Investigating Characteristics of the Nonrecidivating Psychopathic Offender

Abstract

Psychopathic offenders are at higher risk to violently reoffend than nonpsychopathic offenders; however, about one in four psychopathic offenders are not reconvicted for a violent offense even over extended follow-ups. The characteristics of nonrecidivating psychopathic offenders (NRPs) remain underexamined. In a sample of 123 offenders with a minimum Hare Psychopathy Checklist–Revised (PCL-R) score of 25, community follow-up of 5 years, and 4 months in a violence reduction program, 65 men receiving a violent conviction within 5 years (recidivating psychopathic offenders [RPs]) were compared with 58 who had not (NRPs). Comparatively, NRPs were older at release, had better community support, and were rated lower violence risk. NRPs also had significantly lower PCL-R Factor 2 scores, but higher Factor 1 scores. Post hoc analyses revealed prominent NRP characteristics included exploitative personality traits and a predilection for instrumentally violent crimes. Findings are discussed in terms of their implications for the risk assessment and treatment of psychopathic offenders.

Keywords

nonrecidivating psychopathic offender PCL-R treatment recidivism psychopathy

Since the mid-1990s, there has been a host of research demonstrating that psychopathic offenders are high volume violent offenders with lengthy criminal careers. Indeed, offenders scoring highly on the tool most commonly used to assess psychopathy, the Hare Psychopathy Checklist–Revised (PCL-R; Hare, 1991, 2003) and its derivatives, commit more crimes than the average criminal offender (Hare & McPherson, 1984; Wong, 1984); have higher rates of violent recidivism (Leistico, Salekin, DeCoster, & Rogers, 2008); have higher rates of institutional misconduct, including violence (Guy, Edens, Anthony, & Douglas, 2005); and are more likely to fail on conditional release (Hart, Kropp, & Hare, 1988). Not only did the PCL-R usher in a new “benchmark” for risk assessment (particularly for violent recidivism), the construct of psychopathy also introduced a theoretical framework to assist in understanding the habitual offender; that is, individuals who possess psychopathic personality traits (e.g., lack of empathy, grandiosity, manipulative nature) are more likely to possess the respective behavioral characteristics (e.g., impulsivity, need for stimulation, parasitic lifestyle) and to engage in a more active criminal lifestyle.

A well-replicated oblique two-factor solution to the PCL-R has demonstrated that the personality and behavioral aspects of psychopathy are distinct yet related. PCL-R factor scores can be grouped into four domains, with Factor 1 being disaggregated into Interpersonal (e.g., superficiality, grandiosity) and Affective (e.g., callous/lack of empathy, shallow affect) facets, and Factor 2, Lifestyle (e.g., impulsivity, irresponsibility) and Antisocial (e.g., poor behavioral controls, criminal versatility) facets. Psychometric research provides particularly strong support for the predictive accuracy of Factor 2 in general (Leistico et al., 2008), and the incremental validity of the Antisocial facet in particular, for a range of recidivism outcomes (Walters, Knight, Grann, & Dahle, 2008). While Factor 1 and the Interpersonal and Affective facets have tended to be less consistent predictors of recidivism, research has demonstrated Factor 1, particularly the Affective facet, to be important responsivity indicators. Specifically, high Affective facet scores have been linked to increased treatment noncompletion (Olver & Wong, 2011) and decreased therapeutic progress (Olver, Lewis, & Wong, 2013).

While the PCL-R has contributed major inroads for forensic assessment, there are important caveats associated with its use. For instance, some researchers have found evidence for inconsistent field reliability of the PCL scales (as well as other structured forensic tools) within forensic evaluation contexts, particularly Factor 1 items (Edens, Boccaccini, & Johnson, 2010; Murrie et al., 2009). Cox, Clark, Edens, Smith, and Magyar (2013) also found that mock jurors were more likely to support a death verdict when they perceived the individual to have pronounced interpersonal and affective characteristics of psychopathy. In short, notwithstanding the major inroads contributed by the psychopathy construct and the PCL-R to improving risk assessment and clinical practice, such research highlights the need for professional discretion and care when employing the tool in forensic evaluation contexts. This is particularly salient given the connotations of the label “psychopath” and the heterogeneity that exists even among the group of individuals with this classification.

