Monique Tardif a *
and Hubert Van Gijseghem b

Respectivement, sexologue, Département de sexologie, UQAM, Institut Philippe Pinel
et psychologue, École de psycho-éducation, Université de Montréal

“Do pedophiles have
a weaker identity structure
compared with
nonsexual offenders ?”

In Child Abuse & Neglect, 25 (2001) 1381-1394. Pergamon.

Abstract

Introduction

Study objectives

Materials and methods

Subjects

Instruments

Hypotheses

Research procedure

Results

Abuse profile

Psychometric results

Discussion

Acknowledgements

References

Resumé / Resumen

ABSTRACT

Objective : The aim of this study was to determine whether pedophiles have a weaker identity structure compared with nonsexual offenders.

Method : The recruitment process secured the participation of 87 male adult subjects, divided into three groups : 27 pedophiles who abused male victims (X = 38.6 years), 30 pedophiles who abused female victims (X = 35.5 years), 30 nonsexual offenders (X = 29.8 years). The concept of identity was examined on the basis of two factors : body image limits and ego identity. Two objective tests, the Minnesota Multiphasic Personality Inventory (MMPI) and the Ego Identity Scale, and one projective test, the Rorschach scored according to the Fisher and Cleveland method, were used.

Results : A multivariate analysis of covariance, at a significance threshold of p < .05, indicated that pedophiles who abused female victims and pedophiles who abused male victims have more fragile body image limits, as measured by the penetration score, and present a higher level of social introversion (Si scale) than do nonsexual offenders. Moreover, pedophiles who abused male victims have a weaker ego (Es scale) than do pedophiles who abused female victims and nonsexual offenders. In other respects, no significant intergroup difference emerged in terms of body image limits as measured by the barrier score and of strength of ego identity (Ego Identity Scale).

Conclusions : Conceptual and empirical elements related to body image and ego identity are addressed to shed light on the potential disturbances in the identity of pedophile subjects. These results imply not only that certain impairments exist at the level of the basis structures of the personality, but also that these impairments should be taken into account in formulating a program and devising a therapeutic process for pedophiles. © 2001 Elsevier Science Ltd. All rights reserved.

Keywords : Pedophilia ; Identity ; MMPI ; Rorschach ; Social introversion

[1382]

INTRODUCTION

The development of the ego depends on a series of identifications and maturational stages that lay the groundwork for a stable identity and adapted psychosocial functioning (Erikson, 1968 ; Bilsker & Marcia, 1991 ; Rasmussen, 1964). It has been ventured that individuals with a poorly structured ego are more prone to relational conflicts, thereby compromising their social adaptation (Erikson, 1968 ; Bilsker & Marcia, 1991). These statements concur with the major social maladjustment generally observed in pedophiles, which suggests that ontogenic factors may contribute to the emergence of pedophilic tendencies. The aim of this study was to determine whether the identity structure of pedophiles is less solid than that of nonsexual offenders in tenus of body image and ego identity.

Body image

The first form of the ego is the skin ego. It is constructed from body functions and sensations, primarily those experienced through the skin as the membrane that marks the boundary between the individual and the outside world (Anzieu, 1974 ; 1994). The anaclitic link between body parts and the symbolic meaning placed on them depends on the individ-ual's interactions with others (Fisher & Cleveland, 1968). During the mother-child separation process when the need is felt to conceive the body as a protective barrier in the face of outside dangers, the corporal limits serve to ensure ego-integrity and reflect how an individual with a more or less stable identity perceives himself in the external world (Anzieu, 1974 ; Mahler, Pine, & Bergman, 1975).

Body image and the projective method

The unstructured nature of the Rorschach inkblots allows for representations of the body image, which differentiates the ego from the nonego, and the self from the environment. Making this distinction is intimately connected to the development of the ego (Anzieu, 1974 ; Boizou, Chabert, & Rausch de Traubenberg, 1978 ; Fisher & Cleveland, 1968 ; Sanglade, 1983). Moreover, according to Fisher (1986), whether an individual has feelings of vulnerability in the face of the outside world depends on the unconscious notion of the penetrability of the body. The Fisher and Cleveland rating system covers two determinants, namely, the barrier score, which refers to well-defined limits and protective functions, and the penetration score, which is associated with broken limits, the function of penetration, and the deterioration of surfaces and entities.

Body image and delinquency

Research has shown that persons who develop a propensity for acting out or for social deviance present weaker body image limits than do individuals who are better socially adapted (Fisher, 1986). A study by Megargee (1965) indicated that nonsexual offenders obtained significantly lower barrier scores (p < .001) than did the nondelinquent normative groups in Holtzman, Thorpe, Swartz, and Herron (1961). Furthermore, an intragroup comparison [1383] revealed that barrier scores were significantly lower (p < .001) among delinquents who committed more aggressive offenses. Generally speaking, however, empirical research that has employed the Fisher method (Fisher, 1986) does not lend itself readily to interstudy comparisons, owing to differences in samples, testing conditions and methodology. Nonetheless, these studies have demonstrated the link between delinquency and poorly defined body image limits and, consequently, a poorly integrated ego.

