Sexual Sadism: Closing in on an Elusive Diagnosis

Sexual Sadism: Closing in on an Elusive Diagnosis
Julia E. McLawsen, M.A.

University of Nebraska, Lincoln

Rebecca L. Jackson, Ph.D.

Pacific Graduate School of Psychology & Washington State Special Commitment Center

The diagnosis of sexual sadism is awash in a sea of controversy. Research examining the reliability and validity of this diagnosis has yielded discrepant results regarding 1) characteristics and behaviors associated with sexual sadism, as well as 2) how mental health professionals understand and apply this diagnosis. Contradictory research findings raise questions not only about how mental health professionals understand and apply this diagnosis, but also about whether they should. In this article, we review recent empirical literature that has offered mixed evidence for the existence of sexual sadism as a valid diagnostic construct. Although further research must be done before we can rest assured that sexual sadism is a meaningful category, we contend that it is premature to abandon this diagnosis. We conclude this article with recommendations for practitioners.
Overview of Sexual Sadism
Sexual sadism is a paraphilia that is chiefly characterized by deriving sexual pleasure from inflicting pain or humiliation on others. Although sexual sadism shares features with BDSM (an emblematic acronym derived from the terms “bondage and discipline,” “dominance and submission,” and “sadism and masochism”), BDSM is distinct from sexual sadism as it refers to a subculture in which sexual conduct occurs with the explicit consent of all involved partners. Sexual sadism is rare within the general population, and it is difficult to determine how common it is among sex offenders, with prevalence estimates ranging from 2%-5% (Quinsey, Chaplin, & Varney, 1981) to 50% (Langevin, 2003). This surprising range is likely due variability across study samples, as well as diverse methods of defining and diagnosing sexual sadism, two issues that we will examine more closely in this article.
What Makes a Diagnosis Meaningful?
“Diagnosis” originates from the Greek term for through/between (dia) and knowledge (gnosis). In the domain of clinical psychology, diagnosis refers to a process through which a mental health professional discerns specific information about a patient, and uses this information to determine which diagnostic category (or categories) most accurately characterizes the patient’s constellation of signs and symptoms. According to the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR), the resulting diagnosis fulfills two functions (American Psychiatric Association [APA], 2000). First, it represents a label that communicates relatively standardized information about a patient. Second, it represents “the first step in a comprehensive evaluation” (APA, 2000, p. xxxiv) whereby mental health professionals determine what interventions are likely to ameliorate the symptoms that warranted the diagnosis in the first place. A diagnosis that fulfills these criteria makes it possible for mental health professionals to share information with one another, and contribute to a body of knowledge that is used to develop effective treatment and management strategies. On the other hand, a diagnosis that fails to fulfill either of these functions clearly offers nothing more than an empty construct with little utility. Clearly, the utility of the diagnostic process – not to mention the resulting diagnosis – depends on how reliably mental health professionals make diagnostic decisions.
With regard to sexual sadism, Marshall and Kennedy (2003) have written extensively about the difficulty of operationalizing the DSM criteria for this “elusive diagnosis” (p. 15). Their article characterizes the diagnostic criteria as vague, and contends that one of the primary challenges in making this diagnosis stems from the diagnostician’s need to rely on inference when determining whether a patient inflicts suffering to fulfill a sexual motive. No one but the patient knows the answer to this question (and, in some cases, it may be the case that not even the patient understands his own motives), and there are clear reasons why a patient might not want to admit that he derives sexual arousal from inflicting suffering on others.
Empirical Literature Offers Mixed Support for Sexual Sadism
Although relatively few studies have examined the validity of sexual sadism, their varied results create considerable controversy. In this section, we examine two contentious issues that emerge from the sexual sadism literature. First, do characteristics associated with sadism actually differentiate between sadists and non-sadists? And second, do mental health professionals understand and apply this diagnosis consistently?
