The Relationship Between Pathological Dissociation, Self-Injury and Childhood Trauma in Patients with Personality Disorders Using Taxometric Analyses

This article was downloaded by: [New York University]On: 24 February 2015, At: 06:53Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Journal of Trauma & DissociationPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/wjtd20

The Relationship Between Pathological Dissociation,Self-Injury and Childhood Trauma in Patients withPersonality Disorders Using Taxometric AnalysesMarianne Goodman MD a , Daniel S. Weiss PhD b , Vivian Mitropoulou MA a , Antonia New MD a

, Harold Koenigsberg MD a , Jeremy M. Silverman PhD a & Larry Siever MD aa Department of Psychiatry , Mount Sinai School of Medicine and the Bronx VA MedicalCenter , New York, NYb Langley Porter Psychiatric Institute , San Francisco, CAPublished online: 17 Oct 2008.

To cite this article: Marianne Goodman MD , Daniel S. Weiss PhD , Vivian Mitropoulou MA , Antonia New MD , HaroldKoenigsberg MD , Jeremy M. Silverman PhD & Larry Siever MD (2003) The Relationship Between Pathological Dissociation,Self-Injury and Childhood Trauma in Patients with Personality Disorders Using Taxometric Analyses, Journal of Trauma &Dissociation, 4:2, 65-88, DOI: 10.1300/J229v04n02_05

To link to this article: http://dx.doi.org/10.1300/J229v04n02_05

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of theContent. Any opinions and views expressed in this publication are the opinions and views of the authors, andare not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon andshould be independently verified with primary sources of information. Taylor and Francis shall not be liable forany losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use ofthe Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

The RelationshipBetween Pathological Dissociation,Self-Injury and Childhood Trauma

in Patients with Personality DisordersUsing Taxometric Analyses

Marianne Goodman, MDDaniel S. Weiss, PhD

Vivian Mitropoulou, MAAntonia New, MD

Harold Koenigsberg, MDJeremy M. Silverman, PhD

Larry Siever, MD

ABSTRACT. Objective: Dissociative phenomena exist on a spectrumranging from psychological absorption to highly symptomatic disrup-tions of identity and memory. A statistical methodology called taxometric

Marianne Goodman, Vivian Mitropoulou, Antonia New, Harold Koenigsberg,Jeremy M. Silverman, and Larry Siever are affiliated with the Department of Psychia-try, Mount Sinai School of Medicine and the Bronx VA Medical Center, New York,NY.

Daniel S. Weiss is affiliated with the Langley Porter Psychiatric Institute, San Fran-cisco, CA.

Address correspondence to: Marianne Goodman, MD, Bronx VA Medical Center,130 West Kingsbridge Road, Bronx, NY 10468 (E-mail: [email protected]).

The authors wish to acknowledge Drs. Rachel Yehuda and Daphne Simeon for theirhelpful comments on earlier drafts of the paper and Dr. Tenko Raykov who assistedwith statistical consultation.

Parts of this paper were presented at Biological Psychiatry, Chicago, IL, May 1999.

Journal of Trauma & Dissociation, Vol. 4(2) 2003http://www.haworthpress.com/store/product.asp?sku=J229

2003 by The Haworth Press, Inc. All rights reserved.10.1300/J229v04n02_05 65

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

analysis has established a set of indicators that identify patients who havepathological dissociation, a qualitatively different form of dissociativephenomena. Using taxometric methodology, this study examines the re-lationship of pathological dissociation to personality diagnosis and self-directed injury, including suicide and history of childhood abuse, in asample of outpatients with personality disorders.

Method: Patients were recruited from advertisements or referred fromlocal clinicians. Participants completed a diagnostic interview and ratingscales for dissociation, self-injury and childhood trauma. Pathologicaldissociation was identified using the Dissociative Experiences Scale-Taxon (DES-T; Waller, Putnam, & Carlson, 1996). Membership in thepathological dissociation taxon was established by calculating Bayesianposterior taxon membership probabilities; the method advocated byWaller, and compared to an approximation, used widely in the literature,based simply on the unweighted mean of the DES-T items.

Results: Overlapping, but not identical groups of patients were identi-fied, indicating that the two methods are not interchangeable in this sam-ple of personality disordered individuals. Surprisingly, no associationswere detected between indices of childhood trauma and membership inthe pathological dissociation taxon nor for the high dissociators identi-fied through the approximation method.

Conclusions: This study serves as a replication of the ability to detectpathological dissociation as measured by the DES-T. Nonetheless, thefailure to confirm our hypotheses regarding an association betweenpathological dissociation, childhood trauma, and personality diagnosisraise a challenge to some parts of existing etiologic theories. [Article cop-ies available for a fee from The Haworth Document Delivery Service:1-800-HAWORTH. E-mail address: <[email protected]> Website:<http://www.HaworthPress.com> 2003 by The Haworth Press, Inc. All rightsreserved.]

KEYWORDS. Pathological dissociation, childhood trauma, personalitydisorder, taxometric analysis

INTRODUCTION

Dissociation, as defined by the fourth edition of the Diagnostic andStatistical Manual of Mental Disorders (DSM-IV; American PsychiatricAssociation, 1994) is the “disruption of the usually integrated functionsof consciousness, memory, identity, or perception of the environment”(p. 477). The Dissociative Experiences Scale (DES) (Bernstein & Putnam,

66 JOURNAL OF TRAUMA & DISSOCIATION

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

1986) is the most widely used instrument for the identification of disso-ciation in clinical and research populations, and the vast majority of theresearch literature on the clinical import of dissociative phenomena hasused the DES. Thus, careful examination of the clinical implications ofstudies using the DES can help clarify important issues in the role ofdissociation in psychopathology, including the personality disorders.

The traumagenic model of dissociation (Putnam, 1995) is widely ac-cepted in the literature and posits that trauma influences the expressionof dissociative phenomena. Dissociation is considered a psychologicaldefense that shields the individual from the full impact of a trauma. The“states of consciousness model” of dissociation proposes that more se-vere dissociative phenomena such as memory disturbance and fragmen-tation of the self derive from trauma induced disruptions of the ability tomodulate states of consciousness and integrate the self across these dis-parate states (Putnam, 1991). In support of these models, a history ofchildhood sexual, physical or emotional abuse has been associated withincreased dissociative tendencies in both clinical (Chu & Dill, 1990;Herman, Perry, & van der Kolk, 1989; Ogata et al., 1990; Ross, Joshi, &Currie, 1991) and non-clinical samples (Briere & Runtz, 1988; Irwin,1994). Van IJzendoorn and Schuengel’s (1996) meta-analysis of theDES examining the relationship between dissociation and PTSD washighly significant with a robust effect size (d = 0.75; N = 1099). More-over, in the same review, 26 studies were described that examined therelationship between dissociation and physical or sexual abuse. Ahighly significant but medium combined effect size of 0.52 was re-ported (N = 2108, d = 0.42 for physical abuse, d = 0.42 for sexual abuseor both abuses d = 0.58). Though some have suggested that these studiessupport the notion that posttraumatic stress disorder (PTSD) be re-conceptualized as a dissociative disorder (Tampke & Irwin, 1999), thisview overlooks those cases of PTSD that develop with no dissociativephenomena. Moreover, even though recent research has shown thatperitraumatic dissociation is a significant predictor of the developmentof PTSD (Ozer, Best, Lipsey, & Weiss, 2001), the strength of these rela-tionships leaves considerable room for individual variation. Thus, al-though dissociation may be strongly related to these forms of abuse, theinevitability of dissociative disorder or PTSD following such exposurehas not been demonstrated. As well, it is still an open question as towhether certain psychopathology, when present, is typically precededby a history of abuse, even though many may take this latter notion asestablished.