Ego identity

A subjective sense of identity (mind and body) grows out of a familiar and relational environment and must lead contemporaneously to preservation of the self as a unique and coherent entity and to stable relational changes with respect to others (Graafsma, 1994 ; Marcia, 1994). Fantasies, too, play a major role in the emergence of identity. In other words, identity is not based solely on the concrete reality of individuals, as unconscious dynamics are also involved (Graafsma, 1994).

Relationship to the self and identity development

Sexual offenders are often diagnosed with a personality disorder which suggests, among other things, rigid and poorly adapted functioning and serious identity problems (Kernberg, 1984 ; Proulx, Aubut, Perron, & McKibben, 1994). Where pedophiles are concerned, the integration and reciprocal developmental functions of ego identity appear to be wanting. In its role as a conciliator, the ego can make use of functions that must sustain individuality and self-esteem while striving to resolve conflicts. The propensity of delinquents to act out denotes an inadequate control of urges by the ego.

Throughout the individuation process, the creation of an integrative and global sense of identity appears to depend on a series of temporary, partial and altered identifications. According to Erikson (1968), the acquisition of a more structured identity is finalized in adolescence with the integration of gender, physical maturation, formal thought and the requirements related to social functioning. A diffused identity is common among apathetic, slothful, and socially isolated individuals who attempt to avoid ail contact with others. Those driven by a frenetic search for contact with others are often in a constant seducer mode (Marcia, 1994). Generally speaking, these individuals have difficulty sustaining stressful situations. They are more likely to conform to external demands, and their relationships may prove shallow or superficial (Marcia, Waterman, Matteson, Archer, & Orlofsky, 1993). Moreover, they have low self-esteem and a rather low level of moral development. Ail of these characteristics are often found in pedophiles. Their acting out of sexual urges and fantasies reflects the regressive nature of their ego.

Ego identity and delinquency

Ego identity plays a major role in the evaluation and handling of threatening, anxiogenic, and conflictual situations, aimed at adopting appropriate social behavior. In this connection, adolescents in the general population have been shown to have a stronger ego identity [1384] (significance level of p < .05) than do delinquent adolescents, as measured via Rasmussen's Ego Identity Scale (Protinsky, 1988). Along similar lines, Seghorn and Boucher (1980) reported that most of the 50 sexual offenders in their study had difficulties of adaptation, as well as a lack of mastery over their lives and a lack of control over aggressive and sexual urges, which seemed to be related to relational problems, impulsiveness, and a low frustration threshold.

Relationship with others and identity development

A study by Meeus and Dekovic (1995) demonstrated that identity in adolescence, as measured by the Utrecht-Groningen Identity Scale (U-GIDS), is mostly influenced by peers. Parents apparently have only a secondary positive influence. According to these researchers, the relational identity in a sample of subjects 12 years to 24 years old increased with age. This process of relational restructuring is centered on notions of symmetry and equality and apparently begins with frequenting a group of peers and, subsequently, finding a partner.

Samet and Kelly (1987), for their part, demonstrated that relationships with the opposite sex diminished the risk of identity confusion. In this regard, the culmination of identity development remains the acquisition of basic social skills, the affirmation of a sexual identity, and the ability to experience sexual intimacy. This attests to the integrative function of the ego in response to the needs of the id. However, peer pressure to frequent the opposite sex can prove stressful for individuals who are unable to meet this challenge.

Numerous articles have dealt with the deficits that pedophiles present regarding social adaptation and the establishment of interpersonal relationships, compared with the general population (Stermac & Quinsey, 1985 ; Overholser & Beck, 1986). Drawing on attachment theory, Marshall, Hudson, and Hodkinson (1993) suggested that sexual delinquents, due to an unsteady and nonconfident type of attachment to their parents, fail to develop adequate peer relationships, a stage that would guarantee the development of intimate relationships in adulthood. It is generally recognized that these individuals have difficulty maintaining, or even initiating, relationships with other adults, which at times takes on the form of extreme isolation (Marshall, Earls, Segal, & Darke, 1983 ; Quinsey, 1986). Often the tendency to idealize children and their world is accompanied by a negative perception of adults, which explains the avoidance attitude of sexual delinquents (Howells, 1979 ; Segal & Stermac, 1990).