Issue #1: Do Characteristics Associated with Sexual Sadism Differentiate Between Sadistic and Non-sadistic Offenders?
There is disagreement about whether attributes associated with a diagnosis of sexual sadism actually differentiate between individuals with and without this diagnosis. Some researchers contend that personal characteristics and offense behaviors (Dietz, Hazelwood, & Warren, 1990; Gratzer & Bradford, 1995) associated with sexually sadistic offenders fail to differentiate sex offenders with and without a sexual sadism diagnosis (Marshall, Kennedy, & Yates, 2002). Contrary to what theory would predict, Marshall, Kennedy, and colleagues found that non-sadists sex offenders’ offense behaviors, self-reported fantasies, and phallometric arousal appeared more deviant than sadistic sex offenders’.
Other researchers, however, have identified a handful of characteristics and offense behaviors that appear to reliably differentiate between offenders with and without a sexual sadism diagnosis (Jackson, Richards, McCraw, & Koenen, 2006, March; McLawsen, Jackson, Vannoy, Gagliardi, & Scalora, 2008). Jackson and colleagues drew upon historical and contemporary sexual sadism literature to devise a list of characteristics and offense behaviors. Using a sample of 78 civilly committed sexually violent predators, half of whom were diagnosed with sexual sadism, they found that six of the listed offense behaviors reliably related to a diagnosis of sexual sadism, and all of these behaviors had to do with violence inflicted before or during sexual climax. Based on these results, Jackson and colleagues proposed that violence may be a necessary precursor to sexual arousal in sadistic sex offenders. This “arousal hypothesis” suggests violent acts committed before or during sexual climax may enhance arousal, and therefore may be relevant to diagnosing sexual sadism. On the other hand, it is unlikely that violent acts committed after sexual climax would be sexually motivated; in which case, such acts would not be relevant to diagnosing sexual sadism. Given the conventional understanding of sexual sadism as a disorder in which one derives sexual pleasure from inflicting pain or humiliation on others, the arousal hypothesis has clear appeal. However, McLawsen and colleagues (2008) found that professionals did not rate behaviors associated with the arousal hypothesis as more indicative of sexual sadism – that is, professionals do not appear to place differential values on pre- and post-climax behaviors when diagnosing sexual sadism.
McLawsen and colleagues (2008) investigated how professionals who had experience working with sex offenders understood the relative importance of behaviors associated with sexual sadism as opposed to behaviors associated with general (i.e., non-sadistic) sexual offending. This study identified four conceptualizations of sexual sadism: 1) behaviors found to occur “frequently” during sadistic offenses; 2) DSM diagnostic criteria; 3) a composite of behaviors indicative of control/domination/power, cruelty/torture, sexual mutilation, and humiliation/degradation (based on Marshall, Kennedy, Yates, and Serran’s (2002) finding that experts tended to consider these themes as particularly relevant to diagnosing sexual sadism); and, 4) Jackson and colleagues’ “arousal hypothesis.” Using a seven-point Likert-type scale, professionals rated 62 behaviors according to their “essentialness” for a diagnosis of sexual sadism. Professionals rated behaviors associated with the first three conceptualizations as significantly more essential to making a diagnosis of sexual sadism. In addition, professionals gave higher ratings to behaviors associated with “sadistic” sexual offenses, and lower ratings to behaviors associated with non-sadistic sexual offenses. These results suggest that there may be some characteristic and offense behaviors and that can be used to differentiate between sadistic and non-sadistic sexual offenses.
Issue #2: Do Mental Health Professionals Understand and Apply This Diagnosis Consistently?
Second, several studies have cast doubt on how diagnoses of sexual sadism are rendered. Marshall, Kennedy, Yates, and Serran (2002) found that diagnosticians relied on idiosyncratic methods to diagnose sexual sadism. In addition, several studies have revealed that mental health professionals diagnose sexual sadism with low levels of agreement (Levenson, 2004; Levenson & Morin, 2006; Packard & Levenson, 2006). Marshall, Kennedy, Yates, et al. (2002) invited 24 psychiatrists, “deemed to be expert in forensic diagnoses” (p. 669), to rate twelve vignettes according to their fit for a diagnosis of sexual sadism. The vignettes described 12 sexual offenders who, in an earlier study (Marshall, Kennedy, et al., 2002), had been diagnosed as either a sadist or non-sadist. Diagnostic agreement resulted in a kappa of 0.14, or an agreement rate of 21.7%, well below what the authors consider an acceptable level of agreement (90%; see also, American Educational Research Foundation, 1999; McDermott, 1988; Murphy & Davidshofer, 1998).