Psychopathy and Nonrecidivism

Despite being a group broadly classified as high risk, empirically and theoretically it is also apparent that not every psychopathic offender is destined to violently reoffend. Although the majority of psychopaths do recidivate within a matter of months or a few years after release, a consistent minority of psychopaths (20%-30%) do not reoffend (at least as verified by official criminal records), even after long periods of follow-up time (Hare, 1996; Hart et al., 1988; Quinsey, Rice, & Harris, 1995; Serin & Amos, 1995; Wintrup, Coles, Hart, & Webster, 1994). Consistently, research demonstrates that about one in four psychopathic offenders released from custody do not receive any further violent convictions regardless of the length of follow up inherent to the research design. Although the construct of psychopathy and the PCL-R provide a good foundation for understanding a psychopath’s tendency to use violence, they speak little to identifying features of those psychopathic offenders who may eventually abandon it (i.e., does not violently recidivate). Given that the nonrecidivating psychopathic offender (NRP) consistently reemerges across psychopathy research, it would seem valuable to better understand and identify his core features, both to improve risk assessment, and perhaps more importantly, to potentially inform violence reduction strategies with psychopathic or other high-risk offenders.

The NRP

In their book chapter review, Wong and Burt (2007) presented some data examining differences in criminogenic need between 65 psychopathic offenders who violently recidivated after 5 years in the community (labeled recidivating psychopaths or RPs) compared with 58 psychopathic offenders who had not (NRPs). All the men in their sample (featured in the present study) had participated for at least 4 months in an intensive correctional treatment program. In this work, safeguards and precautions were undertaken to maximize the accuracy of the RP/NRP distinction based on the problems inherent in “nonrecidivism.” Perhaps most notably, the absence of a known offense can be readily attributed to a lack of detection (rather than a lack of commission) of a new violent offense and strategies were adopted to help mitigate this concern (e.g., accessing multiple offender databases to rule in or rule out violent recidivism).

In the present study sample, Wong and Burt (2007) reported that RPs and NRPs had moderate to high scores across the majority of static and dynamic items on the Violence Risk Scale (VRS; Wong & Gordon, 1999-2003), a violence risk assessment and treatment planning tool; however, several important differences were evident between the two groups in their risk and need profiles. Across the static items, RPs were significantly younger at age of treatment admission, younger at age of first violent conviction, had more juvenile convictions, used violence more pervasively over their life spans, and came from more disrupted and chaotic childhood backgrounds. Although fewer differences were evident across the dynamic items, RPs were rated as having greater problems in the domains of interpersonal aggression, violence during incarceration, and more frequently being released to high-risk situations.¹

Correlates of Nonrecidivism in Psychopathic Offenders

Age

One of the most potent and reliable covariates of criminal or violent behavior is the offender’s age at the time they are released from custody (Lloyd, Mair, & Hough, 1994). Many sources document a sharp reduction in criminal offending beginning between the ages of 35 and 40. This effect has commonly been referred to as the “burnout” or “maturing out” effect in that almost all offenders appear to abandon their criminalized and violent lifestyles as they enter middle age. Conceptually, psychopathy is personality based, and as such it may be theoretically expected that psychopathic violence is somewhat “immune” to the burnout effect that is observed with other high-risk offenders.

The relationship between psychopathy, age, and general or violent recidivism has not been consistent, with some investigations finding no evidence for such effects with psychopathic offenders (Harris, Rice, & Cormier, 1991), while others have found age at release to predict most recidivism outcomes with psychopathic offenders (Olver & Wong, 2015) and still others finding that age effects may be limited to the frequency and severity of general offending patterns (but not violence) with psychopathic offenders (Hare, McPherson, & Forth, 1988). Harpur and Hare (1994), using a sample of offenders ranging in age from 16 to 70 at the time of PCL or PCL-R assessment, found that Factor 2 scores decreased sharply with age, whereas Factor 1 scores remained stable. The authors concluded that the age-related changes in the psychopath’s antisocial behavior are not necessarily paralleled by changes in the egocentric, manipulative, and callous traits that are “fundamental” to psychopathy. Consistent with the “maturing out” effect, Harpur and Hare speculate that it may be the impulsivity and sensation-seeking aspects of Factor 2 that are most influenced by age. Overall, the relationship between psychopathy, age, and recidivism is not well understood empirically or theoretically.

