Issues in the diagnosis of sexual sadism
William L. Marshall1, Steven .J. Hucker2
1 Rockwood Psychological Services, Canada
2 University of Toronto
[Sexual Offender Treatment, Volume 1 (2006), Issue 2]
This paper summarizes our research to date on sexual sadism. Our initial review of the literature revealed confusion over diagnostic criteria. Our first empirical study showed that experienced forensic psychiatrists did not accurately employ many of the important diagnostic criteria while our second study demonstrated that internationally-renowned forensic psychiatrists could not reliably apply the diagnosis. On the basis of these observations we developed a Sexual Sadism Scale that we are now in the process of evaluating.
Key Words: Sexual sadism, diagnosis, reliability
Sexual sadists represent a real threat to the community in terms of their risk to reoffend but also in terms of the harm they will cause should they reoffend. Researchers and clinicians working with sexual offenders have yet to produce a combined index of risk that includes both the likelihood of reoffending and the likelihood of harm to a potential victim. In addition to this problem, the issue of whether or not a sexual offender meets criteria for sexual sadism has serious implications for decision makers (e.g., the courts, prison authorities, parole boards). Failure by clinicians to identify a true sadist might result in the offender’s release from custody when he is in fact a real threat to the community. On the other hand, diagnosizing a sexual offender as a sadist when he is not might result in continued or extended incarceration thereby jeopardizing the offender’s rights. It is clear from these considerations that the diagnosis of sexual sadist (or the finding that an offender is not a sadist) has serious implications for both the proper protection of the community and for the rights of identified offenders.
As a result of our concerns about these matters, we took the first step of reviewing the extent literature (Marshall & Kennedy, 2003). Unfortunately this review raised more concerns than it solved. We found that while most authors indicated they used the criteria specified by either the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorder (DSM), or the World Health Organization’s International Classification of Diseases (ICD), to diagnose their subjects, in fact the criteria they actually specified did not comply with either of these systems. Each researcher chose an idiosyncratic list of criteria which typically included some features from both DSM and ICD, but also included other features not mentioned in either of these texts.
From our review we identified at least 35 features that had been employed in one or another study as criteria for sexual sadism. The most common features related to the use of violence (including murder or mutilation), attempts to humiliate or degrade the victim, the exercise of power, control, dominance, or enslavement, ritualistic features associated with careful pre-planning of an attack, cruelty or torture, abduction of the victim and transport to a pre-selected location, bondage, prior history of cruelty to others or to animals, anal sex, and post-mortem sex or mutilation. Other features that were mentioned by some authors included keeping trophies from, or records of, the attack, crossdressing in the history of the offender, cannibalism, use of sadistic pornography, use of weapons in the attack and strangulation of the victim. The only common feature to all the reports was that sexual arousal to the identified criteria was seen as essential.
This notion that it is sexual arousal to various features of the attacks that is crucial to the diagnosis has a long history. Krafft-Ebing (1886) defined sadism as “the experience of sexual, pleasurable sensations … produced by acts of cruelty” (p. 109). Both DSM-IV-TR (American Psychiatric Association, 2000) and ICS-IV (World Health Organization, 1992) see sexual arousal to certain features as essential to the diagnosis. While this may seem reasonable in order to define any form of sexual deviance, it does present problems for diagnosticians since it is not clear how sexual arousal to the features is determined. In the absence of an admission by the offender, which in our experience is unlikely, the diagnostician must either infer sexual arousal from the information he/she has available or employ phallometry to detect such arousal. The degree to which inferences must be made, reduces the likely reliability of a diagnosis a fact that the authors of DSM-III (American Psychiatric Association, 1980) noted as their justification for moving away from a theory-based approach to the specification of more observable features to serve as diagnostic criteria. As yet no one has developed satisfactory specific stimuli for phallometric testing designed to detect sexual arousal to sadistic acts, although some have inferred sadistic tendencies from arousal to scenes of forced sex (Barbaree, Seto, Serin, Amos & Preston, 1994; Langevin et al., 1985; Seto & Kuban, 1996).