Levenson (2004) addressed the main limitation of Marshall, Kennedy, Yates, et al.’s study by examining the agreement of psychiatric diagnoses between two or more independent psychological evaluations of Florida inmates, and found a kappa for sexual sadism of 0.30. Packard and Levenson (2006) re-analyzed Levenson’s (2004) data using raw proportions of agreement, odds and risk ratios, and conditional probability estimates as alternative methods to examine diagnostic agreement, and reported odds ratios indicating that two diagnosticians were 67 times more likely to agree with each other about a sexual sadism diagnosis than disagree. Relative risk ratios demonstrated a 53-fold increased likelihood that a second evaluator would arrive at the same diagnostic conclusion as the first diagnostician. However, likely due to its extremely low base rate (less than 15%) within the sample, Packard and Levenson found low positive predictive power for sexual sadism. Specifically, when a subject received a diagnosis of sexual sadism, there was a one in five probability that a second diagnostician would also diagnose that same subject with sexual sadism.
In contrast to the four studies described above, McLawsen and colleagues (2008) demonstrated that professionals who work with sex offenders in a variety of capacities reliably discriminate between sadistic and non-sadistic sexual offense behaviors. While participants’ areas of expertise (e.g., therapy, assessment, and legal) did not affect rating patterns, participants’ diagnostic experience yielded a surprising effect (McLawsen, Vannoy, Jackson, Gagliardi, & Scalora, 2008, October). Participants with less diagnostic experience tended to ascribe higher levels of importance to sadistic behaviors. Considering these behaviors as representative of the diagnostic criteria for sexual sadism, this finding suggests that professionals with less experience tended to evaluate behaviors in closer accordance with diagnostic criteria. These results suggest that professionals may share a common understanding of what behaviors indicate sexually sadistic offenses.
Making Sense of These Equivocal Results
Although preliminary evidence suggests that certain behaviors may differentiate between sadistic and non-sadistic sexual offenses, further research is needed to demonstrate whether these findings hold true across multiple samples. And although it appears that professionals may share an understanding of what behaviors indicate sadistic versus non-sadistic sexual offenses, it remains uncertain whether such a shared understanding translates into diagnostic reliability. Whereas McLawsen and colleagues (2008) had participants rate behaviors on a continuous Likert-type scale, making a diagnosis requires dichotomous (yes/no) decision-making. Compared to dichotomous decision-making tasks, Likert-type scales yield increased reliability (Finn, 1972; Lissitz & Green, 1975; McMordie, 1979). McLawsen and colleagues might have found lower levels of reliability if they had asked participants to rate behaviors using dichotomous response options, rather than seven-point scales.
In addition, it is important to note that that McLawsen and colleagues examined how professionals ascribe importance to behaviors associated with sexual sadism. This is related to, but different from, examining diagnostic consistency. Thus, it is reasonable to say that McLawsen and colleagues’ findings call into question – but do not contradict – studies suggesting that professionals diagnose sexual sadism with low levels of agreement.
In conclusion, it appears that we can respond to the two identified issues by acknowledging that 1) there is preliminary evidence to suggest that certain behaviors may differentiate between sadistic and non-sadistic sexual offenses, and 2) whereas mental health professionals agree, at least under some conditions, on what characteristics and behaviors are indicative of sadistic sexual offenses, it may be the case that agreement at the behavioral level does not necessarily translate into diagnostic consistency. Presently, we do not yet have sufficient evidence to reach firm conclusions about the construct validity of sexual sadism. This means that professionals should proceed with caution when it comes to making this diagnosis and/or incorporating it into decisions.
