The French Version of the Defense Style Questionnaire

Transcription

1 Regular Article Psychother Psychosom 1998;67:24 30 Ch. Bonsack J.N. Despland J. Spagnoli Département Universitaire de Psychiatrie Adulte (DUPA) Lausanne, Suisse The French Version of the Defense Style Questionnaire OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO Key Words Defense mechanisms Rating scales Test validity Test reliability Personality measures OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO Abstract Background: Bond et al. developed the Defense Style Questionnaire (DSQ), a self-questionnaire that aims at empirically measuring conscious derivatives of defense mechanisms. The original analysis yielded 4 factors called Defense Styles (DS). DSQ discriminates between mature and immature defense styles. Objectives: Determine if the French version of DSQ has (a) face validity, (b) a similar structure to the original version, (c) internal consistency, (d) grouping of defense mechanisms into clinically pertinent defense styles, (e) evidence of nonpatients using more mature defense styles, (f) correlation with Defensive Functioning Scale (DFS) (DSM-IV) [2]. Methods: Reliability and validity study on 82 control subjects and 140 patients, 59 among them evaluated for defensive level. Results: Factor analysis of controls sample yielded 4 factors ranging from immature to mature defense styles. DSQ scores on factor I (maladaptive style) are significantly higher in outpatients than in controls. Maladaptive style score correlates with clinical evaluation of defensive level of functioning (DFS). Conclusion: Psychometric features of the French version are similar to the original scale, although minor differences in individual defense mechanisms are present. Factor I (maladaptive defense style) remains more stable than other factors, accounts for most of variance contribution, has high internal consistency and applies to behaviors, i.e. conscious derivatives of defense mechanisms that can be easily identified. The French version of DSQ is (a) an easy and economical way to rate immature defense style in populations of neurotic and borderline patients and (b) further provides a hierarchical grouping of defense mechanisms in defense styles. OOOOOOOOOOOOOOOOOOOOOO Introduction Defense mechanisms is a psychoanalytic concept that has been well operationalized for empirical studies [1]. Indeed, it is the first psychoanalytic concept to be included in the fourth Diagnostic and Statistical Manual of Mental Disorder as a proposed axis for further study. In this manual, defense mechanisms are considered equivalent to coping mechanisms and defined as automatic psychological processes that protect the individual against anxiety and from the awareness of internal or external dangers or stressors [...]. The individual defense mechanisms are divided conceptually and empirically into related groups that are referred to as Defense Levels [2]. Considering coping and defense styles as equivalent can be confusing, and others prefer to differentiate both concepts, coping mechanismes as being flexible and positively oriented toward adaptation to reality and defense ABC Fax karger@karger.ch S. Karger AG, Basel /98/ $15.00/0 This article is also accessible online at: Charles Bonsack, MD Policlinique Psychiatrique A Sévelin 18 CH 1004 Lausanne (Switzerland) Tel , Fax , Charles.Bonsack@inst.hospvd.ch

2 mechanisms as being more rigid and more oriented toward inner conflicts [3]. From a theoretical point of view, these two concepts are quite different, but from an experimental point of view, more mature defense styles and coping styles can be considered as equivalent. Laplanche and Pontalis [4] argued that S. Freud s choice of the word mechanisms is intended, from the outset, to indicate the fact that psychical phenomena are so organized as to permit scientific observation and analysis [5]. However, consensus on observations proved to be difficult and definitions were more linked to metapsychology or theoretical conceptions. A. Freud [6] has defined mechanisms of defense very close to Freudian psychoanalytic theory, whereas M. Klein [7] described more primitive defense mechanisms like splitting and projective identification. At that time, defense mechanisms were mainly considered as pathological mechanisms which should be analyzed and overcome. Study of patients who were not accessible to classical psychoanalysis, however, led to the consideration of defense mechanisms also as coping mechanisms, which are more or less adapted reactions to stressors. In addition, Vaillant [8] showed that defense mechanisms can be ranked empirically in a hierarchy along two dimensions: (a) immaturity-maturity and (b) psychopathology-mental health. Defense mechanisms are partly unconscious process and cannot be observed without some minimal inferences. There are four methods that have been used for the empirical assessment of defense mechanisms: clinical interview, self-report questionnaire, percept-genetic