Goodman et al. 67

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

Recently a distinction between pathological and non-pathologicaldissociation has been made based on statistical methods designed to de-tect qualitative rather than quantitative differences by identifying a cat-egorical classification known as a taxon (Waller, Putnam, & Carlson,1996). Pathological dissociation is a typological construct and wasfound to comprise experiences such as depersonalization, and dis-sociative amnesia; phenomena definitive of the DSM-IV dissociativedisorders.

The statistical methodology of taxometrics (Meehl, 1995; Waller &Meehl, 1998) is based on the conjecture that there are distinct classes orcategories of patients that share important phenotypic and/or genotypiccharacteristics that are not shared in people who are not part of the class.Meehl (1995) uses the example of headache and stiff neck, both preva-lent phenomena, to clarify that it is the unusual co-occurrence of thesephenomena that identifies a taxon-in this case meningitis. However, theconcept may also apply to more purely behavioral features. Taxometricanalyses were undertaken with the DES (Waller et al., 1996), and yieldedboth evidence of and a method for detecting individuals who exhibitpathological dissociation. This method yields two distinct classes orgroups; one group comprises members of the taxon who demonstratepathological dissociative symptoms while the remainder are non-taxonmembers. This conceptualization explicitly moves the concept of path-ological dissociation away from a dimension or trait model to a discreteor typological construct and distinguishes pathologic forms of dissocia-tion from more “normal” dissociative states such as daydreaming.

The significance of the approach to the identification of a taxon inpsychopathology is that it moves the definition of the diagnostic entityaway from an arbitrary set of phenotypic characteristics (the current ap-proach in DSM-IV), to whatever characteristics distinguish taxon fromnon-taxon group members, regardless of whether these phenotypic orgenotypic characteristics are part of the symptom picture. In this way,phenomena that are currently seen as definitive of a disorder may endup only being evidentiary of the disorder. This is a radically differentunderstanding of signs and symptoms, and the clinical implications arestrongest in the identification of members of the taxon class, in whomthere may be considerably less phenotypic similarity than is now thecase for the DSM-IV disorders.

The pathological dissociation taxon previously identified by Walleret al. (1996) is based on eight items from the DES (see Appendix). Ex-amples of items from the DES that mark membership in the dissociativetaxon include finding one self in a place and not knowing how you got

68 JOURNAL OF TRAUMA & DISSOCIATION

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

there, experiencing your surroundings as unreal, feeling like two differ-ent people, not recognizing family or friends, feeling as if your bodydoes not belong to you, and hearing voices in your head. Waller et al.(1996) termed the set of items indicating membership in the taxon theDES-T.

Non-pathological dissociative phenomena, by contrast, are experi-ences such as daydreaming, or becoming absorbed in thought or activ-ity. Scores on the DES items not in the DES-T have a normal distributionin the general population and share statistical characteristics of a per-sonality dimension (Tellegen & Atkinson, 1974). Sample DES items ofnon-pathological dissociation include-becoming absorbed in a movieand becoming unaware of other events, driving in a car and not remem-bering the trip, finding a familiar place strange, the ability to ignorepain, and talking out loud to oneself when you are alone.

Members of the pathological dissociation taxon are identified usingthe items of the DES-T and calculating Bayesian posterior taxon mem-bership probabilities based on weights derived from the taxometricanalysis (Waller et al., 1996). There has been sufficient confusion re-garding this step that the International Society for the Study of Dissoci-ation website has added a link to clarify the procedure. Nonetheless,research has been published using an approximation to this methodol-ogy based on the simple unweighted mean of the eight DES-T items(Tampke & Irwin, 1999; Irwin 1999; Simeon et al., 1998; Jang Paris,Zweig-Frank, & Livesley, 1998). We are not aware of any reports todate that compare the findings of pathological dissociation taxon mem-bership as calculated by Bayesian posterior probabilities with an ap-proximation method, which utilizes the unweighted mean of individualDES-T items.

A meta-analysis of the DES reported seven studies on personality dis-order participants yielding a median DES score of 16.8 (van IJzendoorn &Schuengel, 1996) compared to the mean score of normal participants of11.6 ± 10.6. Studies focusing on borderline personality disorder (BPD)in particular, yield higher average DES scores ranging from 19.5 ± 16.36(Brodsky, Cloitre, & Dulit, 1995) to 25 ± 21.19 (Shearer, 1994) for indi-viduals carrying that diagnosis. Patients with BPD have a wide range ofdissociative experiences, as evidenced by a recent study of 290 BPD pa-tients, in which 68% of patients reported moderate to high levels ofdissociative symptoms based on the Dissociative Experiences Scale(DES) (Zanarini, Ruser, Frankenburg, & Hennen, 2000). The impor-tance of dissociation to borderline phenomena was reflected by the ad-

Goodman et al. 69

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

dition in DSM-IV of a ninth diagnostic criterion that includes “severedissociative symptoms” (American Psychiatric Association, 1994, p. 654)to the diagnosis of BPD. Dissociation in BPD has been associated withimpulsive aggressive acts, and in particular self-injury (Kemperman etal., 1997). Additionally, traits from another cluster A personality disor-der, schizotypal personality disorder (SPD) have been reported to be as-sociated with DES scores (Spiegel & Cardena, 1991). There are alsoreports of a relationship between schizotypy and dissociative tenden-cies (Bauer & Power, 1995; Irwin, 1998). However, the literature todate on the relationship of dissociation to personality disorder diagnosishas not focused on pathological dissociation nor applied taxometricmethodology.

The literature and clinical practice support the belief that childhoodtrauma is an important etiology of pathological dissociation (Startup,1999). Data from an adolescent twin study estimated a zero heritabilityof dissociative pathology, but found that shared environmental influ-ences (e.g., chaotic home environment) accounted for 45% of the varianceof DES-T scores (Waller & Ross, 1997). In another study, dimensionsof childhood trauma (Irwin, 1999) positively predicted pathologicaldissociation but not non-pathological dissociation. In a study of patientswith depersonalization, childhood interpersonal trauma and in particu-lar emotional and sexual abuse was predictive of high levels of dissocia-tion (Simeon, Guralnik, Schmeidler, Sirof, & Knutelska, 2001). Therole of trauma in the pathological dissociation of personality disorders,especially BPD, is an understudied phenomenon, despite considerablestudy on the relationship between trauma and BPD (e.g., Trull, 2001a;Trull, 2001b; Zelkowitz, Paris, Guzder, & Feldman, 2001; Heffernan &Cloitre, 2000; Zanarini, 2001; Paris, 1998; Salzman, 1998; Zanarini etal., 1997; Zweig-Frank, Paris, & Guzder, 1994a; Links & van Reekum,1993; Herman, Perry, & van der Kolk, 1989).