Recommendations for Practitioners
Although the extant body of literature paints an inconclusive picture of sexual sadism, the available information can still be useful for guiding diagnostic practices and risk assessment.
As Marshall and Hucker (2006a) point out, poor diagnostic reliability does not imply the absence of a meaningful construct. For example, if there are, indeed, certain behaviors that indicate sexual sadism, then poor diagnostic reliability might reflect deficient diagnostic practices (e.g., the need to rely on inferences of what motivates a patient’s behavior), rather than a useless diagnostic category. To this effect, Marshall and Hucker propose a behaviorally-based checklist that could improve diagnostic decision-making. Citing Marshall, Kennedy, Yates et al. (2002), Marshall and Hucker note that even though expert diagnosticians made diagnoses of sexual sadism with poor interrater reliability, the experts generally agreed on what features were important to diagnosing sexual sadism.
Based on findings that professionals reliably discriminated between sadistic and non-sadistic offense behaviors, McLawsen and colleagues (2008) arrived at a similar conclusion to Marshall and Hucker (2006a). A diagnostic aide that encourages professionals to make dichotomous decisions by considering behaviorally-based criteria along a dimensional scale would promote a more systematic approach to diagnostic decision-making. Such an instrument might help professionals translate agreement at the behavioral level into diagnostic consistency. While the field waits for such a checklist, professionals who evaluate the suitability of a diagnosis of sexual sadism may wish to consider the relevance of violence committed for reasons other than gaining the victim’s compliance. These behaviors would provide theory-based support for a sexual sadism diagnosis.
It is unclear how sexual sadism affects an offender’s likelihood of sexual reoffense. One study found that sexual sadism might represent a risk factor. among offenders older than 40 years, those diagnosed with sexual sadism showed less of a decline in sexual reoffending rates when compared to non-sadistic rapists and pedophiles (Dickey, Nussbaum, Chevolleau, & Davidson, 2002). However, Marshall and Hucker (2006b) note that recidivism literature fails to support the commonly-held assumption that sexual sadists pose a high risk of reoffense. Yates, Hucker, and Kingston (2008) illustrate Marshall and Hucker’s point by citing a study by Langevin (2006), which found that only 32% of sexual sadists appeared high-risk on either the Violence Risk Appraisal Guide (VRAG; Quinsey, Harris, Rice, & Cormier, 2006) or Sex Offender Risk Appraisal Guide (SORAG; Quinsey et al., 2006). However, Yates and colleagues mention Doren’s (2002) cautionary advice about the limitations of using actuarial instruments with groups on which the instruments have not been validated. The low prevalence rates of sexual sadism, in addition to the questionable diagnostic practices that some studies have used to identify groups of sexual sadists, raise questions about whether it is appropriate to use actuarial risk assessment instruments with this group.
Until extant risk assessment instruments are validated with sexual sadists, or appropriate instruments are developed for use with this group, then professionals should exercise caution when using actuarial risk assessment instruments to predict recidivism risk posed by sexual sadists. For example, an evaluator could acknowledge the potential limitations of using extant risk assessment procedures with sexual sadists by first explaining that risk assessment instruments are calibrated on group estimates, and then identifying unique patient characteristics that indicate why a patient might not be represented by the instrument’s validation sample. An evaluator could also advise that, because of unavoidable limitations to the risk assessment procedures, the risk assessment results do not necessarily provide a valid estimate of the patient’s risk for recidivism.
* * *
In the six years that have passed since Marshall and Kennedy (2003) pronounced sexual sadism “elusive,” research has gradually begun to shed light on this construct. While the emerging picture remains relatively obscure, we are now at a point where we can piece together the available information and build a rudimentary understanding of what factors discriminate between sadistic and non-sadistic sexual behavior. Eventually, it will be possible to apply this knowledge of relevant behavioral criteria and devise tools that will facilitate reliable and valid diagnostic decisions.
References