technique and projective test [9, 11]. According to Skodol and Perry [10], methods of rating of interview material (such as the Defense Mechanisms Rating Scales (DMRS) and the Clinical Assessment of Defense Mechanisms [11]) have placed a premium on the objective rating of defenses for research purpose. Projective testing is a powerful method of revealing defense mechanisms in a regressive situation, and of anticipating the patient s reaction to a psychoanalytic setting. However, this regressive potential does not reflect current defensive functioning and can be quite different from clinical evaluation. The use of selfreport questionnaires, such as Bond s Defense Style Questionnaire (DSQ) [12], raised much controversy about their ability to self-report partly unconscious mechanisms [1]. These are considered as less valid instruments, but are also much more simple and economical to use. They have the advantage of potentially enhanced reliability due to the elimination of observer judgment, but have the disadvantage of accessing only conscious derivatives of defenses [10]. Bond s DSQ (1984 version) has 88 items, each of them attributed to one among 24 individual defense mechanisms. Original analysis of the DSQ [5] yielded 4 groups of defense mechanisms which were called defense styles: Maladaptative action pattern, Image distorting style, Self-sacrificing style and Adaptative defense style. This questionnaire has been translated and validated in different languages and environments [13 17]. According to Bond, DSQ scores correlate with global mental health, physical health, level of ego development, marital and career satisfaction, and happiness, but are not generally associated with broad categories of psychiatric diagnosis [18]. However, the correlations with the other three methods are usually quite low [19]. Objectives: Our objective was to study the reliability and validity of the French version of DSQ with simple clinical measures and to compare its psychometric features with the original and other versions. We tested the following hypotheses: (a) factor structure and internal consistency are similar to the original version, (b) DSQ discriminates between control subjects and patients, and (c) DSQ scores correlate with clinical evaluation of defensive functioning level. Subjects and Methods Methods of Ratings Diagnosis was based on ICD 10. Global level of functioning was assessed using the Global Assessment of Functioning scale (GAF) [2]. The Defensive Functioning Scale (DFS) was used as a structured clinical evaluation of defensive levels.the DFS scale is a proposed axis for further study in the DSM-IV. Ratings are based on a three step process: (a) identify individual defense mechanisms following the glossary s definitions, (b) classify defense mechanisms according to their use by the patient, and (c) rate the current defensive level. Defense mechanisms are grouped in 7 hierarchical defensive levels. Each level has characteristic defense mechanisms, for example in the action level there is acting out, withdrawal, help rejecting complaining and passive aggression [2]. Bond s DSQ (1984 version) has 88 items, each of them attributable to one of 24 individual defense mechanisms. Answers are on a 9-degree Likert scale from Strongly disagree to Strongly agree. The number of questions relating to each defense mechanism varies from 1 to 9 (median = 3). The subject s score for a defense mechanism is the mean of scores of items attributed to this mechanism. Samples Our samples are 82 control subjects from the general population, and 140 outpatients who voluntarily completed the DSQ. Sociodemographic data of the whole sample are shown in table 1. Control subjects were mainly clients of a drugstore and students. In this sample there was a net predominance of females (79%), with an average age of 33 years. 140 patients were volonteers from outpatients institutions of the Adult Psychiatric Department of Lausanne (Policlinique Psychiatrique A (90) et B (7), Clinique du Vallon (30), Département Autonome de Médecine Psychosociale et Centre d intervention thérapeutique brève (13)). Females and males were evenly distribut- French Version of DSQ Psychother Psychosom 1998;67:

3 Age Sex Table 1. Sociodemographic characteristics of the whole sample Mean SD Female Male Patients (n = 140) Controls (n = 82) (47.2%) 66 (52.8%) 19 subjects filled out the DSQ anonymously (79.5%) 16 (20.5%) ed in the patient s sample. The clinical sample is a subsample of 59 patients from PPUA (table 2). This clinic has an open consultation for any outpatients having mental disorders, from adjustment disorders to schizophrenia. Treatments are psychodynamically oriented, without excluding medication or social support. Most psychiatrists are residents in training and have a clinical supervision. Results Reliability We examined 3 aspects of DSQ reliability: (a) individual defense mechanisms associated with factors structure, i.e. grouping of defense derivatives in defense