Few studies have examined the relationship between dissociationand history of childhood traumatic experience, personality disorder di-agnosis, and expression of suicide and self-mutilating behaviors. Shearer(1994) noted that in female patients with BPD, a higher level of dis-sociative symptoms was associated with an increased frequency ofself-mutilating behaviors. Furthermore, childhood sexual and physicalabuse, as well as adult sexual assault, were significant predictors ofDES scores. These patients were female inpatients with BPD and highlevels of lifetime co-morbid Axis I pathology and Dissociative disorderdiagnosis. In an outpatient cohort of personality disordered individuals,the DES score correlated with self-mutilation but not with childhood

70 JOURNAL OF TRAUMA & DISSOCIATION

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

trauma variables (Zweig-Frank, Paris, & Guzder, 1994c). Because inthese studies the measurement of dissociation was based on a purely di-mensional approach using total DES scores, it is unclear to what degreepathological dissociation, as identified by the taxometric approach, isimplicated.

The present study examines the relationship between pathologicaldissociation defined by appropriate taxometrics methodology using theDES-T, a history of self-injurious behavior, and history of childhoodtraumatic experience in participants with personality disorders. We hy-pothesized the following: (1) patients with BPD would show signifi-cantly higher levels of pathological dissociation compared to non-BPDpatients; (2) members of the pathological dissociation taxon woulddemonstrate an increased incidence of self-directed aggression (i.e.,history of suicide attempts or self-mutilation) compared to those with-out pathological dissociation; and (3) patients with pathological disso-ciation would report an increased frequency of childhood physical andsexual trauma.

Finally, we were interested in whether the use of a simple unweight-ed mean of DES-T items and ad hoc cut-off (designated hereafter as thehigh DES-T group) would closely replicate the membership of thepathological dissociation taxon when calculated using posterior Bayesiantaxon membership probabilities. The former method does not use datafrom any particular study to modify the decision about a particular par-ticipant. If the score is above the cut-off the participant is a member; ifbelow, not a member. Additional data would have no impact on decid-ing who had pathological dissociation, unless the data were able tomodify the cut-off score. The Bayesian approach works in probabilities.Before any data are collected, the investigator has a subjective probabil-ity about who is a member of the taxon and who is not. After data arecollected, new probabilities (called posterior because they are calcu-lated after data are collected) can be calculated using Bayes theoremthat updates the probability of taxon membership in light of data aboutthe DES items’ taxonicity.

To compare these two approaches, we conducted analyses to illumi-nate this issue. It is important to investigate if the current standard in theliterature, of assuming that an “approximation” (i.e., use of the un-weighted mean of individual DES-T items with an unchanging arbitrarycut-off) yields similar findings to the more statistically rigorous meth-odology of calculating Bayesian posterior probabilities is correct. Webelieve that this is an invalid assumption.

Goodman et al. 71

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

METHODS

Participants

Participants consisted of 95 personality disorder patients (62 males,33 females) age 18-60 (average age 38.4 ± 9.9), meeting criteria for anaverage of 2.3 ± 1.5 personality disorder diagnoses of which 37 patientsmet criteria for BPD. The racial and ethnic distribution of the samplewas as follows: 66% Caucasian, 18% African American, 13% Hispanicand 3% other. The majority of our research participants were recruitedby advertisement for a comprehensive study of personality disordersfrom which we report data only in part, and greater than 80% are mem-bers of the community (mean education level 15 ± 2.6) functioningbetter than typical help-seeking participants. Forty-eight percent (48%)of the sample met criteria for history of major depressive disorder(MDD), 24% of the sample met criteria for current MDD, 20% met cri-teria for history of PTSD, and 8% met criteria for PTSD at the time ofthe evaluation. Participants were excluded if they met criteria for sub-stance abuse in the last six months. Twenty percent of the sample, how-ever, met criteria for a lifetime history of substance abuse.

Approximately 20% of the participants were recruited from theBronx VAMC’s and the Mount Sinai Medical Center’s Psychiatric Out-patient Clinics, or from referrals from outside mental health profession-als. All participants gave written informed consent after receiving acomplete description of the study. All participants received a medicalevaluation including history, physical and neurologic examination, andlaboratory testing. Patients with evidence of serious systemic illnesswhich might affect central nervous system functioning such as diabetes,hypertension, autoimmune illness, renal, liver, or cardiac disorderswere excluded. Additionally, patients with neurological impairment ora history of severe head trauma with loss of consciousness were ex-cluded. Participants were free from substance abuse for at least 6months and did not currently take medication for medical or psychiatricreasons.

Psychiatric Assessments

Participants completed extensive psychiatric and diagnostic assess-ment by trained raters, utilizing the SCID (Structured Clinical Interviewfor DSM-III) for Axis I disorders (Spitzer & Endicott, 1978) and the

72 JOURNAL OF TRAUMA & DISSOCIATION

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

SIDP-III-R (Structured Interview for DSM III-R Personality Disorder)for Axis II disorders (Pfohl & Zimmerman, 1989) (kappa = .81 forBPD). Unfortunately, data on Axis I dissociative disorders was not col-lected as the SCID-D (Structured Interview for DSM IV DissociativeDisorders) was not included, until recently, in the assessment batteryfor participants in this research program. Potential participants were ex-cluded if they met DSM-III-R (American Psychiatric Association,1984) criteria for schizophrenia or any schizophrenia-related psychoticdisorder or for bipolar (Type I) affective disorder. Meeting criteria forpast major depressive disorder did not exclude potential participants be-cause this is a common comorbidity in cluster B personality disorders.

Dissociation was measured by the Dissociative Experiences Scale(DES), extensively reviewed in the introduction. The DES has beenshown to have good internal consistency (Cronbach’s alpha = .95), andimpressive test-retest reliability = .79 to .96 (Carlson & Putnam, 1993).Respondents are asked to estimate the amount of time they experienceeach item using a response format of 0-100%. Total DES score are cal-culated as the mean response over all 28 items. The DES-Taxon (DES-T)score is based on DES item numbers 3, 5, 7, 8, 12, 13, 22, and 27 (seeAppendix). Pathological dissociation taxon membership was deter-mined by calculating posterior taxon membership probabilities basedon the weights for each DES-T item reported by Waller & Ross (1997).We also determined membership in a high DES-T group by taking theunweighted mean of the DES-T items and assigning all those above acut-off score (in this case, above 13 based on the finding of Simeon et al.(1998) to that group.

Suicide and self-mutilation information was gathered using theSIDP-III-R (Pfohl and Zimmerman, 1989) and Linehan’s “ParasuicideInterview”(Linehan, Wagner, & Cox, 1989) and assigned to categories.Suicide attempts were characterized in terms of intent, and the means bywhich the attempt was made. Acts of self-mutilation were defined asself-injurious behaviors that were not suicidal in intent and further char-acterized as having resulted in tissue damage or not. Estimates ofinterrater reliability for these ratings have not yet appeared in the litera-ture.

Childhood abuse experience was assessed using the Childhood TraumaQuestionnaire (CTQ; Bernstein & Fink, 1998), short form, a 28 itemself-report questionnaire that assess a broad range of abusive experi-ences including childhood sexual, physical and emotional abuse, andphysical and emotional neglect. The response format of the items is a5-point Likert-type scale with options ranging from “never true” to

Goodman et al. 73

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

“very often true.” Each sub-scale is derived from five items with scoresranging from 5-25. For example, on the physical abuse sub-scale, ascore of 20 is derived form answering “often true” to “having bruisesfrom being hit by a family member,” “requiring medical care fromphysical beatings,” “having a teacher, neighbor or doctor notice the ef-fects of beatings,” “punishment with a belt, board or other hard object,”and “having the belief that one was physically abused.” The 70 itemmeasure has been shown to have good internal consistency (Cronbach’salpha = .79-.94), and good test-retest reliability = .80-.83 (Ford, 1996).

