DA
Volumes of left 31% and right 29% amygdala and left 17%
and right 11.0% hippocampal volumes were reduced
when compared with HCs, but correlated with PTSD
symptom severity rather than DA/DID symptoms.
Vermetten et al., 2006 15 DID, 23 HC MRI determined hippocampal and
amygdala volume
SCID-D diagnosed DID Hippocampal volume 19% less in DID but confounded by
age differences. Amygdala volume 32% less in DID, but
only the effect of right amygdala volume still significant
after covarying age.
Weniger et al., 2008 13 DID or DA, 25 HC, 10
PTSD
MRI scan of amygdala and
hippocampal size
SCID-D diagnosed DID
and DA
Neither amygdala nor hippocampal volumes differed
between the DID/DA group and HC.
Functional imaging
Brand et al., 2009 14 DA, 19 HC PET scan acquired during eyes-closed
resting state
DA diagnosis made
according to DSM–IV
DA participants showed less metabolism in the right
inferolateral PFC.
Elzinga et al., 2007 16 DID or DDNOS, 16 HC fMRI scanning during verbal n-back
test of working memory
DES Working memory nonsignificant. DID/DDNOS greater
response as a function of increasing task difficulty,
relative to HC, within left anterior PFC, DLPFC, and
parietal lobe (BA 40).
(table continues)
571
TRAUMA AND FANTASY MODEL OF DISSOCIATION
Table 5 (continued)
Study Sample description Measures and method
Measures of dissociation
and diagnosis Results of interest in brief
Felmingham et al.,
2008
12 PTSD displaying
dissociative reactions,11
PTSD who did not
dissociate
fMRI during fearful and neutral faces
presented consciously or
nonconsciously
CADSS Conscious presentations: Dissociatives had less response in
right superior (BA 8), left middle, right inferior (BA 45),
and medial (BA 6) frontal cortex, but more response in
left ventral ACC (BA 25). Nonconscious presentations:
Dissociatives had greater response in left pallidum,
bilateral amygdala, bilateral insula, and left thalamus.
Hopper et al., 2007 27 PTSD fMRI while hearing trauma scripts CADSS, PDEQ Increasing state dissociation associated with increasing
MPFC response and right superior temporal cortex
response, and decreasing right anterior insula, right
inferior frontal, and left superior temporal cortex
response
Lanius et al., 2002 7 PTSD, 10 HC fMRI while hearing trauma scripts. All
participants with PTSD experienced
state dissociation during scanning
DES and CADSS PTSD exhibited greater response in ACC, right MPFC,
right inferior frontal gyrus, right precuneus, and right
middle and superior temporal gyrus. PTSD exhibited less
response in left superior temporal gyrus, left
parahippocampal gyrus, and right middle frontal gyrus.
Lanius et al., 2005 10 PTSD, 10 HC fMRI while hearing trauma scripts. All
PTSD experienced state dissociation
during scanning
CADSS Greater correlation in HC with left superior frontal cortex,
right parahippocampal gyrus, and right superior occipital
cortex. Greater correlation in PTSD within right middle
frontal cortex, right insula, right cuneus, right superior
temporal cortex, and left superior parietal cortex.
Luda¨scher et al., 2010 15 women with BPD (of
whom 10 comorbid PTSD)
Listened to scripts of either a
personalized dissociation-inducing
situation or a neutral situation during
fMRI. Postscan, dissociation and
pain sensitivity assessed
DSS-4 Dissociation was higher and pain sensitivity was lower after
dissociation-inducing script. High DSS scores
characterized by frontolimbic activation pattern with
increased BOLD signal in the left inferior frontal gyrus
(BA 9) during script in contrast to activation of right
middle frontal gyrus (BA 46) during neutral script.
Medford et al., 2006 10 DD, 12 controls FMRI scanning done while participants
read, then presented with a
recognition test of aversive and
neutral sentences
DD diagnosis according
to DSM–IV
No differences observed in recognition accuracy or reading.
During recognition tests, DD patients exhibited less
response in medial frontal cortex (BA 9,8), orbitofrontal
cortex (BA 11), precuneus (BA 7), and cerebellum.
Phillips et al., 2001 6 DD, 5 HC Viewed aversive and neutral scenes
during fMRI; later judged the
aversiveness of scenes
PSE, DES HC rated aversive pictures as more aversive than neutral
pictures; DD did not. DD exhibited less response in left
insula, bilateral ACC (BA 24/32), occipital cortex (BA
18), lingual gyrus (BA 10), superior temporal gyrus (BA
22/42), and left inferior parietal cortex (BA 40) during
aversive relative to neutral scenes.
Reinders et al., 2006 11 DID Listened to neutral and trauma-related
scripts in an NIS and TIS while HR,
BP, HRV, and rCBF assessed while
undergoing PET
SCID-D diagnosed DID No differences between NIS and TIS during neutral
memory. During traumatic memory, TIS showed greater
activation in sub-cortical areas including bilateral
amygdala, caudate, and left insula. NIS showed greater
activation in cortical areas, including bilateral parietal
cortex, precuneus, MPFC, and ACC
Reinders et al., 2003 11 DID Listened to neutral and trauma scripts
in an NIS and TIS while HR, BP,
HRV, and rCBF assessed while
undergoing PET
SCID-D diagnosed DID No differences between NIS and TIS during neutral
memory processing. During traumatic memory, TIS had
greater activation in left insula and parietal operculum.
NIS showed greater activation in bilateral parietal cortex,
middle frontal cortex, and right MPFC.
572 DALENBERG ET AL.
Table 5 (continued)
Study Sample description Measures and method
Measures of dissociation
and diagnosis Results of interest in brief
Simeon et al., 2000 8 DD, 24 HC Read and recalled word lists SCID-D diagnosed DD DD exhibited less response in right superior (BA 22) and
middle temporal gyrus (BA 21), and more response in
parietal cortex (BA 7B and 39) bilaterally. DES
correlated with response in BA 7B. Greater response in
occipital cortex (left BA 19) in DD patients. DD had
increased response in sensory association cortex.
Veltman et al., 2005 11 high dissociatives, 10
nondissociatives
Verbal working memory tasks (n-back
and Sternberg tasks) while
undergoing fMRI
DIS-Q High dissociatives had greater response in left DLPFC
during both tasks with increasing task difficulty.
Note. Results of a literature review of psychobiological studies of dissociation. Journal articles published in the English language since 1995 were identified by keyword and abstract searches of the
PsycINFO and PubMed databases with the following terms: dissociation/dissociative,biological/psychobiological/psychophysiological/neuroimaging/heart rate/cortisol/skin conductance. Reference
sections of identified articles were also reviewed. To be summarized in the table above, studies had to examine state or trait dissociation as an independent variable in a between-group or within-group
(correlational) design; studies of peritraumatic dissociation were not included. Effect sizes are reported when they were retrievable directly from the references cited. MZ ⫽monozygotic; DZ ⫽
dizygotic; CBCL ⫽Child Behavior Checklist; h
2
⫽heritability; c
2
⫽shared environmental influence; DES-T ⫽Dissociative Experiences Scale taxon score; DES ⫽Dissociative Experiences Scale;
OCD ⫽obsessive compulsive disorder; 5-HTTLPR ⫽serotonin transporter polymorphism; CTQ ⫽Childhood Trauma Questionnaire; COMT ⫽functional catechol-O-methyltransferase Val158Met
polymorphism; MPQ ⫽Multidimensional Personality Questionnaire; CSA ⫽child sexual abuse; HR ⫽heart rate; ADES-T ⫽Adolescent Dissociative Experiences Scale taxon score; SCR ⫽skin
conductance response; PTSD ⫽posttraumatic stress disorder; HRV ⫽heart rate variability; DSS ⫽Dissociation Tension Scale–Acute; PDEQ ⫽Peritraumatic Dissociative Experiences Questionnaire;
SCID-D ⫽Structured Clinical Interview for DSM–IV Dissociative Disorders; DD ⫽depersonalization disorder; HC ⫽healthy controls; CDS ⫽Cambridge Depersonalization Scale; PSE ⫽Present
State Examination; NE ⫽neuroepinephrine; EPI ⫽epinephrine; NPY ⫽neuropeptide Y; CADSS ⫽Clinician-Administered Dissociative States Scale; SASRQ ⫽Stanford Acute Stress Reaction
Questionnaire; HADSI ⫽Hopkins Augmented Dissociative Symptom Inventory; DST ⫽dexamethasone suppression test; TSST ⫽Trier Social Stress Test; BP ⫽blood pressure; MRI ⫽magnetic
resonance imaging; DA ⫽dissociative amnesia; DID ⫽dissociative identity disorder; PET ⫽positron emission tomography; DSM–IV ⫽Diagnostic and Statistical Manual of Mental Disorders (fourth
edition); PFC ⫽prefrontal cortex; DDNOS ⫽dissociative disorder not otherwise specified; fMRI ⫽functional magnetic resonance imaging; DLPFC ⫽dorsolateral prefrontal cortex; ACC ⫽anterior
cingulate cortex; MPFC ⫽medial prefrontal cortex; BPD ⫽borderline personality disorder; DSS-4 ⫽Dissociation Tension Scale; BOLD ⫽blood oxygen level dependent; NIS ⫽neutral identity state;
TIS ⫽traumatic identity state; rCBF ⫽regional cerebral blood flow; DIS-Q ⫽Dissociation Questionnaire.
