[...]
PCL–R weighs antisocial behavior as strongly as—if not more strongly than—traits of emotional detachment in assessing psychopathy. Without a history of violent or criminal behavior, even an individual with pronounced interpersonal and affective traits of psychopathy is unlikely to surpass the PCL–R’s threshold score for diagnosing psychopathy.
[...]
The business success of “snakes in suits” (Babiak & Hare, 2006) contradicts the notion that classic criminal behavior is central to psychopathy
[...]
Hare (1996a) has long spoken of the psychopaths among us who infiltrate political, law enforcement, government, and other social structures: “Thanks to Hare, we now understand that the great majority of psychopaths are not violent criminals and never will be. Hundreds of thousands of psychopaths live and work and prey among us” (Hercz, 2001, ¶ 11). The two-factor model poorly identifies this “great majority of psychopaths” who escape contact with the legal system or simply express their psychopathic tendencies in a manner that does not conflict with the law.
[...]
This suggests that callous and unemotional traits—not necessarily antisocial behavior—are key features of psychopathy. Moreover, there is no compelling longitudinal evidence that psychopathy is highly stable from childhood to adulthood. Indeed, of adolescents with extremely high scores (upper 5%) on a PCL-derived measure of psychopathy during adolescence, less than one third were classified as psychopathic by a PCL measure in early adulthood (Lynam, Caspi, Moffitt, Loeber, & Stouthamer-Loeber, 2007). Measures that focus more exclusively on emotional detachment might yield greater stability estimates (see Vincent, 2002)
[...]
Recently, criminal behavior was deemed “the ultimate criterion for a measure of psychopathy” (Williams & Paulhus, 2004, p. 774). An iterative process of looping between theory and research is necessary to avoid further slippage toward the notion that psychopathy is merely a violent variant of antisocial personality disorder. Why has the field largely forgone this iterative process to embrace PCL–R measurement models as psychopathy?
Informed speculation about this issue is needed to identify potential barriers to future scientific progress. Of several potential explanations, three interacting factors that powerfully drive contemporary interest in the PCL–R and psychopathy emerge as most compelling to us.
First, the modern justice context has created a strong demand for identifying bad, dangerous people.
Second, the PCL measures, which include many malignant personality traits, have been shown to be useful in predicting violent and criminal behavior. In fact, the PCL–R has been lauded as an “unparalleled” single predictor of violence (Salekin et al., 1996, p. 211; cf. Cooke & Michie, 2006), and scholars have cautioned, “psychopathy [i.e., PCL–R scores] is such a robust and important risk factor for violence that failure to consider it may constitute professional negligence” (Hart, 1998, p. 133).
Third, this link between the PCL and violence has supported a myth that emotionally detached psychopaths callously use violence to achieve control over and exploit others. As far as the PCL is concerned, this notion rests on virtually no empirical support. The field’s pseudo-operationalism, then, may largely be a product of the fact that the PCL–R has proven useful in ways that make conceptual sense and fit modern societal demands.
The way in which the PCL–R is applied is consistent with this interpretation. First, although the PCL measures were developed to diagnose psychopathy, they overwhelmingly are applied as violence risk-assessment instruments. In a survey of 71 diplomates in forensic psychology, Tolman and Mullendore (2003) found that the psychological test used most often by diplomates to assess violence risk was the PCL–R. Indeed, diplomates used the PCL–R more than twice as often as well-validated risk-assessment instruments (Quinsey, Harris, Rice, & Cormier, 1998; Webster, Douglas, Eaves, & Hart, 1998).
Second, the PCL measures are used relatively often in legal proceedings, typically as a prosecution tool to meet the demands of sentencing (DeMatteo & Edens, 2006). To counter such strong practical forces toward pseudooperationalism, the process of understanding psychopathy must be separated from the enterprise of predicting violence. The field implicitly interprets the PCL measures’ association with violence as an indication that emotionally detached psychopaths use violence to prey upon others. This interpretation arguably explains the widespread use of the PCL measures.