Missing Data

Three participants were missing one of the DES-T items. A value foreach participant’s missing DES-T item was computed by averaging theremaining seven DES-T items. Probability scores were calculated withand without missing data without significant changes to assignment ofthe remaining 92 participants. Eight participants were missing onenon-DES-T item from the DES and two participants had two non-DES-T items missing. The mean value of the remaining DES items wasimputed for these values.

Two participants, both non-taxon members, were missing CTQ scaledata and were eliminated from those analyses.

Data Analysis

Frequency distributions of categorical variables were examined (i.e.,DES-T membership, diagnosis, suicide history, self-mutilation history,and gender) and associations were analyzed using chi-square. Mean dif-ferences in continuous variables such as CTQ scores and age were ana-lyzed using student’s t-test. Although we have directional hypotheses,two-tail tests at the p < .05 level were used.

To control for the inflated Type 1 error rate associated with multiplesignificance tests, the false discovery rate procedure developed byBenjamini and Hochberg (1995) was used to determine an adjusted sig-nificance threshold where appropriate. For each effect (i.e., BPD diag-nosis, SPD diagnosis, self mutilation and suicide history, and separateanalysis for six CTQ trauma variables), the actual p-values from thetests of each of the subscales were rank-ordered from smallest to larg-est. A cutoff significance threshold was calculated by multiplying 0.05by the rank number of each subscales result and dividing by the numberof subscales tested. For each effect, the p-value from each subscale re-

74 JOURNAL OF TRAUMA & DISSOCIATION

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

sult was then compared against the threshold p-value until a comparisonwas found that did not exceed the threshold. This p-value then becamethe significance level threshold for that particular effect.

Power analysis. We anticipated an unbalanced assignment of partici-pants to taxon and nontaxon groups. Taking the harmonic mean of asplit of 15 and 80 yielded an approximate n of 25 per group. With alphatwo-tailed at .05, this sample division has power of .80 to detect largeeffects and power of .64 to detect a medium effect size as defined byCohen (1988).

RESULTS

The mean DES score of the 95 participants was 17.4 ± 14.5 with arange from 0 to 79. The mean DES-T score was 10.0 ± 12.5 with a rangefrom 0 to 67. Posterior taxon probabilities ranged from 0 to 1 with a meanof 0.14 ± .32. Additionally, the base rate estimate was 0.16, with a typicaltaxon distribution of a sample with a low base rate (Meehl & Yonce, 1996,p. 1107). Fifteen individuals had probability scores greater than or equalto 0.66 and were classified as “definite” members of the pathological dis-sociation taxon (following the example of Woodward, Lenzenweger,Kagan, Snidman, & Arcus, 2000). Of the remaining 80 participants, 68had probability scores less than or equal to 0.0004 and an additional 12participants had probability scores between 0.024 and 0.063. These par-ticipants were designated as non-taxon members (see Figure 1).

There was no statistically significant difference in the racial/ethnicdistribution between taxon and non-taxon members; however, partici-pants who were taxon members had significantly fewer years of educa-tion 13.5 ± 2.1 compared to the non-taxon members (15.6 ± 2.6; t = 3.4,p < .01). There were no statistically significant differences in the preva-lence of MDD, current or history of PTSD, and history of substanceabuse among taxon and non-taxon members. There were significantlymore women in the taxon (9 of 15) compared to the non-taxon group (24of 80) (χ2 = 4.5, df = 1, p < .03, see Table 1). Membership in the patho-logical dissociation taxon was not significantly associated with BPD orSPD diagnoses (χ2 = 1.5, df = 1, ns, χ2 = 1.7, df = 1, ns, effect size 0.18and 0.17, respectively), nor was it associated with a history of suicideattempt or self injury (χ2 =.21, df = 1, ns; χ2 = .79, df = 1, ns, effect sizes0.07 and �0.14, see Table 1). Patients with BPD did not show signifi-cantly higher levels of pathological dissociation as compared to non-BPD patients (χ2 = 1.6, df = 1, ns).

Goodman et al. 75

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

Surprisingly, there was no significant group difference betweentaxon and nontaxon members on childhood abuse variables includingsexual, physical, and emotional abuse, emotional neglect, and totalscore as measured by the Childhood Trauma Questionnaire (CTQ). Ef-fect sizes for CTQ variables ranged from 0.21 to 0.45 (see Table 1). Ex-amination of frequency tables for each childhood abuse variable yieldedno particular pattern of values for pathological taxon members. More-over, the distribution of taxon members’ scores completely overlappedwith those of non-taxon members. Furthermore, past or present co-mor-bid Axis I diagnoses of major depressive disorder (MDD) and post-traumatic stress disorder (PTSD) were examined and found not to be as-sociated with taxon membership (for MDD χ2 = 2.053, df = 1, ns and forPTSD χ2 = .587, df = 1, ns).

Given that our results did not support our hypotheses, we chose to ex-amine pathological dissociation using the alternative approach noted inthe literature, of approximating taxon membership by use of the mean

76 JOURNAL OF TRAUMA & DISSOCIATION

70

60

50

40

30

20

10

0.00000 .05035 .80876 .99798

Taxon Probability

Num

ber

ofO

bser

vatio

nsFIGURE 1. Frequency distribution on taxon membership probabilities.

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

unweighted score of the DES-T items and a cut off score of 13 (seeMethods). Using this “approximation method,” the same 15 individualsinitially classified as pathological dissociation taxon members wereidentified, but also an additional 10, whose posterior probability scoreswere clearly not in the taxon range (mean = 0.03) with the largest beingonly 0.06. We are defining these 25 individuals as the high DES-T group.

In the analyses contrasting the high DES-T group with the remainderof the sample (see Table 2), several of our initial hypotheses were con-firmed. BPD diagnosis was significantly associated (15 of 37) withDES-T means greater than 13, compared to other personality disorderpatients (10 of 58, χ2 = 6.3, df = 1, p < .02), while SPD diagnosis wasnon-significant. Moreover, DES-T means over 13 were associated witha suicide attempt history (10 of 25 high DES-T group members had asuicide attempt, compared to 11 of 70 non-high mean DES-T individu-als; χ2 = 6.3, df = 1, p < .01) and a history of self-mutilation (13 of 25high mean DES-T individuals had a history of self-mutilation as com-pared to 15 of 70 non-high mean DES-T individuals; χ2 = 8.29, df = 1,p < .004). The severity of self-mutilation as assessed by the presence oftissue damage, however, was not significantly associated with averageDES-T scores over 13. There were significantly more women identified

Goodman et al. 77

TABLE 1. DES Taxon Members versus Non-Taxon Members Using BayesianProbability Method

Taxonmembers

Non-Taxonmembers

Effectsize

p value Total sample

Gender 6 males9 females

56 males24 females

0.31 0.03 62 male33 females

DES total 39.0 ± 15 12.5 ± 8.9 2.15 - 17.4 ± 14.5Mean probability 0.88 ± .13 0.007 ± .002 9.49 - 0.15 ± .33Age (years) 36.9 ± 8.5 38.8 ± 10.1 0.2 0.50 38.5 ± 9.8BPD diagnosis 8 (53%) 29 (36%) 0.18 0.26 37 (39%)SPD diagnosis 6 (40%) 19 (24%) 0.17 0.21 25 (26%)Suicide history 4 (27%) 17 (21%) 0.07 0.74 21 (22%)Self mutilation 2 (13%) 19 (24%) �0.14 0.51 21 (22%)CTQ-sexual abuse 8.7 ± 4.6 7.3 ± 4.3 0.31 0.25 7.5 ± 4.3CTQ-physical abuse 11.3 ± 6.1 8.9 ± 4.0 0.47 0.06* 9.2 ± 4.4CTQ-emotional abuse 15.4 ± 4.9 14.3 ± 5.7 0.21 0.49 14.4 ± 5.5CTQ-physical neglect 9.7 ± 3.5 8.2 ± 3.1 0.45 0.09 8.4 ± 3.1CTQ-emotional neglect 14.4 ± 5.4 16.6 ± 5.0 0.42 0.15 16.2 ± 5.1CTQ-total score 60.0 ± 16.4 55.3 ± 16.3 0.29 0.36 56.0 ± 16.3