573
TRAUMA AND FANTASY MODEL OF DISSOCIATION
trauma exposure and response to both idiographic and standard-
ized stimuli (e.g., McTeague et al., 2010).
A growing number of studies have examined cortisol response
as a measure of stress reactivity and functioning of the
hypothalamic–pituitary–adrenal axis in individuals as a function of
dissociative symptoms. However, most studies to date have exam-
ined peritraumatic dissociation only (e.g., Ladwig et al., 2002;
Neylan et al., 2005; Nixon et al., 2005). Higher (Simeon et al.,
2007), lower (Schechter et al., 2004), and null effects have been
observed for basal cortisol in comparisons of individuals high
versus low in dissociative symptoms. Cortisol reactivity to psy-
chological stressors was decreased in response to combat training
as a function of state dissociation (Morgan et al., 2001), but
Simeon et al. (2007, 2008) did not find decreased (or increased)
cortisol reactivity in response to the Trier Social Stress Test.
Finally, Koopman et al. (2003) observed increased salivary corti-
sol in individuals reporting greater trait dissociative symptoms
only 1 day (but not immediately or 2 days) after being interviewed
about traumatic life events. Discrepant findings across studies may
suggest that patterns of arousal differentiating high and low dis-
sociators within PTSD groups may change over time. In addition,
future studies of cortisol reactivity to psychological stressors as a
function of trait dissociation should examine the extent that indi-
viduals experience state dissociation in response to the stressor.
The significance of documented psychophysiological and neu-
roendocrine correlates of self-reported dissociative symptoms can
be interpreted from either the FM or the TM perspective. FM
theorists can maintain that objective psychophysiological re-
sponses to stimuli reminiscent of trauma may represent solely
individuals’ belief that they have experienced trauma, a belief that
may be unfounded in reality. For example, McNally et al. (2004)
found that heart rate, SCR, and left frontal electromyography
increased more significantly in individuals believing themselves to
be alien abductees than in comparison volunteers when exposed to
auditory recounting of alien abduction experiences. However, Mc-
Nally et al. did not distinguish between high and low dissociation
groups in their analysis.
In comparison, TM theorists may note that behavioral and
psychophysiological responses observed in reportedly traumatized
dissociative subjects closely match those often observed in animals
within the context of inescapable predatory threat, a behavioral
pattern referred to in the animal literature as tonic immobility
(Bracha & Maser, 2008; Bracha, Ralston, Matsukawa, Williams, &
Bracha, 2004; Marx, Forsyth, Gallup, Fuse´, & Lexington, 2008;
Moskowitz, 2004). Within the state of tonic immobility, an animal
takes upon itself an outwardly passive defensive response involv-
ing inhibition of movement, muscular rigidity or limpness, and
evidently unfixed concentration (e.g., unfocused gaze, eye clo-
sure), a behavioral and psychophysiological state that has been
associated with increased analgesia.
These characteristics bear a resemblance to certain dissociative
states as discussed above (Frewen & Lanius, 2006; Nijenhuis,
Vanderlinden, & Spinhoven, 1998). Tonic immobility to date has
been examined primarily in its relevance to trauma and PTSD as
opposed to dissociative symptoms specifically, although research-
ers have discussed its particular relevance to dissociative symp-
toms in PTSD (Bovin, Jager-Hyman, Gold, Marx, & Sloan, 2008;
Fiszman et al., 2008; Heidt, Marx, & Forsyth, 2005; Humphreys,
Sauder, Martin, & Marx, 2010; Rocha-Rego et al., 2009). Further-
more, psychometrically measured tonic immobility correlates with
dissociative symptoms (Abrams, Carleton, Taylor, & Asmundson,
2009). In short, the animal literature on tonic immobility affords a
translational model informing the psychophysiological study of
dissociative symptoms. These studies support the basic principle of
the TM that traumatic stress plays a causal role in dissociative
symptoms.
Neuroimaging
Neuroimaging studies have examined emotional processing in
subjects with DPD, and trauma memory and/or pain processing in
individuals with PTSD or borderline personality disorder (BPD)
with prominent dissociative symptoms, with a common finding
being either increased or decreased response in medial prefrontal
cortex and limbic regions accompanying dissociative symptoms
(see Table 5). Phillips et al. (2001) observed less difference in
emotional processing regions of the brain, most notably the insula,
and a greater frontal response, in people with DPD when viewing
valenced photographs.
Among PTSD patients, individuals exhibiting state depersonal-
ization in response to trauma reminders also showed an increased
response within midline anterior regions including the dorsal and
rostral anterior cingulate cortex and the medial prefrontal cortex
(Hopper, Frewen, Sack, Lanius, & Van der Kolk, 2007; Lanius et
al., 2010, 2005, 2002). In comparison, null effects were observed
for the contrast of encoding emotional relative to neutral sentences
in 10 participants with DPD, although during a subsequent recog-
nition test for emotional words, healthy controls activated the
medial prefrontal cortex more so than individuals with DPD (Med-
ford et al., 2006). Less response within medial prefrontal cortex
was also observed in PTSD patients reporting dissociative symp-
toms in response to threatening facial expressions (Felmingham et
al., 2008). Increased midcingulate and insula response in patients
with BPD and comorbid PTSD was observed in conjunction with
reduced pain sensitivity during script-induced dissociative states
(Luda¨scher et al., 2010).
Thus, functional neuroimaging studies increasingly implicate a
frontocingulate and limbic basis for positive symptoms of disso-
ciative disorders and dissociative symptomatology, most notably
those of depersonalization and analgesia. Recently, neuroimaging
studies have also sought to investigate the basis of negative symp-
toms of dissociation, including dissociative amnesia and interiden-
tity amnesia. Findings in 14 individuals with dissociative amnesia
tested with fluorodeoxyglucose PET in a resting state showed
decreased metabolism within the right inferolateral prefrontal cor-
tex (Brand et al., 2009). These findings complement a neuropsy-
chological case series showing reduced response in the frontotem-
poral cortex typically within the right hemisphere, in individuals
whose amnesia was documented to have been provoked by trau-
matic and/or stressful events (review by Staniloiu & Markowitsch,
2010; see also Staniloiu, Markowitsch, & Brand, 2010). Ver-
metten, Schmahl, Lindner, Loewenstein, and Bremner (2006) ob-
served reduced volume of the hippocampus and amygdala in
individuals with DID. This result was not replicated in a subse-
quent study, where brain morphological changes were reported to
be associated with a PTSD diagnosis, not with a dissociative
disorder diagnosis without PTSD (Irle, Lange, Sachsse, & Weni-
ger, 2009; Weniger, Lange, Sachsse, & Irle, 2008). However, in
574 DALENBERG ET AL.
these latter studies, only four of 13 trauma-exposed individuals
met SCID-D diagnostic criteria for DID. Most met diagnostic
criteria for dissociative amnesia, and no data are reported on which
dissociative patients met diagnostic criteria for PTSD. Accord-
ingly, further studies will be needed to more completely elucidate
whether the Vermetten et al. findings can be better explained by
comorbid PTSD, by DID, or by both disorders.
The reviewed neuroimaging studies were not designed to ad-
dress the present question regarding the degree to which dissocia-
tive symptoms represent the product of traumatic life events rather
than fantasy proneness. Complicating this matter, quantitative re-
views of the neuroimaging literature demonstrate that the neural
architecture facilitating the human capacity for episodic memory
(including the medial and ventrolateral prefrontal cortex, medial
and lateral temporal cortex, retrosplenial and posterior cingulate
cortices, temporoparietal junction, and cerebellum; Svoboda,
McKinnon, & Levine, 2006) overlaps significantly with that me-
diating our ability to imagine ourselves taking part in fantasized
events as assessed in studies of prospection and mental time travel
(particularly within medial regions; see Spreng, Mar, & Kim,
2009; Spreng, McKinnon, Mar, & Levine, 2009).
Although the preceding studies address the neural basis of our
faculties for episodic memory and imagination as most germane to
the TM and FM of dissociation, additional research has examined
the neural correlates of working memory function in dissociative
disorders. In comparison with PTSD samples, which generally
show working memory impairment, some studies of dissociative
symptoms (e.g., de Ruiter, Phaf, Elzinga, & Van Dyck, 2004) and
dissociative disorders (Elzinga et al., 2007) have shown preserved
or enhanced working memory ability in comparison with healthy
controls (de Ruiter, Elzinga, & Phaf, 2006; de Ruiter et al., 2004;
cf. Amrhein, Hengmith, Maragkos, & Hennig-Fast, 2008; Terhune
et al., 2011). Increased response within dorsolateral prefrontal
cortex in individuals with high trait dissociation (Veltman et al.,
2005) and DID (Elzinga et al., 2007) during working memory
tasks has also been observed under nonemotional conditions. Neu-
roimaging studies are needed to evaluate whether working mem-
ory ability in individuals with dissociative disorders sustains emo-
tional and/or symptom challenge (e.g., Chiu, Yeh, Huang, Wu, &
Chiu, 2009) and relates to the concept of psychological resilience.