A study by Langevin, Wright, and Handy (1989) of different types of sexual delinquents who were subdivided into those who were sexually abused and those who were not, revealed that the former tended to score higher on Social introversion (Si) on the MMPI (t = 1.76, p < .10). These sexually abused subjects also experienced more aggressiveness in their contacts with others and had more sexual contacts with peers. Shoor, Speed, and Bartelt (1966), for their part, reported that adolescent pedophiles were socially isolated, had enormous difficulty relating to others in a significant fashion, did not belong to a group of peers, and exhibited poor social functioning at best. These factors denote not only deficits in social skills but also the presence of aggressive and sexual behaviors likely to elicit rejection.

More specifically, Blaske, Borduin, Henggeler, and Mann (1989) demonstrated that, compared with nonsexual delinquents and nondelinquents, adolescent sexual delinquents [1385] manifested impaired functioning at the emotional level with respect to peer relationships. It has been reported that about 70% of pedophiles consider that they did not have a close friend during childhood or adolescence (Saunders, Awad, & White, 1986 ; Tingle, Barnard, Rob-bins, Newman, & Hutchinson, 1986). This deprived them of a relationship appropriate to their age and capacity for intimacy (Deisher, Wenet, Paperny, Clark, & Fehrenbach, 1982 ; Overholser & Beck, 1986 ; Segal & Marshall, 1985). The study by Van Ness (1984) revealed that adolescent sexual offenders had more difficulty reaching a level of social competence than did other types of delinquents or adolescent normals. Generally, sexual offenders have deficits in social skills, but this clinical observation does not specify the type and nature of the developmental factors that apply to pedophiles who abused female victims and pedophiles who abused male victims.

STUDY OBJECTIVES

The aim of this study was to determine the following : (1) whether pedophiles have a weaker ego-integrity with respect to body image and ego identity, compared with nonpedophiles ; and (2) whether pedophiles who abused male victims (PM) have a weaker ego-integrity with respect to body image and ego identity, compared with pedophiles who abused female victims (PF).

Materials and methods

Subjects

The subjects were recruited from three sources : outpatient services of the Pinel Institute of Montreal, the Probation Program of Montreal, and Amorce, a self-help group for pedophiles. The study involved a sample of 82 subjects divided into the following three groups : 27 pedophiles who abused male victims (PM), 26 pedophiles who abused female victims (PF), and 29 nonsexual offenders (NS). Two subjects whose victims were aged over 16 years and one offender who did not complete the tests were replaced. The protocols for four PF subjects and one NS subject were withdrawn after the computerized correction of the MMPI suggested a validity problem. Ail the sexual and nonsexual offenders admitted their offenses. This increased the likelihood that their responses to the tests would be reliable.

The criteria for designating a subject a pedophile were partially drawn from the DSM-IV (1994). Thus, a pedophile subject must have admitted that, over a period of at least 6 months, he entertained sexual fantasies involving children or adolescents aged 13 years or under. Also, the sexual abuser had to be at least 5 years older than his victim. Furthermore, other criteria were added : fantasies must have been associated with an actual sexual abuse of children or adolescents aged 16 years or under. To define more precisely pedophile subjects, specific criteria were applied to exclude subjects who had committed incestuous acts only, isolated pedophilic acts without associated fantasies, other types of paraphilias or acts involving both boys and girls.

[1386]

Nonsexual offenders had committed property crimes and drugs related offenses only and were under probation. These subjects admitted having an active delinquency problem over a period of at least 6 months. Nonsexual offenders were chosen as a control group for their socio-demographic similarities with pedophiles.

Certain inclusion criteria applied to the sample as a whole : over 18 years of age, French-Canadian origin, and at least high-school education. Certain exclusion criteria also applied to all subjects to control for potential bias. These included presence of psychiatric antecedents and current or past recourse to therapy for more than 3 months.

Instruments

Rorschach with Fischer and Cleveland rating method. The Rorschach allows subjects to express their perception of the limits of their body image, be it to describe its solidity and its enveloping or protective function or its fragility. Through their research, Fisher and Cleveland (1968) identified two determinants for rating responses to the Rorschach, namely, the barrier score and the penetration score. The barrier score relates to responses with a reference to a protective surface, which could symbolize perception of body image limits. This includes references to anything that covers or hides something else (e.g., clothing), to particular features of a surface (e.g., spotted skin), and to protective, overhanging or enveloping surfaces (e.g., vase). The penetration score relates to responses that could be the symbolic expression of a subjective feeling that the body provides poor protection and can easily be penetrated. This notion comes through in three types of perceived images : damaged body surface (e.g., injured animal), ways and means of penetrating the inside of a structure or expulsion of something from inside another thing (e.g., window, erupting volcano), and the surface of something perceived as permeable, fragile, deteriorated, and inconsistent (e.g., penetration of limits, cancer, ghost).