* p < .008; significance level threshold per Benjamini and Hochberg (NS) (1995)

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

as high mean DES-T (14 of 25) compared to the non-high mean DES-Tgroup (10 of 70, χ2 = 6.8, df = 1, p < .009). Additional analyses co-vary-ing for female gender were performed for diagnoses, self-directed ag-gression, and trauma indices. Statistically significant results for BPDdiagnosis, self-mutilation, and suicide remained.

Lastly via this method, there was no significant group difference be-tween the high mean DES-T group and the its contrast group on anychildhood abuse variables as measured by the CTQ. Effect sizes for in-dividual CTQ variables ranged from 0.16 to 0.47 (see Table 2), consis-tent with our results for true pathological dissociation taxon members.

To explore the relationship between the two approaches, we calcu-lated the correlation between posterior probability scores and the meanunweighted score of the DES-T items. The resultant coefficient was.725, explaining only 53% of the variance. Although this value is highlysignificant it is not an acceptable level of precision necessary for the de-marcation between groups. To explore if the cut-off score was not ac-ceptable for our sample, we examined the relationship between cut-offscores other than 13 on the parametric mean of the DES-T and probabil-ity scores. There was no cut-off score that made fewer errors in replicat-ing the taxon based on posterior probability scores, further pointing todisparities between methodologies.

78 JOURNAL OF TRAUMA & DISSOCIATION

TABLE 2. DES Taxon Members versus Non-Members as Defined by Using the“Approximation Method” of the Parametric Mean of DES-T Items

High meanDES-T

dissociators

Low meanDES-T

dissociators

p value Size

Gender 11 males14 females

51 males19 females

.009** .30

BPD diagnosis 15 (60%) 22 (31%) .2* .30SPD diagnosis 10 (40%) 15 (21%) .07 .21Suicide history 10 (40%) 11 (16%) .01* .27Self mutilation 13 (52%) 15 (21%) .004* .33CTQ-sexual abuse 8.13 ± 4.1 7.4 ± 4.4 .53 .18CTQ-physical abuse 10.4 ± 5.4 8.8 ± 3.9 .21 .29CTQ-emotional abuse 15.5 ± 4.7 14.2 ± 5.8 .21 .27CTQ-physical neglect 8.8 ± 3.5 8.2 ± 2.9 .66 .19CTQ-emotional neglect 15.3 ± 5.4 16.6 ± 5.0 .27 .29CTQ-total score 58.1 ± 15.2 55.5 ± 16.6 .43 .16

* p < .038; significance level threshold per Benjamini and Hochberg (1995)** p < .05

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

DES total score (see Tables 3 and 4) was significantly associatedwith BPD diagnosis, suicide history, and history of self-mutilation, butnot gender, SPD diagnosis or presence of tissue damage. Pearson corre-lations between DES total scores and individual CTQ trauma sub-scalesand age (see Table 4) revealed only a significant association of physicalabuse and DES-total (r = .272, p < .008).

DISCUSSION

The presence of dissociative phenomena as measured by the totalDES score in our participants (17.4 ± 14.5) is comparable with levels ofother studies examining personality disorders (Brodsky et al., 1995; van

Goodman et al. 79

TABLE 3. DES Total by Gender, Diagnosis, Self Injury and Trauma Indices

DES total p value t df Effect size

Gender M-15.3 ± 13.3 .06 �1.931 93 .41F- 21.2 ± 15.9

BPD diagnosis 0- 13.6 ± 11.3 .004 * �3.036 56 .671- 23.2 ± 17.0

SPD diagnosis 0- 15.4 ± 12.3 .07 �1.853 32 .471- 22.8 ± 18.5

Suicide history 0- 15.5 ± 14.0 .022 * �2.33 93 .561- 23.6 ± 14.7

Self mutilation 0- 14.6 ± 11.7 .03* �2.86 93 .591- 23.6 ± 18.3

* p < .038; significance level threshold per Benjamini and Hochberg (1995)

TABLE 4. Pearson Correlation of DES Total by CTQ Trauma Sub-Scales andAge

Pearson correlation p value coefficients

CTQ-sexual abuse .128 .22CTQ-physical abuse .272 .008*CTQ-emotional abuse .122 .25CTQ-physical neglect .034 .74CTQ-emotional neglect �.180 .09CTQ-total score .100 .34Age �.066 .53

* p < .008; significance level threshold per Benjamini and Hochberg (1995)

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

Ijzendoorn & Schuengel, 1996), and suggests that moderate levels ofdissociative symptomatology are present in these individuals.

Our study yielded several important findings. First, contrary to ourhypotheses, pathological dissociation strictly defined by use of taxometricsand Baysian posterior probabilities using the weights reported by Wallerand Ross (1997) is not associated with personality diagnosis, history ofchildhood trauma, or self-directed injury in this population of personal-ity disordered individuals. Second, simplified non-weighted combina-tions of the DES-T items do not produce the same results as the use ofBayesian posterior probabilities, challenging the current literature’s useof this methodology. Third, this study serves as at least a partial replica-tion of the pathological dissociation taxon as evidenced by a base rateestimate of .16 and the distribution of Bayesian posterior probabilitiesin the sample (see Figure 1).

Use of the unweighted mean of DES-T items resulted in a groupcontrast that supported some of our hypotheses, i.e., personality disor-dered participants with a BPD diagnosis who self-mutilate, and have ahistory of suicide. Pathological dissociation defined through Baysianposterior probabilities yielded only gender and lower level of educationas correlates of taxon membership. This suggests that individuals withBPD who have high levels of dissociative experiences, evidence a mix-ture of non-pathological and pathological phenomena. Moreover, thismixed dissociation is also associated with self-directed aggression. Thispremise is consistent with our findings using total DES score and mirrorthe current literature which measures dissociation in a dimensionalfashion (Wilkinson-Ryan & Westen, 2000; Zanarini, Ruser, Franken-burg, Hennen, & Gunderson, 2000; Zweig-Frank, Paris, & Guzder1994a, 1994b, 1994c). Further research is necessary to delineate thecharacteristics of members of the pathological dissociation taxon. Apost-hoc analysis examining separately the individual criteria for BPDand taxon membership yielded a significant association between fear ofabandonment and membership in the pathological dissociation taxon(χ2 = 5.82 df = 1, p < .016). This suggests that future research examinevariables such as quality of attachment, interpersonal relating, and loss.