American Educational Research Foundation. (1999). Standards for educational and psychological testing. Washington, DC: American Psychological Association.

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders, fourth edition, text revision. Washington, DC: Author.

Dickey, R., Nussbaum, D., Chevolleau, K., & Davidson, H. (2002). Age as a differential characteristic of rapists, pedophiles, and sexual sadists. Journal of Sex & Marital Therapy, 28(3), 211-218.

Dietz, P. E., Hazelwood, R. R., Warren, J. (1990). The sexually sadistic criminal and his offenses. Bulletin of the American Academy of Psychiatry and the Law, 18(2), 163-178.

Doren, D. M. (2002). Evaluating sex offenders: A manual for civil commitments and beyond. Thousand Oaks, CA: Sage.

Finn, R. H. (1972). Effects of some variations in ratingscale characteristics on the means and reliabilities of ratings. Educational and Psychological Measurement, 32, 255-265.

Gratzer, T., & Bradford, J. M. W. (1995). Offender and offense characteristics of sexual sadists: A comparative study. Journal of Forensic Sciences, 40(3), 450-455.

Jackson, R. L., Richards, H. J., McCraw, J. E., & Koenen, M. (2006, March). Sexual sadism: In word and deed. Paper presented at “American Psychology-Law Society,” St. Petersburg, FL, U.S.A.

Levenson, J. S. (2004). Reliability of sexually violent predator civil commitment criteria. Law & Human Behavior, 28, 357-368.

Levenson, J. S., & Morin, J. W. (2006). Factors predicting selection of sexually violent predators for civil commitment. International Journal of Offender Therapy and Comparative Criminology, 50(6), 609-629.

Lissitz, R. W., & Green, S. B. (1975). Effect of the number of scale points on reliability: A Monte Carlo approach. Journal of Applied Psychology, 60, 10-13.

MacCulloch, M., Snowden, p., Wood, P., & Mills, H. (1983). Sadistic fantasy, sadistic behavior, and offending. British Journal of Psychiatry, 143, 20-29.
Marshall, W. L., & Hucker, S. J. (2006a). Severe sexual sadism: its features and treatment. In R. D. McAnulty and M. M. Burnette (Eds.), Sex and sexuality: Sexual deviation and sexual offenses (Vol. 3, pp. 227-250). Westport, CT: Praeger Perspectives.
Marshall, W. L., & Hucker, S. J. (2006b). Issues in the diagnosis of sexual sadism. Sexual Offender Treatment, 1(2). Available online: http://www.sexual-offender-treatment.org
Marshall, W. L., & Kennedy, P. (2003). Sexual sadism in sexual offenders: An elusive diagnosis. Aggression and Violent Behavior, 8(1), 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: Journal of Research & Treatment, 14(4), 301-311.
Marshall, W. L., Kennedy, P., Yates, P., & Serran, G. (2002). Diagnosing sexual sadism in sexual offenders: Reliability across diagnosticians. International Journal of Offender Therapy and Comparative Criminology, 46(6), 668-677.
McDermott, P. A. (1988). Agreement among diagnosticians or observers: Its importance and determination. Professional School Psychology, 3(4), 225-240.
McLawsen, J. E., Jackson, R. L., Vannoy, S. D., Gagliardi, G. J., & Scalora, M. S. (2008). Professional perspectives on sexual sadism. Sexual Abuse: A Journal of Research and Treatment, 20, 272-304.
McLawsen, J. E., Vannoy, S. D., Jackson, R. L., Gagliardi, G., & Scalora, M. J. (2008, October). Professional perspectives on sexual sadism: The role of diagnostic experience. Poster presented at “Association for the Treatment of Sexual Abusers,” Atlanta, GA, USA.
McMordie, W. R. (1979). Improving measurement of death anxiety. Psychological Reports, 44, 975-980.

Murphy, K. R., & Davidshofer, C. O. (1998). Psychological testing principles and applications (4^th ed.). Hillsdale, NJ: Prentice Hall.

Packard, R. L., & Levenson, J. S. (2006). Revisiting the reliability of diagnostic decisions in sex offender civil commitment. Sexual Offender Treatment, 1(3). Available online: http://www.sexual-offender-treatment.org

Quinsey, V. L., Chaplin,T. C., & Varney, G. (1981). A comparison of rapists’ and non-sex offenders’ sexual preference for mutually consenting sex, rape, and physical abuse of women. Behavioral Assessment, 3, 127-135.

Quinsey, V. L., Harris, G. T., Rice, M. E., & Cormier, C. A. (2006). Violent offenders: Appraising and managing risk (2^nd ed.). Washington, DC: American Psychological Association.
Yates, P. M., Hucker, S. J., & Kingston, D. A. (2008). Sexual sadism: Psychopathology and theory. In D. R. Laws, & W. T. O’Donohue (Eds.), Sexual deviance: Theory, assessment, and treatment (2^nd ed.). New York: Guilford.