Methodological Concerns
Statistical Partialing as a Test of Causal Theories
The studies in Table 1 do vary in effect size, with Qvalues
(reflecting heterogeneity of effect size) statistically significant for
each trauma type. The methodologies within the nonoutlier studies
(effect sizes .20–.47) include validated questionnaires, interviews,
self-reports, reports by others, and chart reviews. The statistical
significance of the overall weighted rs (as well as the rarity of
nonsignificant studies in general) supports the consistency of the
trauma–dissociation connection across samples; heterogeneity of
effect size suggests the presence of mediators or moderators af-
fecting the strength, not the presence, of the correlation.
The most common response in FM literature is to suggest a set
of potential confounds recommended for covariation to clarify the
causal mechanism, such as Giesbrecht et al.’s (2008) list of “fam-
ily pathology, general psychological distress, and specific variants
of psychopathology associated with dissociation, such as eating
disorders, impulsivity, and schizotypal traits” (p. 632). The impor-
tance of individual or family variables as potentially important
context variables related to the impact of trauma might be cited as
a justification for covariation of such variables, as in Nash, Hulsey,
Sexton, Harralson, and Lambert (1993) and Nash, Neimeyer,
Hulsey, and Lambert (1998). The conclusion then is offered that
some other factor, such as family environment, carries more ex-
planatory power than trauma, or, in the extreme, that the TM
simply does not fit the data.
A study by Mulder, Beautrais, Joyce, and Fergusson (1998) is an
example of the use of partial correlation to make these causal
statements. In this research, a large and representative sample was
assessed on dissociation, clinical pathology, and history of phys-
ical and sexual abuse (not included in Table 1 because the full DES
was not used). At the zero-order level, physical and sexual abuse
were related to high endorsement of dissociative symptoms and to
clinical pathology. When physical abuse, sexual abuse, and clinical
pathology were regressed simultaneously on dissociation, the ef-
fect for sexual abuse became statistically nonsignificant (but the
effect for physical abuse and clinical pathology remained statisti-
cally significant). Mulder et al. concluded that “the influence of
sexual abuse was due to its associations with current psychiatric
illness and with childhood physical abuse” (p. 806), stating that
their findings suggest that “any causal influence of childhood
sexual abuse on dissociation is likely to be indirect and mediated
by more general linkages between childhood sexual abuse and
risks of mental disorder” (p. 809).
The problem with this reasoning is that it assumes foreknowl-
edge of the causal flow within the model. Clinical pathology (such
as anxiety or depressive disorder) may provide a vulnerability to
dissociation; may be a genetic, cultural, or familial covariate of
dissociation proneness; may be an associated feature of dissocia-
tion; or may be a secondary reaction to dissociation. Statistical
control will not provide distinctions between these hypotheses (see
Briere & Elliott, 1993, for a trauma-relevant discussion). As Ped-
hazur (1982) noted, partial correlation is inappropriate if X and Y
are correlated causes of Z (e.g., if anxiety proneness and trauma
history are correlated and both provide vulnerability for dissocia-
tion), or if X affects Z directly, as well as acting through Y (i.e.,
if the effects of trauma on dissociation do not flow entirely through
the existence of other clinical disorders). Both of these alternatives
are entirely plausible models. As summarized by Fisher (1958),
choice of a set of variables without a clear causal model, followed
by the calculation of zero-order and partial correlation coefficients,
“will not advance us a step towards evaluating the importance of
the causes at work” (p. 190).
Therefore, a lesson that the TM theorists should take from the
overall findings is that alternative model-testing with clearly dis-
tinct hypotheses is needed in the field. We have good evidence that
short-term stress and trauma predict dissociation (De Wachter,
Lange, Vanderlinden, Pouw, & Strubbe, 2006; Nixon et al., 2005).
The offering of more complex models, and the pulling apart of
multiple pathways to dissociation (whether trauma initiated or
not), await further research. Promising candidates for inclusion in
these models include genetic markers that appear to facilitate
stress-pathology relationships (Lochner et al., 2007; Savitz et al.,
2008) and known buffers of trauma-related pathology, such as
social support (E. B. Carlson et al., 2001). It should be emphasized
575
TRAUMA AND FANTASY MODEL OF DISSOCIATION
that each of these additions to the TM are proposed as moderators
of the dissociation–trauma relationship, not as mediators of the
relationship. Future researchers would be better served by designs
that include relevant variables as independent grouping factors
(e.g., intrafamilial vs. extrafamilial abuse, low vs. high family
pathology) so that simple effects and interactions can be examined.
Recent studies have begun to answer this question. Data from
the National Comorbidity Study–Replication report that multiple
forms of childhood adversity, including childhood maltreatment
and family dysfunction, covary strongly together, such that it may
not be possible to separate the effects of maltreatment from a
pathogenic family environment in which multiple forms of neglect
and abuse occur (Green et al., 2010; McLaughlin et al. 2010; Scott,
Varghese, & McGrath, 2010). Trickett et al. (2011) came to a
similar conclusion in their review of the many pathological out-
comes of childhood sexual abuse, including increased dissociation.
These adverse outcomes are difficult to completely parcel out from
the manifold harms caused by the pathogenic family environment
in which childhood sexual abuse, physical abuse, emotional abuse,
and neglect occur.
Should We Discount the Nonobjective Trauma
Studies?
Regarding the issue of objective and subjective measures of
trauma in general, it is certainly true that much research on trauma
is conducted with participants whose traumatic background has not
been independently verified. This, however, is the norm rather than
the exception in most areas of psychology. In comparing nonsmok-
ers with light and heavy smokers on rates of varying diseases,
seldom are there external documents verifying the number of
cigarettes per week actually consumed. Salivary cotinine levels
have been used to document abstinence after intervention, but are
used less now because of the high correspondence between these
levels and self-report (Yeager & Krosnick, 2010). The number of
binging or purging episodes for the bulimic are virtually never
verified, nor is there an objective verification that the fantasy-
prone individual actually spends more time fantasizing. Thus, in a
wide range of fields, it is understood that self-report contains
measurement error, and independent studies are conducted to show
that the criterion-positive group (e.g., alcoholic, sexually abused,
bulimic) is reliably more likely to contain criterion-positive indi-
viduals than the self-reported criterion-negative group.
Unfortunately, longitudinal studies cannot provide a full answer
to the question here, since the individual who first reports sexual
abuse as an adult cannot dependably be labeled as a false report
(even if the same individual denied it as a child), because alterna-
tive hypotheses of shame or fear serving to silence the child from
disclosing abuse are viable possibilities. Twenty-year follow-ups
of a large sample of abused children and matched controls revealed
large omission rates for those asked if they had experienced prior
physical abuse (38%–40% in Widom & Shepard, 1996) and prior
sexual abuse (37% of women in Widom & Morris, 1997). Simi-
larly, 31% of children denied recent anal touch (and 14% denied
vaginal touch) in a follow-up of a doctor’s examination that
contained these features (Saywitz, Goodman, Nicholas, & Moan,
1991). In the control group, 3% of children made false allegations
(answered the question positively when the touch did not occur).
These low rates of false allegation are typical of research asking
direct questions regarding sexual touch in studies with objective
evidence (cf. Bottoms, Najdowski, & Goodman, 2009). Therefore,
although we do not see reason to distrust the relationship between
trauma and dissociation that is seen in the many self-report studies,
we agree that the method of collapsing across individuals with
very different types and severity of trauma, and using single-item
or presence–absence measures of trauma or sexual abuse, over-
simplifies experiences and undermines a more complex under-
standing of the impact of such experiences (cf. E. B. Carlson et al.,
2001).
Conclusion and Research Directions
This article has reviewed the evidence for eight predictions
made by the TM and FM of dissociation. The evidence from all
eight areas more strongly and consistently supports the TM than
the FM, as we summarize below.
Strength of the Trauma–Dissociation Relationship
Prediction 1 was strongly supported in favor of the TM. The
trauma–dissociation relationship is not weak and inconsistently
found, as predicted by the FM, but rather appears reliably in both
clinical and community samples. The incorrect conclusions
reached by FM theorists may be caused by their reliance on
undergraduate samples (e.g., Merckelbach et al., 2002) or on very
small and underpowered studies (e.g., Cima et al., 2001). At this
point, research focus should be directed toward risk factors, mod-
erators, and mediators (Kraemer, Stice, Kazdin, Offord, & Kupfer,
2001) of the trauma–dissociation relationship. Promising avenues
include psychiatric and genetic vulnerability (Lochner et al.,
2007), attachment-related variables (E. A. Carlson, 1998; Ogawa
et al., 1997), social support (E. B. Carlson et al., 2001), shame
(Talbot, Talbot, & Tu, 2004), working memory (de Ruiter et al.,
2004), and executive dysfunction (Malek, 2010), among others. It
is quite plausible that dissociation in the context of varying diag-
noses will manifest differently—an area that is underresearched.
Strength of the Trauma–Dissociation Relationship in
Objective Versus Subjective Reports
Prediction 2 was strongly supported in favor of the TM. The
clearest test of the two models is the comparison of studies using
objective and self-report trauma definitions. If the trauma–
dissociation relationship was caused by fantasized trauma within
the dissociative group, clearly the correlation between the two
measurements would be higher in studies relying solely on self-
report or guardian report than in those supplementing report with
CPS findings, therapist ratings, and/or documentation of the
trauma. In contrast with that prediction, objective studies had
effect sizes equivalent to those found in self-report studies.