Theories of the role of self-directed aggressive behavior in BPD in-clude controlling overwhelming affective states, altering pain percep-tion and provision of psychic numbing, and disconnection from people(Favazza, 1989; Leibenluft, Gardner, & Cowdry, 1987; Stone, 1987).Russ et al. (1996) defines an “analgesic” subgroup of self-mutilatingBPD individuals who have more serious suicide attempts, impulsivity

80 JOURNAL OF TRAUMA & DISSOCIATION

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

and higher dissociative symptomatology compared to BPD patientswho feel pain during their acts of self-injury. This analgesic BPD groupmay overlap with our sample of female self-mutilating BPD partici-pants with non-pathological dissociation. It appears that self-directedaggression is not a behavioral correlate of pathological dissociation inthis sample of patients but may accompany other forms of dissociativeexperience.

Surprisingly, there was no association between membership in thepathological dissociation taxon and the sexual, physical or emotionalabuse and neglect variables of the CTQ. The same results were foundwith the high DES-T group contrast. In our study, only physical abusewas significantly associated with overall DES scores. The presence of arelationship between history of childhood abuse and dissociation is sup-ported by many studies (Chu & Dill, 1990; Herman, Perry, & van derKolk, 1989; Meehl, 1995; Shearer, 1994; van Ijendoorn & Schuengel,1996, Irwin, 1999) and contradicted by two others (Zweig-Frank et al.,1994a, 1994c). Recent literature highlights the role of childhood ne-glect in the development of adult personality disorder symptomatology(Johnson, Smailes, Cohen, Brown, & Bernstein, 2000) in the etiology ofdissociation (Draijer & Langeland, 1999), and in causing profound con-sequences to the developing brain (Glaser, 2000). Nonetheless, neitherabuse nor neglect was associated with pathological dissociation taxonmembership in this cohort of individuals.

The broad acceptance of a traumagenic model for the development ofdissociative symptomatology is not supported by our data for patholog-ical dissociation. However, less severe forms of dissociative experiencemay be more influenced by childhood traumatic experience. This iscontradicted by a recent study (Irwin, 1999) in normal individuals not-ing that pathological but not non-pathological dissociation was pre-dicted by indices of the CTQ. In this study, effect sizes were small (r =.21-.24), the level of dissociative symptoms was considerably lowerthan our study and pathological dissociation was not determined by useof Baysian posterior probabilities. Environmental factors other thanchildhood abuse or neglect, such as parental mental illness or forms ofbiparental failure may also be influencing the development of patholog-ical dissociation.

Conflicting data exists regarding the genetics of pathological dissoci-ation. Waller and Ross (1997) contend that non-pathological dissocia-tion but not pathological dissociation has genetic heritability. Jang andLivesley (1999) report genetic factors account for almost 50% of thevariance in DES-T scores and suggest that “genetic factors underlying

Goodman et al. 81

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

personality disorders are also related to dissociative capacity” (p. 350).These personality traits include: cognitive dysregulation, affective labilityand suspiciousness (Jang & Livesley, 1999); traits common to individu-als with SPD or BPD. However, this study, which examined generalpopulation twin pairs, did not report levels of dissociation at all, sug-gesting that it might have been low in absolute magnitude.

An examination of potential underlying biologic factors of dissocia-tion is needed as it may help differentiate pathological from non-patho-logical dissociation. There do exist important biological underpinningsof BPD symptomatology including aggressive impulsivity and suicidality(Coccaro et al., 1989). Additionally, these BPD symptoms involve ab-normalities in the serotonergic system and serotonin related genes(New, Goodman, Mitropoulou, & Siever, 2002). Interestingly, there ispreliminary work suggesting serotonergic dysregulation in dissociativespectrum symptomatology. Challenge studies with the partial serotoninagonist meta-chlorophenylpiperzaine (m-CPP) induced depersonaliza-tion in a double-blind, placebo controlled condition (Simeon et al.,1995) and psilocybin, a mixed 5-HT receptor stimulating agent producedderealisation phenomena in healthy controls (Vollenweider, Vontobel,Hell, & Leenders, 1999). Taken together, these findings potentially im-plicate the serotonergic system in dissociative phenomena. Future studyaddressing biologic parameters of dissociation both pathological andnon-pathological is warranted.

Limitations of this study include the validity of self-report measuresin assessing past abusive experiences. There is data to support the veri-fication of reported childhood events (Robins & Regier, 1991) but un-der-reporting is still a concern. However, in a meta-analysis of 24studies exploring the relationship between traumatic experiences andthe DES, there was no difference in effect size between studies that as-sessed trauma through interview or questionnaire (van IJzendoorn &Schuengel, 1996). Nonetheless, the possibility exists that with patho-logical dissociation, structured childhood trauma interviews might yieldmore detailed responses, and this study is currently underway.

Furthermore, the composition of our personality-disordered samplediffers from others presented in the literature, a factor that may accountfor our differing results using taxometric analyses. The majority of ourresearch participants were recruited from newspaper advertisement(> 80%), and not from clinical services. Our participants are slightlyolder, predominantly male, do not take medications or abuse sub-stances, and are typically functional members of the community. Thesecharacteristics notwithstanding, 22% of our sample did have a history

82 JOURNAL OF TRAUMA & DISSOCIATION

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

of suicide or self-mutilation attempt. Our sample differs from manyclinical samples reported in the literature, in that the latter are oftenmore functionally impaired, and consequently receiving mental healthservices including psychotropic medications. It is possible that using asample that more closely resembled that in the literature, our findingsbetween the two methodologies might have been different. Future stud-ies need to address in more functionally impaired samples similar tothose seen in clinical settings, whether pathological dissociation as de-fined by the use of Bayesian posterior possibilities differs from resultsusing an approximation. Such additional studies will help to contributeto the understanding of the nature of pathological dissociation. Spe-cifically, the etiological implications of pathological dissociation as ataxon are very different from the implications from pathological disso-ciation being just more extreme versions of the dissociation continuum.

One important way that these etiological implications might be in-vestigated is by studying the effect of certain receptor subtypes of partic-ular neurotransmitter systems (e.g., serotonin) on the pattern of symptomsthat are manifest. Such a finding would be much more likely to emergefrom a taxonomic approach that used these and more clinical variablesthan through standard dimensional or grouping analyses. Moreover, apositive finding of differences in receptor subtypes would be virtuallyimpossible to notice clinically. Thus, the taxometric approach holdsmore promise for identifying a biological etiology of pathological dis-sociation than any other method currently available.

An additional limitation concerns the use of conservative statisticalmethodology. Given that our sample sizes are small, and tests of statisti-cal significance conservative, use of a non-clinical control group mighthave improved the statistical power of the study. Future studies shouldinclude a more homogenous population in terms of personality diagno-sis (SPD or BPD), larger number of participants, employ a non-clinicalcontrol group and attempt to identify other defining characteristics oftaxon membership.

In summary, this study examined dissociation in personality disor-ders with the use of taxon methodology. The goal of the taxometric ap-proach is to ultimately eliminate classification on the basis of a cut offscore on a dimensional variable, and to use a combination of indicatorsto give unambiguous probabilities of membership in or out of the taxon(Tellegen & Atkinson, 1974). We found that gender but not indices ofpersonality disorder diagnosis, nor self-directed aggression predictedmembership in the pathological dissociation taxon. Furthermore, ourresults did not support the traumagenic theory of dissociative tenden-

Goodman et al. 83

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

cies, at least not the most severe aspect of dissociation. Lastly, our datasuggest that the use of an “approximation” methodology, i.e., paramet-ric mean of the DES-T items in place of strict taxometric probabilitycalculations yield results more consistent with a dimensional view ofdissociation.