A growing body of evidence suggests that the PCL measures’ predictive utility for violence cannot be attributed to their assessment of psychopathy per se (see Gendreau, Goggin, & Smith, 2002, 2003; Salekin et al., 1996; Serin, Peters, & Barbaree, 1990; Skeem, Miller, Mulvey, Tiemann, & Monahan, 2005; Skeem & 438 SKEEM AND COOKE Mulvey, 2001).
Instead, the lion’s share of the PCL’s predictive utility is attributable to its old Factor 2 assessment of antisocial behavior and traits that are not specific to psychopathic personality deviation (Walters, 2003b). Independent of an association with old Factor 2, the relation between old Factor 1 emotional detachment and future violence is often found to be insignificant (Harris, Rice, & Quinsey, 1993; M. M. Hicks, Rogers, & Cashel, 2000; Skeem & Mulvey, 2001; cf. Serin, 1996). Despite contentions that old Factor 1 is important too (see Hare & Neumann, 2005; Hemphill, Hare, & Wong, 1998; Vitacco et al., 2005), published studies suggest that the core interpersonal and affective features of psychopathy assessed by the PCL–R have yet to prove their independent value in predicting men’s future violence and criminality (for an interesting exception with women, see Richards, Casey, & Lucente, 2003).
Perhaps more importantly, measures that are designed to systematize ratings of chronic criminal behavior or a criminal lifestyle do as well as, or better than, the PCL–R in identifying persistently criminal offenders (Skilling, Harris, Rice, & Quinsey, 2002) and predicting violent and general recidivism (Cooke, Michie, & Ryan, 2001; Gendreau et al., 2002, 2003; Walters, 2003a).
The predictive utility of the PCL–R and these other measures of criminality seems to be based on two factors. First, indices of past criminal behavior (including violence) naturally are linked with future, like behavior. Recall that information about criminal behavior determines one’s ratings of some PCL–R items and heavily affects one’s ratings of others. Second, ratings of past criminal behavior appear to capture something traitlike that is clinically useful but not specific to psychopathy (Blackburn, 2007; Skeem et al., 2005).
[...]
Given that (a) the core features of psychopathy explain relatively little variance in future violent and other criminal behavior as a whole and (b) there are compelling indications that criminal behavior and social deviance are, at best, epiphenomena of psychopathy (see Parameters of the Debate, above), the field would do well to refine the PCL measures to focus more narrowly on the interpersonal and affective traits of the disorder. As Blackburn (2007) explained, identifying psychopathy with criminality confounds different conceptual domains because personality deviation and social deviance belong in different universes of discourse. Personality deviation is defined within the framework of interpersonal norms, whereas social deviance represents departures from legal or moral rules. To define the former in terms of the latter precludes any understanding of the relationship [between the two]. (Blackburn, 2007, p. 145, citations omitted)
As this quotation suggests, failure to separate the process of understanding psychopathy (which seeks construct identification) from the enterprise of predicting violence (which seeks clinical utility) will generate confusion. One might assume that criminal behavior is central to psychopathy because including such behavior
improves the PCL–R’s ability to predict violence. Indeed, some have argued that the PCL–R four-factor model has incremental utility over the three-factor model i
in “predicting important external correlates of psychopathy” (Neumann et al., 2007, p. 98). Those correlates are violence and aggression. Beyond past criminal behavior, adding such variables as gender, age, or substance abuse to the PCL–R might also improve prediction of violence. Such an improvement would not imply that these characteristics are central to psychopathy. Measures of psychopathy that do not emphasize criminal behavior will probably manifest limited utility in predicting violent and other criminal behavior (e.g., Salekin, Brannen, Zalot, Leistico, & Neumann, 2006; Skeem et al., 2003).
At the same time, the measures may better assess psychopathy, furthering psychologists’ ability to conduct powerful research to better understand its etiology, pathogenesis, and treatment. Clearly, there are measures for assessing risk of future violence at levels that rival or exceed those of the PCL measures (Douglas, Ogloff, Nicholls, & Grant, 1999; Gendreau et al., 2002, 2003; Skilling et al., 2002
Measures of psychopathy are meant to assess an enduring personality disorder marked by emotional detachment. Measures of risk are meant to inform risk management. Researchers should not confuse the two. Thus, we believe that the promise of reintroducing theory outweighs its peril.