Course of Dissociation After Trauma and Trauma
Treatment
Prediction 3 was strongly supported in favor of the TM. If
dissociation were nothing more than a manifestation of cognitive
impairment, executive dysfunction, and/or fantasy proneness, as
FM theorists argue, then the pattern of change in dissociation over
576 DALENBERG ET AL.
time should be slow or nonexistent and unrelated to trauma or
trauma treatment. Instead, as the TM suggests, dissociation drops
over the course of the 1st year after trauma for most individuals
(e.g., Feeny, Zoellner, Fitzgibbons, & Foa, 2000; Feeny, Zoellner,
& Foa, 2000), but is stable over the course of short periods in
clinical populations unless treatment is offered (e.g., Chard, 2005;
Rothbaum et al., 2005). The effect size for treatment is typically
moderate. In future research, there is room for methodological
advancement of these designs, including use of active rather than
wait list controls, larger cell sizes, and use of treatments that target
dissociation specifically versus traumatic aftermath generally.
Dissociation Versus Fantasy Proneness as Predictors of
Trauma History
Prediction 4 was strongly supported in favor of the TM. Given
that fantasy proneness and dissociation do covary, the question
arises as to whether there is statistically significant additional
unique variance accounted for by dissociation over fantasy prone-
ness in the prediction of trauma, as hypothesized by TM theorists.
This was repeatedly shown to be the case. However, there is a
paucity of research on a possible fantasy proneness–dissociation
interaction, a focus that would inform and enhance the understand-
ings within both models. Important research questions would in-
clude investigation of base rates for trauma, dissociative disorders,
BPD, and PTSD as they relate to dissociation, fantasy proneness,
and the fantasy proneness–dissociation interaction; investigation
of differential vulnerability to event suggestibility, imagination
inflation, and source confusion; and the study of physiological and
genetic markers for dissociation.
Dissociation and Suggestibility
Prediction 5 was strongly supported in favor of the TM. The
relationship of dissociation to event suggestibility, source moni-
toring, and imagination inflation was weak and inconsistent. In the
event suggestibility, source monitoring, and imagination inflation
categories, 50%–73% of the results were not statistically signifi-
cant, and the remainder typically had small to moderate effects.
This finding does not support the strong statements made that
dissociative individuals, and particularly dissociative disordered
patients, are at high risk for false memories. In fact, studies that
used clinical or abused populations routinely failed to find the
relationship at all. Instead, the findings are more in keeping with a
model suggesting that the difference is one of degree, and that
dissociative individuals as a group are only slightly more likely, if
more likely at all, to accept a suggestion under pressure (with
average effects sizes supporting the conclusion that dissociation
accounts for 1%–3% of the variance in suggestibility).
Methodological improvements in this area would stem from a
merging of the strengths of the TM and FM theoretical back-
grounds. FM theorists raise the likelihood of fantasized trauma,
less commonly discussed in TM literature, but seldom address the
distinctions between false memory, slightly inaccurate memory,
and acquiescence to pressure (Pezdek & Lam, 2007). The latter
three concepts, although each is of great interest, have quite
different implications in clinical and forensic settings. Future re-
search on suggestibility might use paradigms similar to that of
Hyland (2000) for differentiation of false and slightly inaccurate
memory, or that of Ost et al. (2008) for differentiation of false
memory and acquiescence. Further, it is important to extend the
findings of suggestibility research from undergraduate samples to
clinical and community samples. TM theorists might find valuable
information in the inclusion of fantasy proneness in their predic-
tive designs, but as a potential partial explanation for effects and as
an alternative outcome of trauma exposure.
Dissociation and Narrative Fragmentation
Prediction 6 was supported in favor of the TM, although more
research in the area is necessary for definitive conclusions. As
predicted by the TM, dissociation is related to objective fragmen-
tation of narrative in prospective and retrospective studies of
assault victims (Halligan et al., 2003, Studies 1 and 2), and is
higher in dissociative disordered (in this case, DPD) patients than
in controls (Giesbrecht, Merckelbach, et al., 2010). The prediction
of the FM—that fragmentation is related solely to subjective
fragmentation and not to objective fragmentation—is not sup-
ported. A promising direction for future research, following Hal-
ligan et al. (2003), would be to extend the longitudinal work to
include tests of later memory of trauma. The TM would predict
that fragmented memory would be more difficult to maintain over
time, particularly in accurate temporal form, than cohesive mem-
ory. This prediction fits with experimental literature suggesting
that conscious inhibition of memory leads over time to loss of
memory (Wright, Loftus, & Hall, 2001).
Dissociative Amnesia Versus Trauma History as
Fantasy
Prediction 7 was strongly supported in favor of the TM. In
understanding loss of memory for trauma, the FM suggests that
recovered memory or dissociative amnesia patients are fantasiz-
ers—thus referring to the more extreme writings about the pro-
posed “myth” of dissociative amnesia (e.g., Loftus & Ketcham,
1994). It is important to emphasize that we were unable to locate
a single study supporting this point of view. All research attempt-
ing to locate corroboration for the accounts of trauma from those
recovering from dissociative amnesia have found this corrobora-
tion (e.g., Coons, 1994; Geraerts et al., 2007; Williams, 1995), and
studies comparing the accuracy of continuous and recovered mem-
ory have found, in general, equivalent accuracy (e.g., Dalenberg,
1996; Williams, 1995). In studies from the TM perspective, most
research on the correlation between dissociation and the experi-
ence of recovery of memory has found this relationship to be
statistically significant (e.g., McNally et al., 2000; Melchert,
1999). Research on thought suppression (e.g., Geraerts & Mc-
Nally, 2008) typically supports the TM, but directed forgetting
paradigms, using words as targets for forgetting over short time
intervals, have reported more contradictory results (DePrince &
Freyd, 2001, 2004; Devilly et al., 2007). New paradigms have
hypothesized that neural networks that subserve directed forgetting
may also be those that explain psychogenic loss of memory (e.g.,
Anderson & Green, 2001; Anderson et al., 2004). Studies like
these should be extended to dissociative disordered samples,
abused samples, and community samples varying in dissociation
ability.
577
TRAUMA AND FANTASY MODEL OF DISSOCIATION
Psychobiology of Dissociation as a Regulatory
Response to Trauma
Extant research supports the TM of dissociation as a regulatory
response to fear or other extreme emotion with measurable bio-
logical correlates. The strong caveat here is that, to our knowledge,
most research has not been done with FM and TM theories in
mind, and thus has not included measures of fantasy proneness or
suggestibility. Nonetheless, biological researchers have found
trauma-related theories (e.g., tonic immobility) to be useful in
synthesizing findings from animal and human samples. Compel-
ling alternative heuristics that are not trauma related have yet to
appear.
Summary
Finally, in future research, we recommend the careful analysis
of varying alternative causal models; attempts to differentiate
mediators, moderators, and risk factors; the avoidance of use of
outlier studies to make theoretical arguments; and attention to
measurement issues in all conceptual areas (dissociation, fantasy
proneness and false memory) to further this complicated and
fascinating dialogue. Our review of current research suggests that
trauma and dissociation are connected for psychological and neu-
robiological reasons, and fantasy proneness is not the explanation.
References
Abramowitz, E. G., & Lichtenberg, P. (2010). A new hypnotic technique
for treating combat-related posttraumatic stress disorder: A prospective
open study. International Journal of Clinical and Experimental Hypno-
sis, 58, 316–328. doi:10.1080/00207141003760926
Abrams, M. P., Carleton, R. N., Taylor, S., & Asmundson, G. J. G. (2009).
Human tonic immobility: Measurement and correlates. Depression and
Anxiety, 26, 550–556. doi:10.1002/da.20462
Akyu¨z, G., S¸ar, V., Kugu, N., & Dog˘an, O. (2005). Reported childhood
trauma, attempted suicide and self-mutilative behavior among women in
the general population. European Psychiatry, 20, 268–273. doi:10.1016/
j.eurpsy.2005.01.002
Alexander, P. C., & Schaeffer, C. M. (1994). A typology of incestuous
families based on cluster analysis. Journal of Family Psychology, 8,
458–470. doi:10.1037/0893-3200.8.4.458
Allen, J. G., Coyne, L., & Console, D. A. (1996). Dissociation contributes
to anxiety and psychoticism on the Brief Symptom Inventory. Journal of
Nervous and Mental Disease, 184, 639–641. doi:10.1097/00005053-
199610000-00010
Allen, J. G., Coyne, L., & Console, D. A. (1997). Dissociative detachment relates
to psychotic symptoms and personality decompensation. Comprehensive Psy-
chiatry, 38, 327–334. doi:10.1016/S0010-440X(97)90928-7
Allen, J. G., Fultz, J., Huntoon, J., & Brethour, J. R., Jr. (2002). Patho-
logical dissociative taxon membership, absorption, and reported child-
hood trauma in women with trauma-related disorders. Journal of
Trauma & Dissociation, 3, 89–110. doi:10.1300/J229v03n01_07
Allen, J. G., Huntoon, J., & Evans, R. B. (1999). Complexities in complex
posttraumatic stress disorder in inpatient women: Evidence from cluster
analysis of MCMI–III personality disorder scales. Journal of Personality
Assessment, 73, 449–471. doi:10.1207/S15327752JPA7303_10
Amdur, R. L., & Liberzon, I. (1996). Dimensionality of dissociation in
subjects with PTSD. Dissociation, 9, 118–124. doi:10.1016/0006-
3223(96)84140-0
American Psychiatric Association. (2010). H 05 unspecified dissociative
disorder. Retrieved from http://www.dsm5.org/ProposedRevision/
Pages/proposedrevision.aspx?rid⫽59
Amrhein, C., Hengmith, S., Maragkos, M., & Hennig-Fast, K. (2008).