REFERENCES

American Psychiatric Association (1984). Diagnostic and statistical manual of mentaldisorders iii-revised. Washington, DC: Author.

American Psychiatric Association (1994). Diagnostic and statistical manual of mentaldisorders IV. Washington, DC: Author.

Bauer, A.M., & Power, K.G. (1995). Dissociative experiences and psychopathologicalsymptomatology in a Scottish sample. Dissociation, 8, 209-219.

Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practi-cal and powerful approach to multiple testing. Journal of the Royal Statistical SocietyB, 57, 289-300.

Bernstein, D.P., & Fink, L.A. (1998). Childhood Trauma Questionnaire: A retrospec-tive self-report. San Antonio, TX: Psychological Corp.

Bernstein, E.M., & Putnam, F.W. (1986). Development, reliability and validity of adissociation scale. Journal of Nervous and Menal Disease, 174, 727-735.

Briere, J., & Runtz, M. (1988). Symptomatology associated with childhood sexual vic-timization in a non-clinical adult sample. Child Abuse & Neglect, 12, 51-59.

Brodsky, B.S., Cloitre, M., & Dulit, R.A. (1995). Relationship of dissociation toself-mutilation and childhood abuse in borderline personality disorder. AmericanJournal of Psychiatry, 152, 1788-1792.

Carlson, E.B., & Putnam, F.W. (1993). An update on the dissociative experiencesscale. Dissociation, 6, 16-27.

Chu, J.A., & Dill, D.L. (1990). Dissociative symptoms in relation to childhood physi-cal and sexual abuse. American Journal of Psychiatry, 147, 887-892.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.).Hillsdale, NJ: Erlbaum.

Coccaro, E.F., Siever, L.J., Klar, H.M., Maurer, G., Cochrane, K., Cooper, T.B., Mohs,R.C., & Davis, K.L. (1989). Serotonergic studies in patients with affective and per-sonality disorders. Correlates with suicidal and impulsive aggressive behavior. Ar-chives of General Psychiatry, 46, 587-599.

Draijer, N., & Langeland, W. (1999). Childhood trauma and perceived parental dys-function in the etiology of dissociative symptoms in psychiatric inpatients. Ameri-can Journal of Psychiatry, 156, 379-385.

Dulit, R.A., Fyer, M.R., Leon, A.C., Brodsky, B.S., & Frances, A.J. (1994). Clinicalcorrelates of self-mutilation in borderline personality disorder. American Journal ofPsychiatry, 151, 1305-1311.

Favazza, A.R. (1989). Why patients mutilate themselves. Hospital Community Psychi-atry, 40, 137-145.

84 JOURNAL OF TRAUMA & DISSOCIATION

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

Ford, J. (1996). Psychometric review of Childhood Trauma Questionnaire. In B.Stramm (Ed.), Measurement of stress, trauma and adaptation (pp. 94-95). Lutherville,MD: Sidran Press.

Glaser, D.J. (2000). Child abuse and neglect and the brain–a review. Journal of ChildPsychology and Psychiatry, 41(1), 97-116.

Heffernan, K., & Cloitre, M.J. (2000). A comparison of posttraumatic stress disorderwith and without borderline personality disorder among women with a history ofchildhood sexual abuse: Etiological and clinical characteristics. Journal of Nervousand Mental Disease, 188, 589-595.

Herman, J.L., Perry, J.C., & van der Kolk, B.A. (1989). Childhood trauma in border-line personality disorder. American Journal of Psychiatry, 146, 490-495.

Irwin, H.J. (1994). Proneness to dissociation and traumatic childhood events. Journalof Nervous and Mental Disease, 182, 456-460.

Irwin, H.J. (1998). Dissociative tendencies and the sitting duck: Are self reports of dis-sociation and victimization symptomatic of neuroticism? Journal of Clinical Psy-chology, 54, 1005-1015.

Irwin, H.J. (1999). Pathological and nonpathological dissociation: The relevance ofchildhood trauma. Journal of Psychology, 133, 157-164.

Jang, K.L., & Livesley, W.J. (1999). Why do measures of normal and disordered per-sonality correlate? A study of genetic comorbidity. Journal of Personal Disorders,13(1):10-7.

Jang, K.L., Paris, J., Zweig-Frank, H., Livesley, W.J. (1998). Twin study of dissociativeexperience. Journal of Nervous and Mental Disease, 186, 345-351.

Kemperman, I., Russ, M.J., Clark, W.C., Kakuma, T., Zanine, E., & Harrison, K.(1997). Pain assessment in self-injurious patients with borderline personality disor-der. Psychiatry Research, 70(3): 175-183.

Johnson, J.G., Smailes, E.M., Cohen, P., Brown, J., & Bernstein, D.P. (2000). Associa-tions between four types of childhood neglect and personality disorder symptomsduring adolescence and early adulthood: Findings of a community-based longitudi-nal study. Journal of Personality Disorders, 14, 171-187.

Leibenluft, E., Gardner, D., & Cowdry, R. (1987). The inner experience of the border-line self-mutilator. Journal of Personality Disorders, 1, 317-324.

Linehan, M.M., Wagner, A.W., & Cox, G. (1983). Parasuicide history interview:Comprehensive assessment of parasuicidal behavior. Seattle, WA: University ofWashington.

Links, P.S., & van Reekum, R. (1993). Childhood sexual abuse, parental impairmentand the development of borderline personality disorder. Canadian Journal of Psy-chiatry, 38, 472-474.

Meehl, P.E. (1995). Bootstraps taxometrics: Solving the classification problem inpsychopathology. American Psychologist, 50, 266-275.

Meehl, P.E., & Yonce, L.J. (1996). Taxometric analysis: II. Detecting taxonicity usingcovariance of two quantitative indicators in successive intervals of a third indicator(Maxcov procedure). Psychological Reports, 78, 1091-1227.

New, A.S., Goodman, M., Mitropoulou, V., & Siever L.J. (2002). Genetic poly-morphisms and aggression. In J. Benjamin, R. Ebstein, & R. Belmaker (Eds.), Mo-

Goodman et al. 85

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

lecular genetics and the human personality (pp. 231-244). Washington, DC: AmericanPsychiatric Press.

Ogata, S.N., Silk, K.R., Goodrich, S., Lohr, N.E., Westen, D., & Hill, E.M. (1990).Childhood sexual and physical abuse in adult patients with borderline personality.American Journal of Psychiatry 147, 1008-1013.

Ozer, E.J., Best, S.R., Lipsey, T.L., & Weiss, D.S. Predictors of posttraumatic stressdisorder diagnosis and symptoms in adults: A meta-analysis. Manuscript under re-view.

Paris, J. (1998). Does childhood trauma cause personality disorders in adults? Cana-dian Journal of Psychiatry, 43, 148-153.

Pfohl, B., & Zimmerman, M. (1989). Structured interview for the diagnosis of DSM-III-Rpersonality disorders. Ames, IO: University of Iowa College Medicine.

Putnam, F.W. (1991). Dissociative phenomena. In A. Tasman & S.M. Goldfinger(Eds.), American Psychiatric Press review of psychiatry, Vol. 10 (pp. 145-160).Washington, DC: American Psychiatric Press.

Putnam, F.W. (1995). Development of dissociative disorders. In D. Ciccheti & D.J.Cohen (Eds.), Development psychopathology. Vol. 2. Risk, disorder and adaption.(pp. 581-608). New York: Wiley.

Robins, L.N., & Regier, D.A. (1991). Psychiatric disorders in America. New York:The Free Press.