Correctional Treatment

The historic and current prevailing view is that the behavior of psychopaths is difficult or impossible to modify with traditional forms of treatment, intervention, and management (Hare, 1996; Lösel, 1998; Suedfeld & Landon, 1978). Early controlled studies demonstrated psychopaths showed less clinical improvement, displayed lower motivation, and were more likely to leave treatment early (Ogloff, Wong, & Greenwood, 1990), and that inappropriate interventions may even serve to increase posttreatment violence in psychopathic offenders (Rice, Harris, & Cormier, 1992).

In contrast, many researchers have challenged the skepticism surrounding the treatability of psychopathic offenders, indicating that such assertions are premature. Lösel (1998) noted that previous reviews of treatment effects with psychopathic offenders have used samples that were not necessarily psychopathic and have reviewed treatment interventions that were not modified to address relevant characteristics of violent or psychopathic offenders (i.e., risk, need, and responsivity, also known as RNR; Andrews et al., 1990). Simourd and Hoge (2000) note that “the treatment of psychopaths has an interesting history, short on quantity and long on lore . . . close examination of the treatment studies of psychopaths shows that most treatment efforts have failed to attend to the principles of effective interventions” (p. 270). In a systematic review of 42 psychopathy treatment studies, Salekin (2002) noted that several approaches and modalities appeared effective in reducing the severity of psychopathic traits and behaviors. However, the noted studies have significant shortcomings in addressing or understanding the NRP which include a focus on changing psychopathic personality features rather than reducing criminal risk, lack of a valid operationalization of psychopathy, lack of appropriate control groups, small samples, and/or outdated treatment approaches.

Since the review of Salekin (2002), the quantity and quality of psychopathy treatment research has improved (Salekin, Worley, & Grimes, 2010). Although many studies are correlational in design, they consistently support the notion that evidence informed intervention (i.e., adhering to the RNR principles) is associated with declines in future sexual offending and general violence among psychopathic sex offenders (Langton, Barbaree, Harkins, & Peacock, 2006; Looman, Abracen, Serin, & Marquis, 2005; Olver & Wong, 2009), reductions in general violence among psychopathic nonsexual violent offenders (Olver et al., 2013), and decreases in proxies of recidivism offense severity among treated incarcerated psychopathic offenders, relative to matched untreated controls (Wong, Gordon, Gu, Lewis, & Olver, 2012).

Wong and colleagues (Wong et al., 2012; Wong & Hare, 2005) observe that psychopathic offenders essentially have many of the criminogenic needs that are common of many other high-risk offenders (e.g., antisocial attitudes, negative social influences, substance use problems etc.). Indeed, research has demonstrated the chronic antisocial lifestyle features of psychopathy (i.e., Factor 2) to be a reliable predictor of violence (Yang, Wong, & Coid, 2010), to correlate highly with dynamic risk factors or criminogenic needs (Douglas, Yeomans, & Boer, 2005; Olver & Wong, 2009; Simourd & Hoge, 2000; Wong & Gordon, 2006), and that Factor 2 is susceptible to age-related changes (albeit presently limited to cross-sectional age-based research). In particular, when these aspects of risk associated with Factor 2 are specifically targeted for intervention, as per the need principle, reductions in recidivism have been empirically demonstrated. By contrast, Wong and colleagues observe that Factor 1 is likely best construed as a responsivity issue, that is, how individuals engage with and respond to services. In most regards, Factor 1 traits seem to be the antithesis of a motivated and compliant treatment candidate (e.g., pathological lying, manipulative, grandiose, denying responsibility, superficial, callous). Factor 1 traits interfere with engaging and maintaining therapeutic efforts and therefore undermine the ability to consistently target the relevant criminogenic needs (i.e., elements typical of Factor 2), to effect positive change (Wong et al., 2012; Wong & Hare, 2005).

