Dissociation is generally viewed as a disruption in the usually integrated functions of consciousness, memory, identity, and perception of the environment (DSM-IV). It has widely been associated with traumatic events, in particular with posttraumatic stress disorder (PTSD) (1–3), childhood sexual and physical abuse, and disruptions of parental care (4–6). Traditionally, dissociative symptoms have been related to hysteria (7), but in recent years, they have also been studied in a variety of other clinical conditions including somatization (8, 9), self-mutilation (6), and borderline personality disorder (4–6). Furthermore, dissociation has been associated with other psychological constructs such as hypnotizability (3) and alexithymia (6). However, the basic question of nature and nurture of dissociative symptoms has hardly been answered.

In this study, we used Cloninger’s theoretical concept of temperament and character (10, 11) to evaluate the association of personality traits to dissociative symptoms. The four temperament dimensions (novelty-seeking, harm avoidance, reward dependence, and persistence) are thought to reflect inherited behavior. They were found to be homogeneous and independent from each other and to determine automatic emotional responses and habits toward external and internal stimuli (12). In contrast, character dimensions (self-directedness, cooperativeness, and self-transcendence) are thought to be environmentally determined by integrating biographical experiences to generate a self-concept and a view of others and one’s role as a human being.

Given the well-established association of dissociative phenomena with traumatic experiences, we assumed a relationship to nongenetic, environmentally determined character traits. Because dissociation does not differ between genders, we also hypothesized that the same personality factors are related to dissociative symptoms in men and women, although they may have been exposed to different traumatic events in childhood and adulthood.

We included 191 inpatients and outpatients of our department of psychiatry as well as 41 nonclinical subjects. The group consisted of 131 women (mean age=39.3 years, SD=13.1) and 101 men (mean age=40.0 years, SD=13.6). Their DSM-IV diagnoses consisted of major depression (N=43), anxiety disorders (N=33), obsessive-compulsive disorder (N=27), schizophrenia and schizoaffective disorder (N=32), personality disorder not otherwise specified (N=17), borderline personality disorder (N=12), dissociative disorder (N=10), alcohol dependence (N=9), and somatization disorder (N=8).

All participants were Caucasian and gave written informed consent after receiving a full explanation of the study.

Dissociative symptoms were evaluated by means of the German version of the Dissociative Experience Scale (13). Personality traits were assessed by means of the German version of the Temperament and Character Inventory (14), a 240-item, self-report scale. The SCL-90-R (15), a 90-item, self-report scale, was used to assess current psychopathology (global severity index score).

Regression analysis was used to assess the influence of the Temperament and Character Inventory dimensions and the global severity index score on the Dissociative Experience Scale total score. We calculated t tests (two-tailed) for independent groups, multivariate analysis of variance (MANOVA), and Pearson’s correlations (bivariate) where indicated.

The mean Dissociative Experience Scale scores were 11.3 (SD=11.6) for women and 10.1 (SD=9.1) for men (t=0.84, df=230, p=0.40). The mean global severity index scores were 1.0 (SD=0.8) for women and 0.9 (SD=0.7) for men (t=0.90, df=230, p=0.37). For Temperament and Character Inventory dimensions, no significant gender differences were found with MANOVAs (Wilks’s lambda=0.95, F=1.85, df=7, 223, p=0.08).

In women, the Temperament and Character Inventory dimensions of self-transcendence (standardized regression coefficient beta=0.23, t=3.1, df=129, p=0.002), self-directedness (beta=–0.28, t=–2.5, df=129, p=0.013), and the global severity index scores (beta=0.48, t=4.8, df=129, p<0.001) were predictors for the Dissociative Experience Scale scores (F=10.53, df=8, 121, p<0.001). In men, scores for self-transcendence (beta=0.24, t=2.4, df=99, p=0.02), self-directedness (beta=–0.2, t=–1.7, df=99, p=0.09), and the global severity index (beta=0.44, t=3.6, df=99, p<0.005) predicted the Dissociative Experience Scale scores (F=8.57, df=8, 92, p<0.001). No temperament dimensions showed any significant predictive power. No correlation was observed between self-transcendence and self-directedness scores (r=0.04, N=208, p=0.52).

The present study examined the question of nature and nurture of dissociative phenomena by using Cloninger’s model of personality. Our findings support the hypothesis that in both genders, only environ­mentally determined character factors were related to   dissociation. Low self-directedness and high self-­transcendence scores were significant predictors of dissociation in both sexes.

With Cloninger’s theory of personality, these results seem plausible because character dimensions are thought to develop in response to environmental stimuli during childhood and adolescence (10, 11). Correspondingly, dissociation is considered to reflect abusive childhood experiences and traumatic events in general (1–6). Persons with a low degree of self-directedness are immature, destructive, irresponsible, unreliable, and have a low social integration. These character traits are frequently found in psychiatric patients, particularly in those with borderline personality disorder (4–6), who also show a high degree of dissociation. In contrast, subjects with high self-transcendence scores are patient, satisfied, creative, ingenious, and rich in fantasy.

Cloninger’s theoretical ideas also assume that temperament is largely based on genetic factors (12). Therefore, our findings might suggest that genetic factors play a minor role in dissociative phenomena. However, genetic factors may exist that contribute independently from personality dimensions to a specific biological liability for dissociative experiences (16).

A strength of our results is the stability of the association between the character dimensions and dissociation, even when we controlled for the major influence of global psychopathology (global severity index score) on dissociation.

A limitation of our study was that we were not able to differentiate the putative sources of dissociation (e.g., childhood traumata versus PTSD) and whether or not a specific trauma history is associated with either low self-directedness or high self-transcendence scores.

Moreover, a further biological and psychological validation of Cloninger’s theory is necessary to properly interpret our findings in regard to the nature and nurture of dissociative symptoms. Because of the complexity of the subject, we consider our results to represent a hypothesis concerning the predisposition or the cause of dissociative symptoms; moving beyond the stage of hypothesis building will require much additional research.

Received Aug. 13, 1998; revisions received Jan. 21 and May 18, 1999; accepted May 26, 1999. From the Department of Psychiatry, Ernst-Moritz-Arndt-University Greifswald. Address reprint requests to Dr. Grabe, Department of Psychiatry, Ernst-Moritz-Arndt-University Greifswald, Klinikum der Hansestadt Stralsund, Rostocker Chaussee 70, 18437 Stralsund, Germany