Thus, an individual with a well-constructed ego-integrity will more often visualize better-defined limits, which translates into a high barrier score. A high penetration score, instead, is indicative of weak ego limits and of more fragile body image boundaries than if there were no barrier score. Empirical studies have reported the absence of correlation between the barrier and penetration scores. The final score for a protocol is obtained from the total number of barrier responses and the total number of penetration responses. Certain responses can be rated as both a barrier and a penetration response (e.g., broken vase).

In several studies, the method for rating the concepts of barrier and penetration demonstrated good inter-rater agreement, as evidenced by correlation coefficients of .80 to .90 for each factor. With the exception of one study, test-retest reliability ranged from .65 to .80. To avoid introducing a bias in the data, the barrier and the penetration scores were weighted according to the total number of responses given per protocol.

Ego Identity Scale (EIS). The Ego Identity Scale (Tan, Rendis, Fine, & Porac, 1977) comprises 12 pairs of forced-choice statements that measure the structure of the ego identity based on Erikson's conception (Erikson, 1968). In each pair, one statement relates to ego-integrity and the other to a diffused identity. The authors sought to reduce the social desirability effect by adopting the forced-choice technique and pairing off statements with a [1387] similar level of social desirability. The construct validity rests on a significant correlation between the EIS and four personality variables : Rotter's Locus of Control Scale (.26), Tan et al.'s Intimacy versus Isolation Scale (.25), Tomkin's Polarity Scale (.26), and Schulze's Short Dogmatism Scale (-.20). The reliability coefficient obtained with the even-odd method was .68.

Minnesota Multiphasic Personality Inventory (MMPI). The French-language version of the MMPI translated by Chevrier (1961) was used to verify the scores obtained on the Social introversion (Si) and the Ego strength (Es) scales.

The 68 items of the Es scale measure a general adaptation and personal integration factor. They relate to physical functioning, isolation, religious attitudes, moral stance, personal adequacy and ability to overcome difficulties, phobias, and anxieties. Barron (1953) demonstrated that the homogeneity coefficient of the even/odd items of this scale was .76 and that the test-retest reliability after 3 months was .72. Moreover, Greene (1991) reported no overlap between the items of the Es and Si scales. A high score on the Es scale indicates that the subject is more easily able to overcome stressful situations and that his ego is strong. A low score often attests to problems of the personality.

The Si scale comprises 70 statements that measure features associated with introversion and extroversion in the areas of thought, social activity, and emotional expression. Introversion is found in individuals who avoid social responsibilities and interpersonal relation-ships and who exhibit little interest in others. Extroversion, instead, characterizes individuals who manifest interest in social life and people. A high score on the Si scale indicates a tendency towards social introversion.

HYPOTHESES

Compared with nonsexual offenders, pedophiles will obtain significantly higher scores with respect to the psychic determinants of a fragile identity (penetration score, Si scale) and significantly lower scores with respect to indicators of a solid identity, be it in terms of body image limits (barrier score) and ego identity (Es scale, EIS)

Compared with pedophiles who abused female victims (PF), pedophiles who abused male victims (PM) will obtain significantly higher scores with respect to the psychic determinants of a fragile identity (penetration score, Si scale) and significantly lower scores with respect to the indicators of a solid identity, be it in terms of body image limits (barrier score) and ego identity (Es scale, EIS).

Research procedure

Data collection took place over a 16 month period. Participation was voluntary and independent of any benefit or reward. Testing was conducted on an individual basis by either of the two interviewers from the Centre de psychiatrie légale in Montreal, at the probation offices of the Montreal region or at the Département de Psychologie de l'Université de Montréal. The data from the various psychometric instruments were collected on anonymous [1388] response forms. The testing sequence was the same for ail subjects : MMPI, Rorschach, Clarke Parent-Child Relations, Bem Sex-Role Inventory, and EIS. The interviewer took down socio-demographic information before the start of testing. The collection of data related to offenses and family history took place at the end of testing to avoid the possibility that subjects adopt a more defensive attitude towards the tests. The Rorschach and EIS protocols were corrected manually by the two interviewers. The MMPI protocols were corrected and processed by computer. This project was approved by the human investigation committee namely, le Comité d'Éthique à la recherche de l'Institut Philippe Pinel de Montréal.

Table 1
Data on physical and sexual abuse suffered
by group and age at 1st sexual abuse

* p < .01,

** p < .05,

*** p < .00015.

RESULTS

Abuse profile

The recruitment process secured the participation of 87 subjects, divided into three groups : 27 pedophiles who abused male victims (PM), 30 pedophiles who abused female victims (PF) and 30 nonsexual offenders (NS). The socio-demographic profiles of the three groups did not differ significantly with respect to the following : years of schooling, annual income, sources of income, marriage status, and drug abuse. About one-third of the pedophiles in the sample abused both prepubescent and pubescent victims, whereas two-thirds targeted pre-pubescent victims exclusively. The mean age of PM, PF, and NS was 38.6 years, 35.5 years, and 29.8 years, respectively. However, the age difference between the group of nonsexual offenders and the group of pedophiles who abused male victims proved significant (p < -05), owing to a sampling bias. This forced us to run analyses of covariance (ANCOVA) to attenuate the impact of age on the test scores.