Neuropsychological characteristics of highly dissociative healthy indi-
viduals. Journal of Trauma & Dissociation, 9, 525–542. doi:10.1080/
15299730802226332
Anderson, M. C., & Green, C. (2001). Suppressing unwanted memories by
executive control. Nature, 410, 366–369. doi:10.1038/35066572
Anderson, M. C., Ochsner, K. N., Kuhl, B., Cooper, J., Robertson, E.,
Gabrieli, S. W., . . . Gabrieli, J. D. E. (2004). Neural systems underlying
the suppression of unwanted memories. Science, 303, 232–235. doi:
10.1126/science.1089504
Armstrong, J. G., Putnam, F. W., Carlson, E. B., Libero, D. Z., & Smith,
S. R. (1997). Development and validation of a measure of adolescent
dissociation: The Adolescent Dissociative Experiences Scale. Journal of
Nervous and Mental Disease, 185, 491–497. doi:10.1097/00005053-
199708000-00003
Bagley, C., Rodberg, G., Wellings, D., Moosa-Mitha, M., & Young, L.
(1995). Sexual and physical child abuse and the development of disso-
ciative personality traits: Canadian and British evidence from adolescent
child welfare and child care populations. Child Abuse Review, 4, 99–
113. doi:10.1002/car.2380040207
Barrett, D. (1992). Fantasizers and dissociaters: Data on two distinct
subgroups of deep trance subjects. Psychological Reports, 71, 1011–
1014. doi:10.2466/PR0.71.7.1011-1014
Becker-Blease, K. A., Deater-Deckard, K., Eley, T., Freyd, J. J., Stevenson,
J., & Plomin, R. (2004). A genetic analysis of individual differences in
dissociative behaviors in childhood and adolescence. Journal of Child
Psychology and Psychiatry, 45, 522–532. doi:10.1111/j.1469-
7610.2004.00242.x
Bernstein, E. M., & Putnam, F. W. (1986). Development, reliability, and
validity of a dissociation scale. Journal of Nervous and Mental Disease,
174, 727–735. doi:10.1097/00005053-198612000-00004
Bernstein, I. H., Ellason, J. W., Ross, C. A., & Vanderlinden, J. (2001). On
the dimensionalities of the Dissociative Experiences Scale (DES) and
the Dissociation Questionnaire (DIS-Q). Journal of Trauma & Dissoci-
ation, 2, 101–120. doi:10.1300/J229v02n03_07
Besiroglu, L., Akdeniz, N., Agargun, M. Y., Calka, O., Ozdemir, O., &
Bilgili, S. G. (2009). Childhood traumatic experiences, dissociation and
thought suppression in patients with “psychosomatic” skin diseases.
Stress and Health, 25, 121–125. doi:10.1002/smi.1224
Bohus, M., Haaf, B., Simms, T., Limberger, M. F., Schmahl, C., Unckel,
C., . . . Linehan, M. M. (2004). Effectiveness of inpatient dialectical
behavioral therapy for borderline personality disorder: A controlled trial.
Behaviour Research and Therapy, 42, 487–499. doi:10.1016/S0005-
7967(03)00174-8
Bonanno, G. A., Noll, J. G., Putnam, F. W., O’Neill, M., & Trickett, P. K.
(2003). Predicting the willingness to disclose childhood sexual abuse
from measures of repressive coping and dissociative tendencies. Child
Maltreatment, 8, 302–318. doi:10.1177/1077559503257066
Bottoms, B. L., Najdowski, C. J., & Goodman, G. S. (Eds.). (2009).
Children as victims, witnesses and offenders: Psychological science and
the law. New York, NY: Guilford.
Bovin, M. J., Jager-Hyman, S., Gold, S. D., Marx, B. P., & Sloan, D. M.
(2008). Tonic immobility mediates the influence of peritraumatic fear
and perceived inescapability on posttraumatic stress symptom severity
among sexual assault survivors. Journal of Traumatic Stress, 21, 402–
409. doi:10.1002/jts.20354
Bracha, H. S., & Maser, J. D. (2008). Anxiety and posttraumatic stress
disorder in the context of human brain evolution: A role for theory in
DSM–V?Clinical Psychology: Science and Practice, 15, 91–97. doi:
10.1111/j.1468-2850.2008.00113.x
Bracha, H. S., Ralston, T. C., Matsukawa, J. M., Williams, A. E., &
578 DALENBERG ET AL.
Bracha, A. S. (2004). Does “fight or flight” need updating? Psychoso-
matics, 45, 448–449. doi:10.1176/appi.psy.45.5.448
Brand, M., Eggers, C., Reinhold, N., Fujiwara, E., Kessler, J., Heiss,
W.-D., & Markowitsch, H. J. (2009). Functional brain imaging in 14
patients with dissociative amnesia reveals right inferolateral prefrontal
hypometabolism. Psychiatry Research: Neuroimaging, 174, 32–39. doi:
10.1016/j.pscychresns.2009.03.008
Bremner, J. D., Krystal, J. H., Putnam, F. W., Southwick, S. M., Marmar,
C., Charney, D. S., & Mazure, C. M. (1998). Measurement of dissocia-
tive states with the Clinician-Administered Dissociative States Scale
(CADSS). Journal of Traumatic Stress, 11, 125–136. doi:10.1023/A:
1024465317902
Bremner, J. D., & Marmar, C. R. (Eds.). (1998). Trauma, memory, and
dissociation. Washington, DC: American Psychiatric Press.
Brennen, T., Dybdahl, R., & Kapidzˇic´, A. (2007). Trauma-related and
neutral false memories in war-induced posttraumatic stress disorder.
Consciousness and Cognition, 16, 877–885. doi:10.1016/j.concog
.2006.06.012
Brewin, C. R., Dalgleish, T., & Joseph, S. (1996). A dual-representation
theory of posttraumatic stress disorder. Psychological Review, 103,
670–686. doi:10.1037/0033-295X.103.4.670
Briere, J. (2002). Multiscale Dissociation Inventory. Odessa, FL: Psycho-
logical Assessment Resources.
Briere, J. (2006). Dissociative symptoms and trauma exposure: Specificity, affect
dysregulation and posttraumatic stress. Journal of Nervous and Mental Disease,
194, 78–82. doi:10.1097/01.nmd.0000198139.47371.54
Briere, J., & Elliott, D. M. (1993). Sexual abuse, family environment, and
psychological symptoms: On the validity of statistical control. Journal
of Consulting and Clinical Psychology, 61, 284–288. doi:10.1037/0022-
006X.61.2.284
Briere, J., Scott, C., & Weathers, F. (2005). Peritraumatic and persistent
dissociation in the presumed etiology of PTSD. American Journal of
Psychiatry, 162, 2295–2301. doi:10.1176/appi.ajp.162.12.2295
Brunner, R., Parzer, P., Schuld, V., & Resch, F. (2000). Dissociative
symptomatology and traumatogenic factors in adolescent psychiatric
patients. Journal of Nervous and Mental Disease, 188, 71–77. doi:
10.1097/00005053-200002000-00002
Bryant, R. A. (1995). Autobiographical memory across personalities in
dissociative identity disorder: A case report. Journal of Abnormal Psy-
chology, 104, 625–631. doi:10.1037/0021-843X.104.4.625
Bryant, R. A., Brooks, R., Silove, D., Creamer, M., O’Donnell, M., &
McFarlane, A. C. (2011). Peritraumatic dissociation mediates the rela-
tionship between acute panic and chronic posttraumatic stress disorder.
Behaviour Research and Therapy, 49, 346–351. doi:10.1016/
j.brat.2011.03.003
Carden˜a, E. (2008). Dissociative disorders measures. In A. J. Rush Jr.,
M. B. First, & D. Becker (Eds.), Handbook of psychiatric measures (pp.
587–599). Washington, DC: American Psychiatric Press.
Carden˜a, E., & Carlson, E. (2011). Acute stress disorder revisited. Annual
Review of Clinical Psychology, 7, 245–267. doi:10.1146/annurev-
clinpsy-032210-104502
Carden˜a, E., Koopman, C., Classen, C., Waelde, L. C., & Spiegel, D.
(2000). Psychometric properties of the Stanford Acute Stress Reaction
Questionnaire (SASRQ): A valid and reliable measure of acute stress
reactions. Journal of Traumatic Stress, 13, 719–734. doi:10.1023/A:
1007822603186
Carden˜a, E., & Spiegel, D. (1993). Dissociative reactions to the San
Francisco Bay Area earthquake of 1989. American Journal of Psychia-
try, 150, 474–478.