Community Support

As conditional release can mark the beginning of an offender’s opportunity to function effectively in the community, investigation of variables during conditional release periods with psychopathic offenders may be salient predictors of successful reintegration and the nonrecidivism of psychopathic offenders. Zamble and Palmer (1995) found that the measurement of interpersonal problems, emotional dysphoria, social isolation, criminal peers, lifestyle instability, and length of employment improved their ability to predict recidivism, even after they accounted for PCL-R scores. In an examination of community variables and parole failure, Hart et al. (1988) found that psychopaths scored significantly worse than moderate and low psychopathy groups on these domains, but their community variables (e.g., stable employment, unstable relationships) did not predict recidivism. The extant findings suggest a complex relationship exists between parole failure, community factors, and psychopathy in which the degree of relationship between parole failure and community support may be dependent on the level of psychopathy. It is also important to bear in mind that these studies examined community factors based on information gathered and coded prior to the offenders’ release, rather than the assessment of such variables during the actual release period.

Nevertheless, past findings provide a useful starting point to understand how recidivism or reintegration occurs. For instance, the assessment of positive social interpersonal supports and stability of prosocial employment factors are likely central for successful reintegration with psychopathic and other high-risk offenders. Given the tendency for psychopathic violence to be instrumental (Woodworth & Porter, 2002), establishing a consistent prosocial means of creating an income may be an especially vital component in their reintegration. Moreover, the likelihood of effectively utilizing community resources may be influenced by participation in continued treatment planning, in that a good relapse prevention plan is considered central for the reintegration process (Wong & Hare, 2005). The relapse prevention model incorporates the notion that therapeutic gains can be facilitated or undermined by community factors. As with other high-risk offenders, psychopathic offenders who have strong prosocial relationships in the community, who have a good work ethic and can obtain and maintain employment, who are actively engaged in community treatment, and who have structured release over a longer period of time, would seem more likely to maintain a prosocial lifestyle after parole expiry.

Present Study

While a common research paradigm historically has featured comparisons of psychopathic and nonpsychopathic offenders, some unique directions have begun to highlight risk-relevant differences within a group of exclusively psychopathic violent offenders. The latter design promotes the possibility of discovering heterogeneity and distinctions within psychopaths. The NRP represents a particularly important subgroup of psychopathic offenders. As such, the present study builds on the initial data presented by Wong and Burt (2007) in an effort to provide a more comprehensive understanding of the NRP to inform risk assessment and possible risk reduction strategies with psychopathic offenders.

Potential differences between RPs and NRPs were investigated through the following four hypotheses:

Hypothesis 1: NRPs were anticipated to be significantly older on average at release from custody compared with RPs.

Hypothesis 2: NRPs were anticipated to score significantly higher than RPs on measures associated with progress from violence reduction treatment; specifically, that they would demonstrate larger VRS change scores and longer overall stay in programming. There were no predictions as to the magnitude of possible differences.

Hypothesis 3: NRPs were anticipated to score more favorably than RPs on a series of important community support factors, suggesting that they may have more community supports and/or make better use of such supports.

Hypothesis 4: NRPs and RPs were expected to have different PCL-R profiles at the broad factor level; specifically, NRPs are anticipated to score lower on Factor 2 compared with RPs.

Method

Sample

Sample characteristics for the present study are described in detail in the original study of the NRP (Wong & Burt, 2007) and the reader is directed to that research for a more thorough review of the sample and setting (i.e., the intensive treatment program). In brief, to assess the impact of correctional treatment on postrelease behavior, the sample was comprised of federal male inmates who had been involved (for at least 4 months) in a high-intensity violent offender treatment program. They had also received a PCL-R score of 25 or greater, using guidelines for research and/or file-based diagnoses (Wong, 1988).