Self-reported information regarding family history was not independently verified. Different profiles of childhood abuse emerged from these data (Table 1). A significantly larger proportion of PM were physically abused as children, compared with NS (55.5% vs. 20.0%, p < .01). Although one-third (33.3%) of PF were physically abused as children, the difference with the other two groups did not prove significant. A significantly higher proportion of PM (74%) was also sexually abused, compared with NS (20%). Once again, the group of PF victims stood midway between the other two groups, and differences were [1389] not significant. The age at which sexual abuse occurred was significantly lower for PF (7.82 years), compared with PM (9.42 years at p < .05), and NS (13.2 years).

Table 2
Means and SD for the three groups on SI, ES, Penetration, Barrier, and EIS

Psychometric results

Two psychologists conducted blind analyses of ail the Rorschach protocols. The intraclass correlation to determine inter-rater agreement was .98 for the barrier score and .93 for the penetration score.

To limit the effect on type-I errors (alpha) of running repeated comparison tests, we resorted to multivariate analysis of covariance (MANCOVA) to examine simultaneously the intergroup differences on ail the variables related to ego-integrity. Age of subjects was introduced into the analysis model as a covariable, to attenuate the effect of the heterogeneity of the groups on this dimension. The MANCOVA results revealed a significant difference on three facets of ego-integrity : Es scale (MMPI), Si scale (MMPI), and Rorschach penetration score (F (10, 148) = 2.51, p = .008). Table 2 presents the univariate interpretation indices for this analysis, as well as means and standard deviations. These univariate indices showed that PF and NS presented a better integration of ego strength, compared with PM (F = 4.05, p — .02). Results on the Si scale indicated that PM and PF obtained a higher social introversion score than did NS (F = 6.11, p = .003). Pedophiles (PM and PF) manifested a tendency to perceive body image limits that were more easily penetrated, compared with NS (F = 2.66, p = .08). No intergroup difference emerged regarding results on the EIS (F = 1.43, p = .25) and the barrier scores (F = 97, p = .38). It should be noted that the EIS did not present a satisfactory homogeneity coefficient, which inevitably raises questions about its usefulness. Finally, subjects in the three groups did not differ regarding perception of their body as a protective and clearly delineated entity, as measured by the barrier score.

DISCUSSION

The results of this study suggest that PM and PF, compared with NS, develop a body image that they feel is more easily penetrated. This corroborates our first hypothesis to the effect that a significant difference would emerge between pedophiles and nonsexual offenders on the penetration score. Our second hypothesis predicting a significant difference in [1390] terms of penetration scores between PM and PF was not confirmed. It should be pointed out that the penetration score is better than the barrier score at discriminating between pedophiles and nonsexual offenders. The fact that PM and PF obtained a significantly higher penetration score than did NS could mean that their feeling of vulnerability towards the outside world is more disturbed, that they perceive the outside world to be more intrusive, and that the limits of their ego are fragile. Yet the dimension of body image limits as evaluated by the barrier score did not present any significant differences across the three groups. An unexpected finding was that of a stable and separate identity for ail subjects regardless of group, which suggests that PM and PF are no more disturbed than NS with respect to this dimension.

Where the evaluation of the concept of ego identity is concerned, the analysis of the results indicates that the ego strength (Es) of PF and NS is greater than that of PM. The hypothesis to the effect that pedophiles would have presented a significant difference on the Es scale compared with NS was not confirmed. Pedophiles (PM) manifested a weaker ego than did PF, thereby confirming the significant difference expected between PF and PM. This result suggests that PM have greater difficulty coping with situations perceived as threatening and conflictual, compared with the two other groups. It also reflects a weaker ego and suggests that a higher level of childhood disturbance might have affected ego development. The data related to the abuse profile of the subjects in this study indicates that the problem of sexual and relational abuse was more pronounced among PM than among PF.

On the Si scale, both pedophile groups (PM and PF) obtained significantly higher scores for social introversion than did NS. The hypothesis to the effect that PM would be more socially introverted than PF is not confirmed. These results support the majority of clinical observations and the few empirical studies that reported impairment at the level of the interpersonal relationships, social functioning, and intimate relationships of sexual offenders (Deisher et al., 1982 ; Overholser & Beck, 1986). In this regard, pedophile subjects (PM and PF) present a less solid ego identity compared with NS on various aspects of their relationships with others.