Ross, C.A., Joshi, S., & Currie, R. (1991). Dissociative experiences in the general pop-ulation: A factor analysis. Hospital and Community Psychiatry, 42, 297-301.

Russ, M.J, Crawford, W.C., Cross, L.W., Kemperman, I., Kakuma, T., & Harrison, K.(1996). Pain and self injury in borderline patients: Sensory decision theory, copingstrategies, and loss of control. Psychiatry Research, 63, 57-65.

Salzman, J.P. (1998). Etiology of borderline personality disorder. American Journal ofPsychiatry, 155, 1640-1641.

Shearer, S.L. (1994). Dissociative phenomena in women with borderline personalitydisorder. American Journal of Psychiatry, 151, 1324-1328.

Simeon, D., Guralnik, O., Gross, S., Stein, D.J., Schmeidler, J., & Hollander, E.(1998). The detection and measurement of depersonalization disorder. Journal ofNervous and Mental Disease, 186, 536-542.

Simeon, D., Guralnik, O., Schmeidler, J., Sirof, B., & Knutelska, M. (2001). The roleof childhood interpersonal trauma in depersonalization disorder, American Journalof Psychiatry, 158, 1034-1039.

Simeon, D., Hollander, E., Stein, D.J., DeCaria, C.M., Cohen, L., Saoud, J., Islam, N.,& Hwang, M.N. (1995). Induction of depersonalization by the serotonin agonistmeta-chlorophenylpiperazine. Psychiatry Research, 58, 161-164.

Spiegel, D., & Cardena, E. (1991). Disintegrated experience: The dissociative disor-ders revisited. Journal of Abnormal Psychology, 100, 366-378.

Spitzer, R.L., & Endicott, J. (1978). Schedule for affective disorders and schizophre-nia. New York, NY: New York State Psychiatric Institute.

Startup, M. (1999). Schizotypy, dissociative experiences and childhood abuse: Rela-tionship among self-report measures. British Journal of Clinical Psychology, 38,333-344.

86 JOURNAL OF TRAUMA & DISSOCIATION

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

Stone, M. (1987). A psychodynamic approach: Some thoughts on the dynamics andtherapy of self-mutilating borderline patients. Journal of Personality Disorders, 1,347-349.

Tellegen, A., & Atkinson, G. (1974). Openness to absorbing and self-altering experi-ences (“absorption”), a trait related to hypnotic susceptibility. Journal of AbnormalPsychology, 83, 268-277.

Tampke, A.K., & Irwin, H.J. (1999). Dissociative processes and symptoms of post-traumatic stress in Vietnam veterans. Journal of Traumatic Stress, 12, 733-746.

Trull, T.J. (2001a). Structural relations between borderline personality disorder fea-tures and putative etiological correlates. Journal of Abnormal Psychology, 3,471-481.

Trull, T.J. (2001b). Relationships of borderline features to parental mental illness,childhood abuse, Axis I disorder, and current functioning. Journal of PersonalityDisorders, 15, 19-32.

Van der Kolk, B.A., Perry, J.C., & Herman, J.L. (1991). Childhood origins of self-de-structive behavior. American Journal of Psychiatry, 148, 1665-1671.

Van IJzendoorn, M.H., & Schuengel, C. (1996). The measurement of dissociation innormal and clinical populations: Meta-analytic validation of the dissociative expe-riences scale (DES). Clinical Psychology Review, 16, 365-382.

Vollenweider, F.X., Vontobel, P., Hell, D., & Leenders, K.L. (1999). 5-HT modulationof dopamine release in basal ganglia in psilocybin-induced psychosis in man–a PETstudy with [11C]raclopride. Neuropsychopharmacology, 20, 424-433.

Waller, N.G., & Meehl, P.E. (1998). Multivariate taxometric procedures: Distin-guishing types from continua. Thousand Oaks, CA: Sage Publications.

Waller, N.G., Putnam, F.W., & Carlson, E.B. (1996). Types of dissociation anddissociative types: A taxometric analysis of dissociative experiences. Psychologi-cal Methods, 3, 300-321.

Waller, N.G., & Ross, C.A. (1997). The prevalence and biometric structure of patho-logical dissociation in the general population: Taxometric and behavior geneticfindings. Journal of Abnormal Psychology, 106, 499-510.

Wilkinson-Ryan, T., & Westen, D. (2000). Identity disturbance in borderline personal-ity disorder: An empirical investigation. American Journal of Psychiatry, 157,528-541.

Woodward, S.A., Lenzenweger, M.F., Kagan, J., Snidman, N., & Arcus, D. (2000).Taxonic structure of infant reactivity: Evidence from a taxometric perspective. Psy-chological Science, 11, 296-301.

Zanarini, M.C. (2001). Childhood experiences associated with the development of bor-derline personality disorder. Psychiatric Clinics of North America, 23, 89-101.

Zanarini, M.C., Ruser, T., Frankenburg, F.R., Hennen, J. (2000). The dissociative ex-periences of borderline patients. Comprehensive Psychiatry, 41, 223-227.

Zanarini, M.C., Ruser, T., Frankenburg, F.R., Hennen, J., & Gunderson, J.C. (2000).Risk factors associated with the dissociative experiences of borderline patients.Journal of Nervous and Mental Disease, 188, 26-30.

Zanarini, M.C., Williams, A.A., Lewis, R.E., Reich, R.B., Vera, S.C., Marino, M.F.,Levin, A., Yong, L., & Frankenburg, F.R. (1997). Reported pathological childhood

Goodman et al. 87

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015

experiences associated with the development of borderline personality disorder.American Journal of Psychiatry, 154, 1101-1105.

Zelkowitz, P., Paris, J., Guzder, J., & Feldman, R. (2001). Diatheses and stressors inborderline pathology of childhood: The role of neuropsychological risk and trauma.Journal of the American Academy of Child and Adolescent Psychiatry, 40, 100-105.

Zweig-Frank, H., Paris, J., & Guzder, J. (1994a). Dissociation in female patients withborderline and non-borderline personality disorder. Journal of Personality Disor-ders, 8, 203-209.

Zweig-Frank, H., Paris, J., & Guzder, J. (1994b). Dissociation in male patients withborderline and non-borderline personality disorder. Journal of Personality Disor-ders, 8, 210-218.

Zweig-Frank, H., Paris, J., & Guzder, J. (1994c). Psychological risk factors for dissoci-ation and self-mutilation in female patients with borderline personality disorders.Canadian Journal of Psychiatry, 39, 259-264.

RECEIVED: 10/26/01REVISED: 04/26/02

08/21/02ACCEPTED: 08/21/02

APPENDIX. Pathological Dissociation DES Taxon Items

1. Item #3 – having the experience of finding oneself in a place and havingno idea how one got there

2. Item #5 – having the experience of finding new things among one’sbelongings that one does not remember buying

3. Item #7 – having the experience of seeing oneself as if looking at anotherperson

4. Item #8 – having the experience of not recognizing friends or familymembers

5. Item #12 – having the experience of feeling that other people, objectsand the world are not real

6. Item #13 – having the experience that one’s body is not one’s own

7. Item #22 – having the experience of feeling as though one were twodifferent people

8. Item #27 – having the experience of hearing voices inside their heads

88 JOURNAL OF TRAUMA & DISSOCIATION

Dow

nloa

ded

by [

New

Yor

k U

nive

rsity

] at

06:

53 2

4 Fe

brua

ry 2

015