Offenders who reoffended violently within 5 years of their first community release after treatment were labeled “RPs,” whereas those who did not receive a violent conviction within 5 years were labeled “NRPs.” As such, the final sample was comprised of 123 offenders, of which 58 offenders (47.2% of the sample) had no violent convictions after 5 consecutive years in the community (NRPs) whereas 65 offenders (52.8% of the sample) violently recidivated within 5 years of release (RPs). The sample spent an average of 6.6 months in treatment (SD = 4.44).

Materials

PCL-R

The PCL-R (Hare, 1991, 2003) is a 20-item symptom construct rating scale designed to assess and diagnose psychopathy. Each item is scored on a 0 (not present), 1 (partially present), 2 (present) scale. Possible scores range from 0 to 40, with higher scores indicating the presence of a greater number of psychopathic traits. Historically, cutoffs of 30 (file + interview based ratings) or 25 (for file only ratings and/or for research purposes) have been used to diagnose psychopathy. As mentioned above, the men were identified using the 25-point cutoff. Although an argument may be made that the 25-point cutoff could over-classify individuals as psychopaths, we retained this cutoff given that file-based ratings can sometimes result in underestimating the interpersonal and affective features of the syndrome and to ensure sufficient N. PCL/PCL-R ratings were conducted by trained clinical and research staff as part of directing treatment and assessing risk of offenders during their stay at the treatment program.

VRS

The VRS (Wong & Gordon, 1999-2003) was designed to assess violence risk, identify targets for violence reduction treatment, and assess changes in risk posttreatment. It consists of 20 dynamic and six static items rated on a 4-point (0, 1, 2, 3) ordinal scale. Dynamic factors are conceptualized as treatment targeted risk factors. Those items rated either a 2 or 3 are considered to be appropriate targets for treatment. Using a modified version of Prochaska, DiClemente, and Norcross’s (1992) stages of change model, each dynamic predictor is rated pre- and posttreatment to assess the stage of change (i.e., Precontemplation, Contemplation, Preparation, Action, or Maintenance) the person is at respectively. Progression in treatment from a less advanced to a more advanced stage of change indicates improvement and a reduction in risk for violence for that risk item. Progression from one stage of change to the next more advanced stage is scored as a 0.5 reduction in the pretreatment rating of the item; progression through two stages is scored as a 1.0 point reduction and so on. The total change score is the sum of the change scores for all the dynamic items. Research demonstrates that the VRS has moderate to high predictive accuracy (cf. Rice & Harris, 2005) for future violence in a large general prison inmate sample (Wong & Gordon, 2006), treated high-risk violent offenders (Lewis, Olver, & Wong, 2013), and psychiatric inpatients (Dolan & Fullam, 2007) as well as has demonstrated good convergent validity with respect to other well-established forensic assessment measures (Wong & Gordon, 2006).

VRS ratings were conducted using file information outlined in the previous study (Wong & Burt, 2007). In brief, inter-rater reliability was assessed based on 13 randomly sampled cases, in which 72% of VRS pretreatment inter-ratings were found to be of an exact match, and 95% were an exact match or discrepant by a single point. Eighty-two percent of VRS change ratings were exact matches and 99% were exact or within a single point.

Community Support Rating Scale (CSR)

The CSR was designed by the principal investigator to measure community variables that theoretically may discriminate between RPs and NRPs (see appendix for scoring details). The measure consists of eight items grouped together into three domains: (a) “Community Establishment” consisted of three dichotomous items thought to tap an offender’s ability to initially establish himself in the community and utilize community supports; (b) “Community Utilization” consisted of the offender’s work ethic, attitude toward treatment, and intensity of therapeutic contact, and was thought to measure the offender’s ability to positively maintain any community supports that he established; and (c) the remaining two items were aggregated as a measure of the structure and duration of the offender’s supervisory period and dubbed “Community Supervision.” Internal consistency (KR-20) of the domains was as follows: .52 Community Establishment, .54 Community Utilization, .09 for the two Community Supervision items, and .57 for the CSR total score. Across 11 double coded cases, 75% of the items rated were exact matches and the mean kappa value reflected moderate to good agreement between raters (.69). Data were missing for 23 (2.3%) of the items for CSR ratings. Missing values were replaced with item mean scores, which is a conservative approach as this method minimizes NRP and RP group differences, rather than enhance them.