Certain clinical indicators lead us to believe that pedophiles (PM and PF) present more pronounced and specific ego-integrity impairments than do NS. The relative positions obtained by the three groups regarding the different variables were in line with the predictions made in the study hypotheses. Pedophiles who abused female victims maintained an intermediate position between the two other groups with respect to the three variables that yielded significant differences. However, the nondifferentiation of PF from PM (penetration and Si scores), on the one hand, and of PF from NS (Es), on the other, raises the hypothesis that the tests used in this study are not sufficiently sensitive. It is also possible, however, that no significant difference exists in terms of these variables, and that the results obtained are a proper reflection of the characteristics of the three groups in question.

In conclusion, the analysis of our results indicates that PM and PF develop a more fragile and vulnerable body image compared with NS and that they present more problems in terms of social introversion. Moreover, NS and PF possess greater Ego strength than do PM. These results imply not only that certain impairments exist at the level of the basic structures of the personality, but also that these impairments should be taken into account in formulating a prognosis and devising the therapeutic process for [1391] pedophiles. Moreover, the early identification of these impairments of the identity structure can guide the intervention with adolescents and children who manifest sexually abusive behavior to prevent the development of abusive sexuality and to promote a more healthy developmental process.

These results have suggested that ego impairments and body image have become rooted during infancy and the adolescence of future child molesters. Moreover, in accordance with Ryan (2000b), the identification process implies not only concrete reality but also the psychic development of the concept of self, its relationship with others, and its influence on sexual development and deviance. By targeting these precocious impairments, it is important to understand the role of parenting on a development of sexually abusive behaviors of children on younger children. The interest in this question is mentioned in articles which treat the cycle of aggression and intergenerational transmission of violence (Bentovim, 1996 ; Ryan, 1998 ; 1999) ; poor attachment bonds between parents and children (Ward, Hudson, & Marshall, 1996) ; and negative perceptions of parental figures among pedophiles (Tardif & Van Gijseghem, 1999). Therefore it is important to work with parents of juvenile sexual offenders to understand the dynamic interplay among individual, family, and other relationships, to identify the weaknesses and vulnerabilities of certain categories of subjects and to improve the competency factors in parenting (Ryan, 1998, 1999, 2000a, b).

ACKNOWLEDGMENTS

The authors express their gratitude to Margaret C. Kiely, Ph.D., professor in the Department of Psychology, University of Montreal, for her support and her valuable and constructive comments.

REFERENCES

American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC : Author.

Anzieu, D. (1974). Le moi-peau. Nouvelle Revue de Psychanalyse, 9, 195-208.

Anzieu, D. (1994). Le penser du Moi-peau au Moi-pensant. Paris : Dunod.

Barron, F. (1953). An ego strength scale which predicts response to psychotherapy. Journal of Consulting Psychology, 42, 152-162.

Bentovim, A. (1996). The trauma organized System of working with family violence. In C. Cordess & M. Cox(Eds.), Forensic psychotherapy : crime. Psychodynamics and the offender patient (Vol. 2, pp. 291-311). London : Jessica Kingsley.

Bilsker, D., & Marcia, J. E. (1991). Adaptative regression and ego identity. Journal of Adolescence, 14, 75-84.

Blaske, D., M., Borduin, C. M., Henggeler, S. W., & Mann, B. J. (1989). Individual and peer characteristics of adolescent sex offenders and assaultive offenders. Developmental Psychology, 25 (5), 846-855.

Boizou, M. F., Chabert, C, & Rausch de Traubenberg, B. (1978). Représentation de Soi, identité, identification au Rorschach chez l'enfant et l'adulte. Bulletin de Psychologie, 32 (3, Suppl. 7), 271-277.

Chevrier, J. M. (1961). Inventaire multiphasique de la personnalité (Minnesota). Montréal, Qc : Institut de Recherche Psychologique (trad. de Minnesota Multiphasic Personality Inventory, 1940).

[1392]

Deisher, R. W., Wenet, G. A., Paperny, D. M., Clark, T. F., & Fehrenbach, P. A. (1982). Adolescent sexual offense behavior : the role of the physician. Journal of Adolescent Health Care, 2 (4), 279-286.

Erikson, E. H. (1968). Identity, youth and crisis. New York : Norton.

Fisher, S. (1986). Development and structure of the body image (Vol. 2). Hillsdale, NJ : Erlbaum.

Fisher, S., & Cleveland, S. E. (1968). Body image and personality (2nd ed. rev.). New York : Dover Publications.

Graafsma, T. L. G. (1994). A psychoanalytic perspective on the concept of identity. In H. A. Bosma, T. L. G. Graafsma, H. D. Grotevant, & D. J. de Levita (Eds.), Identity and development : an interdisciplinary approach(pp. 41-61). Thousand Oaks, CA : Sage Focus Edition.

Greene, R.L. (1991). The MMP1-2/MMPI : an interpretative manual. Boston, MA : Allyn and Bacon.