styles, (b) internal consistency of each defense style, (c) comparison with previous studies (see conclusion). Factor Grouping of Individual Defense Mechanisms. To obtain as large a panel of defense mechanisms as possible, we performed factor analysis on the control sample (n = 82). The goal of factor analysis is to reduce the number of variables, i.e. to indicate which defense mechanisms are related to each other and can be grouped in socalled defense styles. Factor analysis of the 24 individual defense mechanisms scores yielded 4 factors (defense styles) ranging from immature to mature (table 3). Factor I could be called maladaptive style, but the remaining factors are far more difficult to name. Factor II is close to neurotic style, factor III to minor distorting style and factor IV to adaptive style. Factor I (maladaptive style) accounts for 15.5% of variance, the other factors respectively for 10, 9 and 7% of variance. Factor analysis of patients (n = 140) yielded the same core defense mechanisms, but factor I includes undoing, passive aggression, autistic fantasy and splitting. Apart from undoing, all these defense mechanisms are usually rated as immature. The other factors are less stable. Internal consistency of each defense style, which tests the degree of relatedness between individual defense mechanisms within each defense style, i.e., reproducibility of measurement across Table 2. Sociodemographic and clinical data for the clinical sample Age Sex Principal diagnosis (CIM-10) GAF Defensive levels (DFS) Mean SD Female Male F1 Substances F2 Psychosis F3 Affective F4 Neurosis F5 Behavioural F6 Personality Z Other conditions Mean SD Defensive dysregulation Action Major image distorting Disavowal Minor image distorting Mental inhibition Patients (n = 59) (52.5%) 28 (47.5%) 4 (7.0%) 11 (19.3%) 7 (12.3%) 20 (35.1%) 1 (1.8%) 10 (17.5%) 4 (7.0%) (15.3%) 11 (18.6%) 9 (15.3%) 14 (23.7%) 12 (20.3%) 4 (6.8%) defense mechanisms within each factor, is sufficient only for factor I (Cronbach s alpha = 0.71). However, these low Cronbach alphas for factors II, III and IV must be cautiously interpreted, because of the low number of individual defense mechanisms in these factors. Validity We tested 3 aspects of the DSQ validity: does the DSQ (a) adequatly probe derivatives of defense mechanisms (content validity), (b) discriminate between patients and probants and correlate with clinical evaluation of defensive functioning level (criterion validity), (c) account for a hierarchical grouping of defense mechanisms within defense styles (construct validity) [20]. Content Validity. 5 independent experts examined DSQ face validity. C.B., J.-N.D., M.S. and J.B. are medical doctors specialized in psychiatry and psychotherapy; Y. de R. has a PhD in psychology. All have experience in clinical psychotherapy and psychotherapy research. C.B, J.-N.D., Y. de R. and M.S. are trained for DMRS. They attributed to each item one of the 24 defense mechanisms used in the DSQ. When available, definitions of DSM-IV [2] were used as guide. Only 5 questions produced no consensus. 3 of these could be suppressed, because they refer to lie, which is not commonly accepted as a defense 26 Psychother Psychosom 1998;67:24 30 Bonsack/Despland/Spagnoli

4 Table 3. Factor analysis of the French version of DSQ Defenses Factor I Factor II Factor III Factor IV Variance, % Cronbach = Acting out 0.60 Help rejecting complaining 0.61 Projection 0.66 Projective identification 0.56 Regression 0.67 Somatization 0.50 Inhibition 0.55 Sublimation 0.59 Reaction formation 0.59 ( Splitting) ( 0.59) Suppression 0.53 Anticipation 0.52 Omnipotence 0.74 Isolation 0.61 ( Pseudoaltruism) ( 0.51) Affiliation 0.71 Anticipation 0.61 Scores below 0.50 are omitted 7 mechanisms, and 2 should be reformulated (No. 7 and 77). The content validity of DSQ could be criticized on 2 points: (a) how is it possible to self-evaluate partly unconscious mechanisms? and (b) definitions of defense mechanisms used for the DSQ were not completly explicit. The first point limits the measurement of defense mechanisms to conscious derivatives of defense only with the following consequences: firstly, DSQ will more easily rate defense mechanisms that lead to clear behavior patterns such as acting out and secondly DSQ can be used only in subjects who are aware of their own behavior. Minor differences in definitions of defense mechanisms would be important if DSQ were intended to measure individual defense mechanisms precisely. At the gross level of self-evaluation items, subtle differences in definitions do not seem relevant. Criterion Validity. We tested criterion validity of the DSQ with the following hypothesis: (a) the maladaptive style score should be higher with patients than with controls, and (b) maladaptive style score should correlate with clinical evaluation of defensive functioning level. Maladaptive style scores were found to be significantly higher in all patients (n = 140) than in the controls (n = 82) (t test mean