Carlson, E. A. (1998). A prospective longitudinal study of attachment
disorganization/ disorientation. Child Development, 69, 1107–1128. doi:
10.1111/j.1467-8624.1998.tb06163.x
Carlson, E. B., & Dalenberg, C. (2011). Including dissociation in the
diagnostic criteria for posttraumatic stress disorder. Manuscript sub-
mitted for publication.
Carlson, E. B., Dalenberg, C., Armstrong, J., Daniels, J. W., Loewenstein,
R., & Roth, D. (2001). Multivariate prediction of posttraumatic symp-
toms in psychiatric inpatients. Journal of Traumatic Stress, 14, 549–
567. doi:10.1023/A:1011164707774
Carlson, E. B., Field, N. P., Ruzek, J. I., Bryant, R. A., Dalenberg, C. J.,
Keane, T. M., & Spain, D. A. (2011). Measuring the dynamics of
psychological phenomena and behavior for research and clinical work
with Electronic Proximal Assessment. Manuscript submitted for publi-
cation.
Carlson, E. B., & Putnam, F. W. (1993). An update on the Dissociative
Experiences Scale. Dissociation, 6, 16–27.
Carroll, J. B. (1945). The effect of difficulty and chance success on
correlations between items or between tests. Psychometrika, 10, 1–19.
doi:10.1007/BF02289789
Chard, K. M. (2005). An evaluation of cognitive processing therapy for the
treatment of posttraumatic stress disorder related to childhood sexual
abuse. Journal of Consulting and Clinical Psychology, 73, 965–971.
doi:10.1037/0022-006X.73.5.965
Chiu, C.-D., Yeh, Y.-Y., Huang, Y.-M., Wu, Y.-C., & Chiu, Y.-C. (2009).
The set switching function of nonclinical dissociators under negative
emotion. Journal of Abnormal Psychology, 118, 214–222. doi:10.1037/
a0014654
Chu, A., & DePrince, A. P. (2006). Development of dissociation: Exam-
ining the relationship between parenting, maternal trauma and child
dissociation. Journal of Trauma & Dissociation, 7, 75–89. doi:10.1300/
J229v07n04_05
Cima, M., Merckelbach, H., Klein, B., Shellbach-Matties, R., & Kremer,
K. (2001). Frontal lobe dysfunctions, dissociation, and trauma self-
reports in forensic psychiatric patients. Journal of Nervous and Mental
Disease, 189, 188–190. doi:10.1097/00005053-200103000-00008
Clancy, S. A., Schachter, D. L., McNally, R. J., & Pitman, R. K. (2000).
False recognition in women reporting recovered memories of sexual
abuse. Psychological Science, 11, 26–31. doi:10.1111/1467-9280.00210
Collin-Ve´zina, D., & He´bert, M. (2005). Comparing dissociation and
PTSD in sexually abused school-aged girls. Journal of Nervous and
Mental Disease, 193, 47–52. doi:10.1097/01.nmd.0000149218.76592.26
Coons, P. M. (1994). Confirmation of childhood abuse in child and
adolescent cases of multiple personality disorder and dissociative disor-
der not otherwise specified. Journal of Nervous and Mental Disease,
182, 461–464. doi:10.1097/00005053-199408000-00007
Coons, P. M., & Milstein, V. (1986). Psychosexual disturbances in multiple
personality: Characteristics, etiology, and treatment. Journal of Clinical
Psychiatry, 47, 106–110.
Courtois, C. A. (1999). Recollections of sexual abuse: Treatment principles
and guidelines. New York, NY: Norton.
Dale, K. Y., Berg, R., Elden, A., Odegard, A., & Holte, A. (2009). Testing
the diagnosis of dissociative identity disorder through measures of
dissociation, absorption, hypnotizability and PTSD: A Norwegian pilot
study. Journal of Trauma & Dissociation, 10, 102–112. doi:10.1080/
15299730802488478
Dalenberg, C. J. (1996). Accuracy, timing, and circumstances of disclosure
in therapy of recovered and continuous memories of abuse. Journal of
Psychiatry and Law, 24, 229–275.
Dalenberg, C. J. (2000). Countertransference and the treatment of trauma.
Washington, DC: American Psychological Association. doi:10.1037/
10380-000
Dalenberg, C. (2006). Recovered memory and the Daubert criteria: Re-
covered memory as professionally tested, peer reviewed, and accepted in
the relevant scientific community. Trauma, Violence, & Abuse, 7, 274–
310. doi:10.1177/1524838006294572
Dalenberg, C., Loewenstein, R., Spiegel, D., Brewin, C., Lanius, R.,
Frankel, S., . . . Paulson, K. (2007). Scientific study of the dissociative
579
TRAUMA AND FANTASY MODEL OF DISSOCIATION
disorders. Psychotherapy and Psychosomatics, 76, 400–401. doi:
10.1159/000107570
Dalenberg, C. J., & Paulson, K. (2009). The case for the study of “normal”
dissociation processes. In P. F. Dell & J. A. O’Neil (Eds.), Dissociation
and the dissociative disorders: DSM–V and beyond (pp. 145–154). New
York, NY: Routledge.
Deese, J. (1959). On the prediction of occurrence of particular verbal
intrusions in immediate recall. Journal of Experimental Psychology, 58,
17–22. doi:10.1037/h0046671
Dell, P. F. (2006). The Multidimensional Inventory of Dissociation (MID):
A comprehensive measure of pathological dissociation. Journal of
Trauma & Dissociation, 7, 77–106. doi:10.1300/J229v07n02_06
DePrince, A. P., Allard, C. B., Oh, H., & Freyd, J. J. (2004). What’s in a
name for memory errors? Implications for ethical issues arising from the
use of the term “false memory” for errors in memory for details. Ethics
& Behavior, 14, 201–233. doi:10.1207/s15327019eb1403_1
DePrince, A. P., & Freyd, J. J. (2001). Memory and dissociative tenden-
cies: The roles of attentional context and word meaning in a directed
forgetting task. Journal of Trauma & Dissociation, 2, 67–82. doi:
10.1300/J229v02n02_06
DePrince, A. P., & Freyd, J. J. (2004). Forgetting trauma stimuli. Psycho-
logical Science, 15, 488–492. doi:10.1111/j.0956-7976.2004.00706.x
DePrince, A. P., Weinzierl, K. M., & Combs, M. D. (2008). Stroop
performance, dissociation, and trauma exposure in a community sample
of children. Journal of Trauma & Dissociation, 9, 209–223. doi:
10.1080/15299730802048603
de Ruiter, M. B., Elzinga, B. M., & Phaf, R. H. (2006). Dissociation:
Cognitive capacity or dysfunction? Journal of Trauma & Dissociation,
7, 115–134. doi:10.1300/J229v07n04_07
de Ruiter, M. B., Phaf, R. H., Elzinga, B. M., & Van Dyck, R. (2004).
Dissociative style and individual differences in verbal working memory
span. Consciousness and Cognition, 13, 821–828. doi:10.1016/
j.concog.2004.08.002
Devilly, G. J., Ciociari, J., Piesse, A., Sherwell, S., Zammit, S., Cook, F.,
& Turton, C. (2007). Dissociative tendencies and memory performance
on directed forgetting tasks. Psychological Science, 18, 212–217. doi:
10.1111/j.1467-9280.2007.01875.x
De Wachter, D., Lange, A., Vanderlinden, J., Pouw, J., & Strubbe, E.
(2006). The influence of current stress on dissociative experiences: An
exploratory study in a non-clinical population. Journal of Trauma &
Dissociation, 7, 87–96. doi:10.1300/J229v07n01_07
Diseth, T. H. (2006). Dissociation following traumatic medical treatment
procedures in childhood: A longitudinal follow-up. Development and
Psychopathology, 18, 233–251. doi:10.1017/S0954579406060135
Dorahy, M. J. (2001). Dissociative identity disorder and memory dysfunc-
tion: The current state of experimental research, and its future directions.
Clinical Psychology Review, 21, 771–795. doi:10.1016/S0272-
7358(00)00068-4
Dorahy, M. J. (2006). The Dissociative Processing Scale: A cognitive
organization activated by perceived or actual threat in clinical dissocia-
tors. Journal of Trauma & Dissociation, 7, 29–53. doi:10.1300/
J229v07n04_03
Dorahy, M. J., Irwin, H. J., & Middleton, W. (2002). Cognitive inhibition
in dissociative identity disorder (DID): Developing an understanding of
working memory function in DID. Journal of Trauma & Dissociation, 3,
111–132. doi:10.1300/J229v03n03_07
Dorahy, M. J., Lewis, C. A., & Wolfe, F. A. M. (2007). Psychological
distress associated with domestic violence in Northern Ireland. Current
Psychology, 25, 295–305. doi:10.1007/BF02915237
Drivdahl, S. B., & Zaragoza, M. S. (2001). The role of perceptual elabo-
ration and individual differences in the creation of false memories for
suggested events. Applied Cognitive Psychology, 15, 265–281. doi:
10.1002/acp.701
Duffy, C. (2000). Prevalence of undiagnosed dissociative disorders in an
inpatient setting (Doctoral dissertation). Available from ProQuest Dis-
sertations and Theses database. (AAT No. 3041898)
Dunlap, W. P., Cortina, J. M., Vaslow, J. B., & Burke, M. J. (1996).