Holtzman, W. H., Thorpe, J. S., Swartz, J. D., & Herron, E. W. (1961). Inkblot perception and personality. Austin, TX : University Texas Press.

Howells, K. (1979). Some meanings of children for pedophiles. In M. Cook & G. Wilson (Eds.), Love and attraction (pp. 57-82). New York : Pergamon Press.

Kernberg, O. F. (1984). Severe personality disorders. New Haven, CT : Yale.

Langevin, R., Wright, P., & Handy, L. (1989). Characteristics of sex offenders who were sexually victimized as children. Annals of Sex Research, 2, 227-253.

Mahler, M. S., Pine, F., & Bergman, A. (1975). The psychological birth of the human infant. New York : Basic Books.

Marcia, J. E. (1994). The empirical study of ego identity. In H. A. Bosma, T. L. G. Graafsma, H. D. Grotevant, & D. J. de Levita (Eds.), Identity and development : an interdisciplinary approach (pp. 67-80). Thousand Oaks, CA : Sage Focus Edition.

Marcia, J. E., Waterman, A. S., Matteson, D. M., Archer, S. L., & Orlofsky, J. (1993). Ego identity : a handbook for psychosocial research. New York : Springer Verlag.

Marshall, W. L., Earls, C. M., Segal, Z., & Darke, J. (1983). A behavioral program for the assessment and treatment of sexual aggressors. In K. D. Craig & R. D. McMahon (Eds.), Advances in clinical behavior therapy (pp. 148-174). New York : Brunner Mazel.

Marshall, W. L., Hudson, S. M., & Hodkinson, S. (1993). The importance of attachment bonds in the development of juvenile sex offending. In H. E. Barbaree, W. L., Marshall, & S. M. Hudson (Eds.), The juvenile sex offender (pp. 164-181). New York : Guilford Press.

Megargee, E. I. (1965). Relation between barrier scores and aggressive behavior. Journal of Abnormal Psychology, 70(4), 307-311.

Meeus, W., & Dekovic, M. (1995). Identity development, parental and peer support in adolescence : results of a national Dutch survey. Adolescence, 30 (120), 931-944.

Overholser, C, & Beck, S. (1986). Multimethod assessment of rapists, child molesters, and three control groups on behavioral and psychological measures. Journal of Consulting and Clinical Psychology, 54, 682-687.

Protinsky, H. (1988). Identity formation : a comparison of problem and nonproblem adolescents. Adolescence, 23(89), 67-72.

Proulx, L, Aubut, L, Perron, L., & McKibben, A. (1994). Troubles de la personnalité et viol : implications théoriques et cliniques. Criminologie, 27, 33-53.

Quinsey, V. L. (1986). Men who have sex with children. In D. Weisstub (Ed.), Law and mental health : international perspectives (Vol. 1, pp. 84-121). New York : Pergamon Press.

Rasmussen, J. E. (1964). Relationship of ego identity to psychosocial effectiveness. Psychological Reports, 15, 815-825.

Ryan, G. (1998). Trends in a national sample of sexually abusive youths. Annual Progress in Child Psychiatry and Child Development, IV, 333-344.

Ryan, G. (1999). Treatment of sexually abusive youth : the evolving consensus. Journal of Interpersonal Violence, 14 (4), 422-436.

Ryan, G. (2000a). Childhood sexuality : a decade of study. Part I—research and curriculum development. Child Abuse & Neglect, 24 (1), 33-48.

Ryan, G. (2000b). Childhood sexuality : a decade of study. Part II—dissemination and future directions. Child Abuse & Neglect, 24 (1), 49-61.

[1393]

Samet, N., & Kelly, E. W. (1987). The relationship of steady dating to self-esteem and sex role identity among adolescents. Adolescence, 22 (85), 231-245.

Sanglade, A. (1983). Image du corps et image de soi au Rorschach. Psychologie Française, 28 (2), 104-111.

Saunders, E., Awad, G. A., & White, G. (1986). Male adolescent sexual offenders : the offender and the offense. Canadian Journal of Psychiatry, 31 (6), 542-549.

Segal, Z. V., & Marshall, W. L. (1985). Heterosexual social skills in a population of rapists and child molesters. Journal of Consulting and Clinical Psychology, 53, 55-63.

Seghom, T. K., & Boucher, R. J. (1980). Sexual abuse in childhood as a factor in adult sexually dangerous criminal offenses. In J. M. Samson (Ed.), Childhood and sexuality : proceedings of the International Symposium (Pt. VII, 705-711). Montreal, Qc : Éditions Études Vivantes.

Shoor, M., Speed, M. H., & Bartelt, C. (1966). Syndrome of the adolescent child molester. American Journal of Psychiatry, 122 (7), 783-789.

Stermac, L. E., & Quinsey, V. L. (1985). Social competence among rapists. Behavioral Assessment, 18, 171-185.