difference = 1.16, d.f. = 220, t value = 6.45, p! ) (fig. 1). There was no significant difference of scores by sex, women having only slightly higher scores than men. Therefore, as the predominance of females in the control sample tended to lower the difference of scores between patients and controls, this difference cannot be due to a sex selection bias. Nevertheless, this comparison remains at a gross level and does not take in account differences in patients self-awareness or well-being. Indeed, it is interesting to compare maladaptive style scores with defensive functioning level (fig. 2). This analysis was performed on 59 patients. Firstly, patients at a level of defensive dysregulation (n = 9) show a strikingly low DSQ maladaptive style score: this reflects either the breakdown of defenses or a lack of self awareness that leads to underestimated self-evaluation of defense derivatives. When these patients at a level of defensive dysregulation are excluded, there is a significant correlation (n=50, Spearman rank correlation = 0.41, p = 0.004) between clinical evaluation of defensive level (DFS) and maladaptive style score. The scores are highest for patients at the action level and diminish to the mental inhibitions level. Patients at the disavowal level show lower than expected scores, corresponding to the predominant use of defense mechanisms such as denial. The correlation rate is similar to correlation rates with other external measures in previous studies, which are between 0.30 and 0.40 for maladaptive style only [15, 17, 18, 21]. These low rates indicate that other French Version of DSQ Psychother Psychosom 1998;67:

5 Fig. 1. Maladaptive style scores in patients (n = 140) and controls (n = 82). Fig. 2. Correlation between DSQ and clinical evaluation of defence functioning level (n = 59). 28 Psychother Psychosom 1998;67:24 30 Bonsack/Despland/Spagnoli

6 factors influence DSQ scores. Table 4 shows asignificant correlation between social functioning evaluation, the GAF score, and clinical evaluation of defensive functioning (DFS). This remains true even when patients at the level of defensive dysregulation are excluded, whereas the most significant difference in GAF scores lies between nonpsychotic (n = 50, mean = 60) and psychotic patients (n = 9, mean = 46) (mean difference = 14, d.f. = 58, t value = 2.8, p = 0.007). Conversely, DSQ scores do not correlate with GAF scores. To take self-awareness into account, we looked at nonpsychotic patients who show a discrepancy between self and clinical evaluation of defensive functioning. Although patients who underrate the DSQ score tend to have lower GAF scores, this difference is not significant. DSQ could be closer to symptoms whereas clinical defensive functioning level seems to be closer to social functioning evaluation. Construct Validity. The empirical hierarchy of defense mechanisms in defense styles corresponds to the prediction of defense theory, which is an element of construct validity. Reciprocally, hierarchical grouping of defense mechanisms in clinically relevant defense styles is an important contribution of the DSQ which confirms that defenses are not randomly distributed in the population. Replications of results show that grouping of defense derivatives beyond factor I is not perfectly stable, and has less solid psychometric features. However, defense derivatives do not cluster randomly, and defense styles remain clinically relevant and ranging from mature to an immature hierarchy. Conclusion Properties of the French version of DSQ are similar to the original version. Factor structure seems close enough to the original version to use the same defense styles and permit comparisons with other studies. As in other validation studies of the DSQ [12 14, 16, 17], we obtained a relatively stable factor I, but the other three were less stable. The first factor contains immature defense mechanisms associated with maladaptive behavior, such as acting out. Defense mechanisms of other factors differ slightly between studies, although grouping remains clinically relevant. The maladaptive style seems therefore to be the most clearly identified subscale within the DSQ. We propose using only the maladaptive style score (factor I) as a global defensive index: (a) it has the most desirable psychometric features, (b) it includes defense mechanisms that demonstrate a clear behavior pattern and are suscep- Table 4. Correlations between maladaptive style score (DSQ), defensive level (DFS) and GAF (n = 50) Spearman rank correlation p value DSQ, DFS DSQ, GAF DFS, GAF Patients at the level of defensive dysregulation are excluded. tible to self-awarenes, and (c) it correlates with the clinical evaluation of defensive functioning level. The low level (0.40) of this correlation suggests that both approaches, clinical and DSQ, reflect other aspects of defense mechanisms and that other elements such as self-awareness must be taken in account. In fact, patients who are aware of their own way of