Meta-analysis of experiments with matched groups or repeated measures
designs. Psychological Methods, 1, 170–177. doi:10.1037/1082-
989X.1.2.170
Dunn, G. E., Ryan, J. J., & Paolo, A. M. (1994). A principal components
analysis of the Dissociative Experiences Scale in a substance abuse
population. Journal of Clinical Psychology, 50, 936–940. doi:10.1002/
1097-4679(199411)50:6⬍936::AID-JCLP2270500619⬎3.0.CO;2-4
Dutra, L., Bureau, J.-F., Holmes, B., Lyubchik, A., & Lyons-Ruth, K.
(2009). Quality of early care and childhood trauma: A prospective study
of developmental pathways to dissociation. Journal of Nervous and
Mental Disease, 197, 383–390. doi:10.1097/NMD.0b013e3181a653b7
Ehlers, A., & Clark, D. M. (2000). A cognitive model of posttraumatic
stress disorder. Behaviour Research and Therapy, 38, 319–345. doi:
10.1016/S0005-7967(99)00123-0
Eich, E., Macauley, D., Loewenstein, R. J., & Dihle, P. H. (1997). Mem-
ory, amnesia, and dissociative identity disorder. Psychological Science,
8, 417–422. doi:10.1111/j.1467-9280.1997.tb00454.x
Eisen, M. L., & Carlson, E. B. (1998). Individual differences in suggest-
ibility: Examining the influences of dissociation, absorption, and a
history of childhood abuse. Applied Cognitive Psychology, 12, S47–S61.
doi:10.1002/(SICI)1099-0720(199812)12:7⬍S47::AID-ACP598⬎
3.0.CO;2-P
Eisen, M. L., Goodman, G. S., Qin, J., Davis, S., & Crayton, J. (2007).
Maltreated children’s memory: Accuracy, suggestibility and psychopa-
thology. Developmental Psychology, 43, 1275–1294. doi:10.1037/0012-
1649.43.6.1275
Eisen, M. L., Morgan, D. Y., & Mickes, L. (2002). Individual differences
in eyewitness memory and suggestibility: Examining relations between
acquiescence, dissociation and resistance to misleading information.
Personality and Individual Differences, 33, 553–571. doi:10.1016/
S0191-8869(01)00172-6
Eisen, M. L., Qin, J., Goodman, G. S., & Davis, S. L. (2002). Memory and
suggestibility in maltreated children: Age, stress arousal, dissociation,
and psychopathology. Journal of Experimental Child Psychology, 83,
167–212. doi:10.1016/S0022-0965(02)00126-1
El-Hage, W., Darves-Bornoz, J.-M., Allilaire, J.-F., & Gaillard, P. (2002).
Posttraumatic somatoform dissociation in French psychiatric outpa-
tients. Journal of Trauma & Dissociation, 3, 59–74. doi:10.1300/
J229v03n03_04
Elliott, D. M. (1997). Traumatic events: Prevalence and delayed recall in
the general population. Journal of Consulting and Clinical Psychology,
65. doi:10.1037/0022-006X.65.5.811
Elzinga, B. M., Ardon, A. M., Heijnis, M. K., de Ruiter, M. B., Van Dyck, R., &
Veltman, D. J. (2007). Neural correlates of enhanced working-memory perfor-
mance in dissociative disorder: A functional MRI study. Psychological Medi-
cine, 37, 235–245. doi:10.1017/S0033291706008932
Elzinga, B. M., Phaf, R. H., Ardon, A. M., & Van Dyck, R. (2003). Directed
forgetting between, but not within, dissociative personality states. Journal of
Abnormal Psychology, 112, 237–243. doi:10.1037/0021-843X.112.2.237
Endo, T., Sugiyama, T., & Someya, T. (2006). Attention-deficit/
hyperactivity disorder and dissociative disorder among abused children.
Psychiatry and Clinical Neurosciences, 60, 434–438. doi:10.1111/
j.1440-1819.2006.01528.x
Erdelyi, M. H. (1990). Repression, reconstruction, and defense: History
and integration of the psychoanalytic and experimental frameworks. In
J. L. Singer (Ed.), Repression and dissociation: Implications for per-
sonality theory, psychopathology, and health (pp. 1–31). Chicago, IL:
University of Chicago Press.
Evren, C., S¸ar, V., & Dalbudak, E. (2008). Temperament, character, and
dissociation among detoxified male inpatients with alcohol dependency.
Journal of Clinical Psychology, 64, 717–727. doi:10.1002/jclp.20485
580 DALENBERG ET AL.
Feeny, N. C., Zoellner, L. A., Fitzgibbons, L. A., & Foa, E. B. (2000).
Exploring the roles of emotional numbing, depression, and dissociation
in PTSD. Journal of Traumatic Stress, 13, 489–498. doi:10.1023/A:
1007789409330
Feeny, N. C., Zoellner, L. A., & Foa, E. B. (2000). Anger, dissociation, and
posttraumatic stress disorder among female assault victims. Journal of
Traumatic Stress, 13, 89–100. doi:10.1023/A:1007725015225
Felmingham, K., Kemp, A. H., Williams, L., Falconer, E., Olivieri, G.,
Peduto, A., & Bryant, R. (2008). Dissociative responses to conscious
and non-conscious fear impact underlying brain function in post-
traumatic stress disorder. Psychological Medicine, 38, 1771–1780. doi:
10.1017/S0033291708002742
Fisher, R. A. (1958). Statistical methods for research workers. Edinburgh,
Scotland: Oliver and Boyd.
Fiszman, A., Mendlowicz, M. V., Marques-Portella, C., Volchan, E.,
Coutinho, E. S., Souza, W. F., . . . Figueira, I. (2008). Peritraumatic tonic
immobility predicts a poor response to pharmacological treatment in
victims of urban violence with PTSD. Journal of Affective Disorders,
107, 193–197. doi:10.1016/j.jad.2007.07.015
Foa, E. B., Molnar, C., & Cashman, L. (1995). Change in rape narratives
during exposure therapy for posttraumatic stress disorder. Journal of
Traumatic Stress, 8, 675–690. doi:10.1002/jts.2490080409
Foote, B., Smolin, Y., Kaplan, M., Legatt, M. E., & Lipschitz, D. (2006).
Prevalence of dissociative disorder in psychiatric outpatients. American
Journal of Psychiatry, 163, 623–629. doi:10.1176/appi.ajp.163.4.623
Forrest, K. A. (1999). Explicit memory and dissociative identity disorder:
The function of one-way amnesia barriers. Dissertation Abstracts Inter-
national: Section B. Sciences and Engineering, 60(1), 0382.
Forrest, K. A. (2001). Toward an etiology of dissociative identity disorder:
A neurodevelopmental approach. Consciousness and Cognition, 10,
259–293. doi:10.1006/ccog.2001.0493
Francia-Martı´nez, M., Roca de Torres, I., Alvarado, C. S., Martı´nez-
Taboas, A., & Sayers, S. (2003). Dissociation, depression and trauma in
psychiatric inpatients in Puerto Rico. Journal of Trauma & Dissociation,
4, 47–61. doi:10.1300/J229v04n04_04
Frewen, P. A., & Lanius, R. A. (2006). Neurobiology of dissociation: Unity
and disunity in mind–body–brain. Psychiatric Clinics of North America,
29, 113–128. doi:10.1016/j.psc.2005.10.016
Freyd, J. J., & Gleaves, D. H. (1996). “Remembering” words not presented
in lists: Relevance to the current recovered/false memory controversy.
Journal of Experimental Psychology: Learning, Memory, and Cogni-
tion, 22, 811–813. doi:10.1037/0278-7393.22.3.811
Freyd, J. J., Klest, B., & Allard, C. B. (2005). Betrayal trauma: Relation-
ship to physical health, psychological distress, and a written disclosure
intervention. Journal of Trauma & Dissociation, 6, 83–104. doi:
10.1300/J229v06n03_04
Gast, U., Rodewald, F., Nickel, V., & Emrich, H. M. (2001). Prevalence of
dissociative disorders among psychiatric inpatients in a German univer-
sity clinic. Journal of Nervous and Mental Disease, 189, 249–257.
doi:10.1097/00005053-200104000-00007
Geraerts, E., & McNally, R. J. (2008). Forgetting unwanted memories:
Direct forgetting and thought suppression methods. Acta Psychologica,
127, 614–622. doi:10.1016/j.actpsy.2007.11.003
Geraerts, E., McNally, R. J., Jelicic, M., Merckelbach, H., & Raymaekers,
L. (2008). Linking thought suppression and recovered memories of
childhood sexual abuse. Memory, 16, 22–28. doi:10.1080/
09658210701390628
Geraerts, E., Merckelbach, H., Jelicic, M., Smeets, E., & Van Heerden, J.
(2006). Dissociative symptoms and how they relate to fantasy proneness
in women reporting repressed or recovered memories. Personality and
Individual Differences, 40, 1143–1151. doi:10.1016/j.paid.2005.11.003
Geraerts, E., Schooler, J. W., Merckelbach, H., Jelicic, M., Hauer, B. J. A.,
& Ambadar, Z. (2007). The reality of recovered memories: Corroborat-
ing continuous and discontinuous memories of childhood sexual abuse.