Tan, A. L., Kendis, R. J., Fine, J. T., & Porac, J. (1977). A short measure of Eriksonian Ego identity. Journal of Personality Assessment, 41, 279-284.

Tardif, M., & Van Gijseghem, H. (1999). La perception des figures parentales des pedophiles hétérosexuels et homosexuels : realité factuelle ou virtuelle. Bulletin de Psychologie, 52 (5), 597-604.

Tingle, D., Barnard, G. W., Robbins, L., Newman, G., & Hutchinson, D. (1986). Childhood and adolescent characteristics of pedophiles and rapists. International Journal of Law and Psychiatry, 9 (1), 103-116.

Van Ness, S. R. (1984). Rape as instrumental violence : a study of youth offenders. Special issue : gender issues, sex offenses and criminal justice : current trends. Journal of Offender Counseling Services & Rehabilitation, 9 (2), 161-170.

Ward, T., Hudson, S. M., & Marshall, W. L. (1996). Attachment style in sex offenders : a preliminary study. Journal of Sex Research, 33 (1), 17-26.

RÉSUMÉ

Objectif : Le but de cette étude fut de determiner si les pédophiles ont une identité plus faible que les non agresseurs.

Méthode : On a recruté 87 hommes pédophiles, répartis en trois groupes : 27 pédophiles ayant agressé des victimes de sexe masculin (X = 38.6 ans), 30 pédophiles abusant des victimes de sexe feminin (X = 35.5 ans) et 30 délinquants non sexuels (X — 29.8 ans). On a examiné leur concept de l'identité à partir de deux aspects : les limites de l'image corporelle et l'identité du moi. On s'est servi de deux tests objectifs, le MMPI et le Ego Identity Scale ; et un test projectif, le Rorschach. Ce dernier a été côté selon la méthode Fisher et Cleveland.

Résultats : Une analyse de covariance dont le seuil critique est de p < .05 indique que les limites de l'image corporelle sont plus fragiles chez les pédophiles hétérosexuels et homosexuels, d'après le score de pénétration. Ils présentent des nivaux d'introversion sociale (échelle Si) plus élèves que les agresseurs non sexuels. De plus, les pédophiles homosexuels ont un moi plus faible (échelle Es) que les pédophiles hétérosexuels et les agresseurs non sexuels. Aucune autre différence importante entre les groupes au point de vue des limites de l'image corporelle et de l'identité du moi n'a été notée.

Conclusions : Des éléments conceptuels et empiriques vis-à-vis l'image corporelle et l'identite du moi sont presentes afin d'identifier des troubles de l'identité communs aux pédophiles. Les résultats de l'étude montrent qu'il existe des faiblesses dans la structure fondamentale de leur personnalité et il est proposé qu'elles soient prises en considération lorsque l'on conçoit un programme et une thérapie pour les pédophiles.

[1394]

RESUMEN

Objetivo : El objetivo de este estudio fue determinar si los pedófilos tienen una estructura de identidad mas debil comparados con los no ofensores sexuales.

Metodo : El proceso de reclutamiento aseguró la participación de 87 sujetos hombres, adultos, divididos en tres grupos : 27 pedófilos quienes abusaron víctimas varones (X — 38.6 anos), 30 pedofilos que abusaron a victimas hembras (X = 35.5 anos), 30 no ofensores sexuales (X = 29.8 anos). El concepto de identidad fue examinado sobre la base de dos factores : limites de la imagen corporal e identidad del yo. Se utilizaron dos pruebas objetivas, el IMPM y la Escala de Identidad del Yo, y una prueba proyectiva, el Rorschach, corregida de acuerdo al método de Fisher y Cleveland.

Resultados : Un análisis multivariado de covarianza, a un umbral de significócion de p < .05, indicó que los heterosexuales y pedófilos que abusaban de las víctimas varones tienen su imagen corporal más frágil, de acuerdo a la medida de penetración, y presentan un nivel mas alto de introversión social (Escala SI) que los no ofensores sexuales. Además, los pedófilos que abusaron de víctimas masculinas tienen un Yo más debil (Escala ES) que los pedofilos que abusaron de víctimas femeninas y los no ofensores sexuales. En otros aspectos, no surgieron otras diferencias significativas intergrupales en terminos de los límites de la imagen corporal medida por el puntaje de barrera y por la fuerza de la identidad de Yo (Escala de Identidad del Yo).

Conclusiones : Se discuten elementos conceptuales y empíricos relacionados con la imagen corporal y la identidad del Yo para esclarecer los potenciales trastornos en la identidad de los sujetos pedófilos. Estos resultados implican no solo que existen ciertos daños a nivel de las estructuras básicas de la personalidad, sino también que estos trastornos deben tomarse en cuenta en la formulación de un programa y la creación de un proceso terapeutico para los pedófilos.