As a result of the confusion we noted in our literature review regarding the criteria used to diagnose sexual sadists, we decided that further research was required. Our first step was to determine how effectively the diagnosis was applied in federal prisons in Canada. We (Marshall, Kennedy & Yates, 2002) examined the records in three prisons of all sexual offenders for whom a psychiatric appraisal was made over the period 1989-1998. From these records we identified evaluations of 59 sexual offenders with 41 being diagnosed as sexual sadists while the remaining 18 were given other diagnoses. It is important to note that the clinicians, whose diagnoses we examined in this study, were all respected and experienced forensic psychiatrists. We then compared those diagnosed as sadists with those who were identified as nonsadists, on the features we derived from our literature review. We found that it was the nonsadists who displayed the so-called sadistic features. For example, 61.6% of those thought not to be sadists but only 24.4% of the sadists, violently beat their victims; similarly, 38.9% of the nonsadists and 9.8% of the sadists tortured their victims. On two composite indices of sadism it was again the nonsadists who appeared most problematic. On composite index derived from offense details, 100% of the nonsadists and 80.5% of the sadists scored in the deviant range; on a composite index based on phallometric responses only 17.1% of the sadists appeared deviant and yet 44.4% of the so-called nonsadists displayed deviant responses.
The results from this first study revealed that the diagnosis of sexual sadism was not being applied in the Canadian prison service in a way that matched any of the criteria identified in the literature. When we examined each diagnostician’s application of the criteria, it was evident that there was not only disagreement across diagnosticians in the criteria they considered relevant, there was no evident consistency within diagnosticians in the criteria they used. As a result we decide to see if a range of international experts might show greater consistency.
Our next study involved extracting, at random, twelve of the cases from our first study, and then producing documents that detailed every aspect of the offenses, the offenders’ life histories, their self-reported sexual fantasies and sexual interests (where available), psychometric test data, and phallometric assessment results. Six of these offenders had been identified as sadists in our first study and six had been given other diagnoses. We (Marshall, Kennedy, Yates & Serran, 2002) then sent these extensive documents to 24 internationally-renowned forensic psychiatrists and asked them to diagnose each case as a sadist or not a sadist. Fifteen psychiatrists returned the complete data. Diagnostic agreement across the psychiatrists proved to be unacceptably low. We employed the kappa coefficient to examine reliability across diagnosticians. This statistic corrects for chance agreement. Our analyses revealed a kappa of 0.14. For relatively important decisions, it is usually accepted that reliability across diagnosticians must reach a kappa of at least 0.9, whereas for unimportant decisions a kappa of at least 0.6 is required (Murphy & Davidshofer, 1998). Clearly the kappa we found was unacceptably low.
In this study we also asked the diagnosticians to rate all the criteria we identified in our literature review, in terms of their relevance for the diagnosis of sexual sadism. On the basis of the ratings provided by these experts, we developed a Sexual Sadism Scale. This scale has 17 items clustered in 4 groups. The first group of 5 items was judged by our experts to be essential to the diagnosis of sexual sadism, so in our scale these items are given the highest weightings. The items in each successive group are given progressively less weight, until the final grouping, which has only 2 items, is given the lowest weightings. We hoped to make the scale minimally dependent upon the diagnostician’s inference or on the offender’s self-report, and for the most part we were successful. Note that except for one item, the scale is indifferent regarding the issue of whether the offender is sexually aroused by the acts described in each item. Most of the items in the scale describe features that can be identified objectively from crime scene details or from detailed police and/or victim reports. One item (“offender is sexually aroused by sadistic acts”) depends on either the offender’s self-report or phallometric assessment results, although it could be inferred from the details of official reports of the offense.
For diagnosticians, our scale can serve to justify the diagnosis of sexual sadism or reveal features that warrant concern. Thus the scale serves both the needs of a categorical classification system, such as DSM or ICD, while at the same time employing the benefits of a dimensional system that has been touted by some as a better approach to diagnosis (Livesley, 2001; Widiger & Coker, 2003). Along with colleagues in several centres, we are in the process of subjecting our scale to empirical analyses with the first step being to establish the inter-rater reliability of the scale.