dealing with stressors could function better socially than those who are not. Psychometric features of defense styles other than the maladaptive style do not allow them to be used as a measure of more mature defense mechanisms. They are, however sufficiently enough relevant clinically to suggest for an empirical hierarchy of defense mechanisms. Patients at a level of defensive dysregulation have strikingly low maladaptive style scores. DSQ cannot therefore be used to evaluate patients with psychotic personality organization. Even if they do not suffer from a clear psychotic disorder, these patients tend to lose reality testing, lack self object differentiation and cannot use even immature defense mechanisms [22]. In conclusion, DSQ does not replace more subtle assessment methods of defense mechanisms, but does provide a simple, reliable and economical rating of maladaptive defense style in borderline and neurotic patients. Other factors that influence self-rating of defense style remain unknown. Self awareness probably plays an important role, and deserve further study [23]. Acknowledgements Many thanks to Jean Bovet for reviewing the manuscript, to Edmond Gilliéron for his support, to Christiane Schnyder, Pierre Cochand and Daniel Peter for their help in collecting data and to Martin Broome for reviewing the English text. We are also grateful to Yves de Roten and Michael Stigler for examining the face validity of the French version of DSQ. French Version of DSQ Psychother Psychosom 1998;67:

7 OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO References 1 Ehlers W, Hettinger R, Paar G: Operational diagnostic approaches in the assessment of defense mechanisms. Psychother Psychosom 1995;63: American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, ed 4 (DSM-IV). Washington, American Psychiatric Association, Parker JDA, Endler NS: Coping and defense; in Zeidner M, Endler NS (eds): Handbook of Coping. New York, Wiley, Laplanche J, Pontalis JB: Vocabulaire de la psychanalyse, ed 9. Paris, Presses Universitaires de France, Bond M: An empirical study of defensive styles: the Defense Style Questionnaire; in Vaillant GE (ed): Ego Mechanisms of Defense: a Guide for Clinicians and Researchers. Washington, American Psychiatric Press, 1992, pp Freud A: The Ego and the Mechanisms of Defense. London, Hogarth Press, Klein M: A contribution to the psychogenesis of manic depressive states. Int J Psychoanal 1935;16: Vaillant GE: Natural history of male psychological health: the relation of choice of ego mechanisms of defense to adult adjustment. Arch Gen Psychiatry 1976;33: Perry JC, Cooper SH: Empirical studies of psychological defenses; in Michels R, Cavenar JOJ (eds): Psychiatry. Philadelphia, Lippincott, 1987, vol Skodol AE, Perry JC: Should an axis for defense mechanisms be included in DSM-IV? Compr Psychiatry 1993;34(2): Ehlers W: The structure and process of defense in diagnosis of personality and psychoanalytic treatment; in Hentschel U, Smith GJW, Draguns JG (eds): The Concept of Defense Mechanisms in Contemporary Psychology: Theoretical, Research, and Clinical Perspectives. New York, Springer, 1993, pp Bond M, Gardner ST, Christian J, Siegel JJ: An empirically validated hierarchy of defense mechanisms. Arch Gen Psychiatry 1983;40: Andrews G, Pollock C, Stewart G: The determination of defense style by questionnaire. Arch Gen Psychiatry 1989;46: Flannery RB, Perry JC: Self-rated defense style, life stress, and health status: An empirical assessment. Psychosomatics 1990;31: Reister G, Fellhauer RF, Franz M, Wirth T, Schellberg D, Schepank H, Tress W: Psychometric measurements of ego defense mechanisms: Correlation between questionnaire and expert rating. Psychother Psychosom Med Psychol 1993;43(1): Sammallahti P, Aalberg V, Pentinsaari JP: Does defense style vary with severity of mental disorder? An empirical assessment. Acta Psychiatr Scand 1994;90: Schauenburg H, Schlüssler G, Leibing E: Empirische Erfassung von Abwehrmechanismen mit einem Selbsteinschätzungsfragebogen (nach Bond et al.). Psychother Psychosom Med Psychol 1991;41: Bond M, Perry JC, Gautier M, Goldenberg M, et al: Validating the self-report of defense styles. J Pers Disord 1989;3(2): Hentschel U, Ehlers W, Peter R: The measurment of defense mechanisms by self report questionnaires; in Hentschel U, Smith GJW, Ehlers W, Draguns JG (eds): The Concept of Defense Mechanisms in Contemporary Psychology: Theoretical, Research, and Clinical Perspectives. New York, Springer, 1993, pp Morley S, Snaith P: Principles of psychological assessment; in Freeman C, Tyrer P (eds): Research Methods in Psychiatry, ed 2. London, Gaskell, Vaillant GE, Bond M, Vaillant CO: An empirically validated hierarchy of defense mechanisms. Arch Gen Psychiatry 1986;43: Kernberg OF: Severe Personality Disorders: Psychotherapeutic Strategies, New Haven, Yale University Press, Ryff CD, Singer B: Psychological well-being. Psychother Psychosom 1996;65: Psychother Psychosom 1998;67:24 30 Bonsack/Despland/Spagnoli