Prediction Based on Ominous Signs

Experience with sexual murderers, particularly compulsive murderers, reveals some common characteristics in their backgrounds and prehomicidal behaviors. Therefore, the presence of such signs should serve as a red flag or should raise what Mulvey and Lidz (1995) refer to as “clinical concerns” about the likelihood that such individuals will commit a sexually motivated compulsive murder in the future. Schlesinger (2001b) has described ten such ominous signs to be used as a guideline for practitioners confronting potential sex murderers. Although not yet empirically grounded, these signs can provide a direction for research with this relatively small criminal population. In fact, Douglas et al. (1999) have pointed out the heuristic benefit of this type of clinical finding; they call for a “collaboration among researchers and clinicians [where it is hoped their partnership will] benefit some patients, convicted persons, and even young children inclined towards early aggression and violence” (p. 149).

The usual signs and behaviors that the violence-prediction research has focused on—mental disorder, threats, prior violence—are not applicable to the potential compulsive sex murderer. Compulsive murderers rarely have a major psychiatric disorder like an overt psychosis (Meloy, 2000; Revitch, 1965; Revitch and Schlesinger, 1981, 1989; Schlesinger, 2000a). In fact, Meloy (2000) notes that “with few exceptions ... most sexual homicide perpetrators are not psychotic and do not have a diagnosable psychotic disorder if evaluated” (p. 7). In addition, as Rappeport (1967) has pointed out, most psychot- ics, contrary to popular belief, are not particularly prone to commit homicide and are rarely assaultive.

Unlike psychiatric patients who are involuntarily committed because they have made a threat to commit violence, compulsive murderers rarely threaten to kill. Unfortunately, the important relationship between threats to kill and carrying out such warnings has rarely been studied except for some early work by MacDonald (1963,1967,1968). In one of the few studies of its kind, MacDonald (1967) found that out of 100 patients admitted to a psychiatric hospital because of homicidal threats, four committed suicide, and only three committed homicide in a 5- to 6-year follow-up. Thus, there seems to be a connection between homicidal threats and suicide as well as homicide, a finding substantiated by Rhine and Mayerson (1973). In any case, relatively rarely does an individual who makes a homicidal threat carry out the act. In cases of compulsive murder, the problem is compounded since these offenders keep their fantasies private and typically do not directly notify others of their inner thoughts.

Moreover, many compulsive murderers have a history neither of interpersonal violence nor of the usual sex crimes and sexual offenses. If a community is terrorized by a series of sex murders or sex assaults, the police often target known sex offenders—most who are child molesters—who have been recently released. However, for many years researchers (Brancale, Ellis, and Doorbar, 1952; Brancale, McNeil, and Vuocolo, 1965; Guttmacher, 1963) have found that most sex offenders are passive, inadequate individuals who may repeat their specific crime but rarely commit a sexual murder. In fact, Revitch (1965) found that only 3 out of 43 (6%) violent sex offenders and sex murderers had a previous history of sexual crimes. Other than fetishism (which is generally not a crime) and voyeurism (which can be a crime but is often considered a misdemeanor), most of the sex murderers described in this book did not have a history of traditional sex offenses. In some cases individuals do commit sex crimes and escalate their acting out until it culminates in sexual murder. But these cases are not typical; most rapists, for example, do not go on to kill their victims (Groth, 1979; Podolsky, 1966; Ressler, Burgess, and Douglas, 1983).

The ten ominous signs that Schlesinger (2001b) described are listed in order of increasing prognostic significance (see Table 10.1). Early in the list are disturbed childhood experiences of abuse and inappropriate maternal (sexual) conduct. Childhood sexual abuse, which is quite prevalent in nonclinical, psychiatric, and other criminal populations, has a little discriminate capacity and therefore accrues little weight. As the list of behaviors advances from maternal sexual conduct to pathological lying to animal cruelty, more serious forms of acting out are involved. These escalating forms of disturbed behavior are typically found in the histories of compulsive murderers, and they accrue more significance accordingly. Firesetting, sexual burglary, unprovoked attacks on women, and ritualistic behavior are increasingly more ominous and therefore should be assigned the most prognostic weight.

In making a prediction, as in arriving at a diagnosis, specific signs, symptoms, behaviors, and test results are usually not sufficient. In the final analysis, clinical judgment must be exercised. As Douglas et al. (2013) have noted, the risk of future violence depends on numerous factors including types of supervision, treatment, motivation to live a noncriminal life and whether the individual will experience adverse life events (Otto, 2000). Malamuth (1986) found predictions of future sexual violence should be based on a dynamic interaction of various indicators, rather than on the static summation of them. Malamuth also argued that the assessments should be placed on a continuum of degrees of inclination to act aggressively. In fact, the statistical cookbook-like approach (Borum, 1996) may

Table 10.1 Ominous Signs (When Seen in Combination) Indicate Risk for a

Potential Sex Murderer

Source: From Schlesinger, L.B., Journal of Threat Assessment, 1,2001, p. 50. With permission.

miss the essence of the subject (Grubin, 1997). Psychometric measures used to help predict future sex crimes—standardized with traditional (nonhomicidal) sex offenders (Witt, Del Russo, Oppenheim, and Ferguson, 1996)—maybe of value, particularly when used in conjunction with the ten ominous signs. However, the clinician needs to remain aware that traditional sex offenders are very different from sexual murderers. Moreover, Litwack and Schlesinger (1999) argue:

The distinction between actuarial [cookbook] and clinical assessments is blurring to the point where there is often no meaningful distinction between the two. For example, good clinical assessment routinely takes into account such actuarial predictors as a history of violence; conversely, the most powerful actuarial predictor in recent studies—the PCL-R [Hare, 1991] score—is, in significant part, a clinical variable, (p. 188)

While the ten ominous signs described in detail in the following sections have not been empirically validated as a group, it is hard to believe that an individual who has engaged in this type of conduct is headed for a good adjustment in later life. Perhaps the person who commits sexual burglaries, sets fires, kills cats, attacks females in an unprovoked manner, and displays some ritualistic behavior may not commit a sexually motivated murder. However, these behaviors in and of themselves (notwithstanding the lack of empirical predictive validity) certainly should raise a major red flag and require some form of intervention. We hope that future efforts will be aimed at validating and weighing the significance of the ten ominous signs. In the meantime, however, clinical judgments and clinical concerns should be expressed without fear of making some false-positive predictions, which are to be expected under the circumstances. With this group of offenders, a false-negative prediction can be particularly devastating and should not be rationalized on the basis of lack of statistical verification.