We hope that other researchers will either employ our scale or develop their own to pursue a more objective dimensional approach to identifying the problems presented by sexual sadists. We also encourage clinicians to use our scale and to provide us with feedback (table 1).
Offender is sexually aroused by sadistic acts
Offender exercises power/control/domination over victim
Offender humiliates or degrades the victim
Offender tortures victim or engages in acts of cruelty on victim
Offender mutilates sexual parts of victim’s body
Offender has history of choking consensual partners during sex
Offender engages in gratuitous violence toward victim
Offender has history of cruelty to other persons or animals
Offender gratuitously wounds victim
Offender attempts to, or succeeds in, strangling, choking, or otherwise asphyxiating victim
Offender keeps trophies (e.g., hair, underwear, ID) of victim
Offender keeps records (other than trophies) of offense
Offender carefully pre-plans offense
Offender mutilates nonsexual parts of victim’s body
Offender engages in bondage with consensual partners during sex
Victim is abducted or confined
Evidence of ritualism in offense
Table 1: Items of the “Sadism Scale”
- American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders (4th ed. Text rev.). Washington, DC: Author.
- Barbaree, H.E., Seto, M.C., Serin, R.C., Amos, N.L., & Preston, D.L. (1994). Comparisons between sexual and non-sexual rapist subtypes: Sexual arousal to rape, offense precursors and offense characteristics. Criminal Justice and Behavior, 21, 95-114.
- Krafft-Ebing, R. von (1996). Psychopathia sexualis. Philadelphia: F.A. Davis.
- Langevin, R., Bain, J., Ben-Aron, M.H., Coulthard, R., Day, D., Handy, L., Heasman, G., Hucker, S.J., Purins, J.E., Roper, V., Russon, A.E., Webster, C.D., & Wortzman, G. (1985). Sexual aggression: Constructing a predictive equation. A controlled pilot study. In R. Langevin (Ed.), Erotic preference, gender identify, and aggression in men: New research studies (pp. 39-76). Hillsdale, NJ: Lawrence Erlbaum.
- Livesley, W.J. (2001). Commentary on reconceptualizing personality disorder categories using trait dimensions. Journal of Personality, 69, 277-286.
- Marshall, W.L., & Kennedy. P. (2003). Sexual sadism in sexual offenders: An elusive diagnosis. Aggression and Violent Behavior: A Review Journal, 8, 1-22.
- Marshall, W.L., Kennedy, P., & Yates, P. (2002). Issues concerning the reliability and validity of the diagnosis of sexual sadism applied in prison settings. Sexual Abuse: A Journal of Research and Treatment, 14, 310-311.
- Marshall, W.L., Kennedy, P., Yates, P., & Serran, G.A. (2002). Diagnosing sexual sadism in sexual offenders; Reliability across diagnosticians. International Journal of Offender Therapy and Comparative Criminology, 46, 668-676.
- Murphy, K.R., & Davidshofer, C.O. (1998). Psychological testing: Principles and applications (4th ed.). Hillsdale, NJ: Prentice Hall.
- Seto, M.C., & Kuban, M. (1996). Criterion-related validity of a phallometric test for paraphilic rape and sadism. Behaviour Research and Therapy, 34, 175-183.
- Widiger, T.A., & Coker, L.A. (2003). Mental disorder as discrete clinical conditions: Dimensional versus categorical classification. In M. Hersen & S.M. Turner (Eds.), Adult psychopathology and diagnosis (4th ed., pp. 3-35). New York: John Wiley & Sons.
- World Health Organization (1992). The ICD-10 classification of mental and behavioral disorders. Geneva; WHO.
William L. Marshall
Rockwood Psychological Services
Suite 403, 303 Bagot Street, Kingston
Ontario, K7K 5W7 Canada