Psychological Science, 18, 564–568. doi:10.1111/j.1467-9280.2007
.01940.x
Geraerts, E., Smeets, E., Jelicic, M., Van Heerden, J., & Merckelbach, H.
(2005). Fantasy proneness, but not self-reported trauma is related to
DRM performance of women reporting recovered memories of child-
hood sexual abuse. Consciousness and Cognition, 14, 602–612. doi:
10.1016/j.concog.2005.01.006
Giesbrecht, T., Geraerts, E., & Merckelbach, H. (2007). Dissociation,
memory commission errors, and heightened autonomic reactivity. Psy-
chiatry Research, 150, 277–285. doi:10.1016/j.psychres.2006.04.016
Giesbrecht, T., Lynn, S. J., Lilienfeld, S. O., & Merckelbach, H. (2008).
Cognitive processes in dissociation: An analysis of core theoretical
assumptions. Psychological Bulletin, 134, 617–647. doi:10.1037/0033-
2909.134.5.617
Giesbrecht, T., Lynn, S. J., Lilienfeld, S. O., & Merckelbach, H. (2010).
Cognitive processes, trauma, and dissociation—Misconceptions and
misrepresentations: Reply to Bremner (2010). Psychological Bulletin,
136, 7–11. doi:10.1037/a0018068
Giesbrecht, T., Merckelbach, H., Kater, M., & Sluis, A. F. (2007). Why dissoci-
ation and schizotypy overlap: The joint influence of fantasy proneness, cogni-
tive failures, and childhood trauma. Journal of Nervous and Mental Disease,
195, 812–818. doi:10.1097/NMD.0b013e3181568137
Giesbrecht, T., Merckelbach, H., Van Oorsouw, K., & Simeon, D. (2010).
Skin conductance and memory fragmentation after exposure to an emo-
tional film clip in depersonalization disorder. Psychiatry Research, 177,
342–349. doi:10.1016/j.psychres.2010.03.010
Gleaves, D. H., Eberenz, K. P., Warner, M. S., & Fine, C. G. (1995).
Measuring clinical and non-clinical dissociation: A comparison of the
DES and QED. Dissociation, 8, 24–31.
Gleaves, D. H., Hernandez, E., & Warner, M. S. (1999). Corroborating
premorbid dissociative symptomatology in dissociative identity disor-
der. Professional Psychology: Research and Practice, 30, 341–345.
doi:10.1037/0735-7028.30.4.341
Goldberg, L. R. (1999). The Curious Experiences Survey, a revised version
of the Dissociative Experiences Scale: Factor structure, reliability, and
relations to demographic and personality variables. Psychological As-
sessment, 11, 134–145. doi:10.1037/1040-3590.11.2.134
Green, J. G., McLaughlin, K. A., Berglund, P. A., Gruber, M. J., Sampson, N. A.,
Zaslavsky, A. M., & Kessler, R. C. (2010). Childhood adversities and adult
psychiatric disorders in the National Comorbidity Survey Replication I: Asso-
ciations with first onset of DSM–IV disorders. Archives of General Psychiatry,
67, 113–123. doi:10.1001/archgenpsychiatry.2009.186
Griffin, M. G., Resick, P. A., & Mechanic, M. B. (1997). Objective
assessment of peritraumatic dissociation: Psychophysiological indica-
tors. American Journal of Psychiatry, 154, 1081–1088.
Gudjonsson, G. H. (1997). Gudjonsson Suggestibility Scales. Hove, Eng-
land: Psychology Press.
Gudjonsson, G. H. (Ed.). (2003). The psychology of interrogations and
confessions: A handbook. West Sussex, England: Wiley.
Hagenaars, M. A., Van Minnen, A., & Hoogduin, K. A. L. (2010). The
impact of dissociation and depression on the efficacy of prolonged
exposure treatment for PTSD. Behaviour Research and Therapy, 48,
19–27. doi:10.1016/j.brat.2009.09.001
Halligan, S. L., Michael, T., Clark, D. M., & Ehlers, A. (2003). Posttrau-
matic stress disorder following assault: The role of cognitive processing,
trauma memory, and appraisals. Journal of Consulting and Clinical
Psychology, 71, 419–431. doi:10.1037/0022-006X.71.3.419
Haraldsson, E. (2003). Children who speak of past-life experiences: Is there a
psychological explanation? Psychology and Psychotherapy: Theory, Research
and Practice, 76, 55–67. doi:10.1348/14760830260569256
Harvey, A. G., & Bryant, R. A. (1999). A qualitative investigation of the
organization of traumatic memories. British Journal of Clinical Psychol-
ogy, 38, 401–405. doi:10.1348/014466599162999
Hauschildt, M., Peters, M. J. V., Moritz, S., & Jelinek, L. (2011). Heart rate
581
TRAUMA AND FANTASY MODEL OF DISSOCIATION
variability in response to affective scenes in posttraumatic stress disor-
der. Biological Psychology, 88, 215–222. doi:10.1016/j.biopsycho
.2011.08.004
He´bert, M., & Tourigny, M. (2010). Effects of a psychoeducational group
intervention for children victims of sexual abuse. Journal of Child &
Adolescent Trauma, 3, 143–160. doi:10.1080/19361521003726930
Heidt, J. M., Marx, B. P., & Forsyth, J. P. (2005). Tonic immobility and
childhood sexual abuse: A preliminary report evaluating the sequela of
rape-induced paralysis. Behaviour Research and Therapy, 43, 1157–
1171. doi:10.1016/j.brat.2004.08.005
Hekkanen, S. T., & McEvoy, C. (2002). False memories and source-
monitoring problems: Criterion differences. Applied Cognitive Psychol-
ogy, 16, 73–85. doi:10.1002/acp.753
Hetzel-Riggin, M. D. (2010). Peritraumatic dissociation and PTSD effects
on physiological response patterns in sexual assault victims. Psycholog-
ical Trauma: Theory, Research, Practice, and Policy, 2, 192–200.
doi:10.1037/a0019892
Hetzel-Riggin, M. D., & Wilber, E. L. (2010). To dissociate or suppress? Predict-
ing automatic vs. conscious cognitive avoidance. Journal of Trauma & Disso-
ciation, 11, 444–457. doi:10.1080/15299732.2010.495376
Holmes, E. A., Brewin, C. R., & Hennessy, R. G. (2004). Trauma films,
information processing, and intrusive memory development. Journal of
Experimental Psychology: General, 133, 3–22. doi:10.1037/0096-
3445.133.1.3
Holmes, E. A., Brown, R. J., Mansell, W., Fearon, R. P., Hunter, E. C. M.,
Frasquilho, F., & Oakley, D. A. (2005). Are there two qualitatively
distinct forms of dissociation? A review and some clinical implications.
Clinical Psychology Review, 25, 1–23. doi:10.1016/j.cpr.2004.08.006
Holowka, D. W., King, S., Saheb, D., Pukall, M., & Brunet, A. (2003).
Childhood abuse and dissociative symptoms in adult schizophrenia.
Schizophrenia Research, 60, 87–90. doi:10.1016/S0920-9964(02)
00296-7
Hopper, J. W., Frewen, P. A., Sack, M., Lanius, R. A., & Van der Kolk,
B. A. (2007). The Responses to Script-Driven Imagery Scale (RSDI):
Assessment of state posttraumatic symptoms for psychobiological and
treatment research. Journal of Psychopathology and Behavioral Assess-
ment, 29, 249–268. doi:10.1007/s10862-007-9046-0
Hornstein, N. L., & Putnam, F. W. (1992). Clinical phenomenology of
child and adolescent dissociative disorders. Journal of the American
Academy of Child & Adolescent Psychiatry, 31, 1077–1085. doi:
10.1097/00004583-199211000-00013
Horselenberg, R., Merckelbach, H., & Josephs, S. (2003). Individual dif-
ferences and false confessions: A conceptual replication of Kassin and
Kiechel (1996). Psychology, Crime & Law, 9, 1–8. doi:10.1080/
10683160308141
Horselenberg, R., Merckelbach, H., Muris, P., Rassin, E., Sijsenaar, M., &
Spaan, V. (2000). Imagining fictitious childhood events: The role of
individual differences in imagination inflation. Clinical Psychology &
Psychotherapy, 7, 128–137. doi:10.1002/(SICI)1099-0879(200005)7:
2⬍128::AID-CPP238⬎3.0.CO;2-Q
Horselenberg, R., Merckelbach, H., Smeets, T., Franssens, D., Peters,
G.-J. Y., & Zeles, G. (2006). False confessions in the lab: Do plausibility
and consequences matter. Psychology, Crime & Law, 12, 61–75. doi:
10.1080/1068310042000303076
Horselenberg, R., Merckelbach, H., Van Breukelen, G., & Wessel, I.
(2004). Individual differences in the accuracy of autobiographical mem-
ory. Clinical Psychology & Psychotherapy, 11, 168–176. doi:10.1002/
cpp.400
Humphreys, K. L., Sauder, C. L., Martin, E. K., & Marx, B. P. (2010).
Tonic immobility in childhood sexual abuse survivors and its relation-
ship to posttraumatic stress symptomatology. Journal of Interpersonal
Violence, 25, 358–373. doi