Violent Criminal History

Given the potency of violent criminal history as a predictor of violent recidivism (Villeneuve & Quinsey, 1995), three such variables were coded: annual rate of pretreatment violent convictions, average sentence length per pretreatment violent conviction, and type of past violent offense. First, pretreatment rate of violent convictions was calculated by dividing the total number of violent convictions incurred prior to treatment by the total number of years the offender was in the community (from the date of his first violent conviction to the date he entered treatment). Time incarcerated was not included in this calculation Second, average sentence length per pretreatment violent conviction was measured by dividing the total number of years an offender had been sentenced for violent offenses during the pretreatment period by the total number of pretreatment violent offenses committed. Third, type of past offenses was examined by grouping pretreatment violent offenses into four nonoverlapping categories: assaults, robbery, sexually violent offenses, and offenses resulting in death; only 12% of the RPs’ offenses and 19% of the NRPs’ offenses remained unclassified supporting the use of this scheme.

Data Analytic Plan

Several planned analyses were performed followed by a series of post hoc analyses that were informed by the a priori analyses. The planned analyses consist of a series of RP–NRP group comparisons (t tests) on key study variables, namely, age at release (Hypothesis 1), treatment-relevant variables, including VRS change score (Hypothesis 2), the CSR and its eight individual items (Hypothesis 3), and PCL-R factor and total scores (Hypothesis 4). The magnitude of RP–NRP group differences are represented by Cohen’s d, an effect size metric reporting differences in standard deviation units where roughly values of .20 correspond to low/small, .50 medium, and .80 large (Cohen, 1992). In light of emergent findings, we turn to a series of post hoc analyses examining group differences in pretreatment violence offense history, specifically, the density of violent offending and the types of past violent offenses. These differences in turn are linked to the men’s PCL-R profiles and an explanatory framework is presented in an attempt to integrate the pattern of findings. We detail the procedure and rationale for post hoc analyses in the Results section as each new set of analyses and corresponding findings are presented.

Results

Hypothesis 1: NRP–RP Age Differences

Means, standard deviations, and planned t-test comparisons for RPs and NRPs across all independent variables are presented in Table 1. In support of Hypothesis 1, NRPs were significantly older at the time of release on average than RPs, with the magnitude of the difference being approximately one half of a standard deviation.

Table 1:

Means, Standard Deviations, and Planned Comparisons (d and t Test) for RPs and NRPs

Variable	RPs (n = 65)		NRPs (n = 58)		d	df	t	p
Variable	M	SD	M	SD	d	df	t	p
Age at release	29.4	5.9	33.0	7.4	−.54	121	3.0	.004
Treatment length	0.6	0.3	0.6	0.4	.00	121	0.6	.560
VRS pre-tx total score	62.3	7.4	57.1	9.2	.63	121	3.4	.001
VRS pre-tx static	14.4	2.8	12.1	3.1	.79	121	4.3	.001
VRS pre-tx dynamic	47.9	6.1	45.0	7.0	.45	121	2.4	.030
VRS change score	2.4	2.6	2.5	2.6	−.04	121	0.4	.700
VRS post-tx total score	59.9	8.2	54.6	10.4	.57	121	3.2	.002
VRS post-tx dynamic	45.5	7.1	42.6	8.3	.38	121	2.1	.040
CSR total	8.0^a	3.7	5.9^b	2.9	.64	74	2.2	.033
Community establishment	1.5^a	1.0	0.7^b	0.8	.89	74	4.0	.001
Community utilization	5.6^a	2.7	4.4^b	2.2	.49	74	1.7	.090
Structure of supervision	1.4^a	1.3	1.0^b	0.9	.37	74	1.6	.110
PCL-R Factor 1	10.2	2.2	11.4	2.7	−.49	121	2.7	.008
PCL-R Factor 2	14.6	1.9	13.8	1.9	.42	121	2.5	.014
PCL-R total	28.7	3.1	29.6	3.3	−.28	121	0.9	.352

Note. Sample sizes for comparisons involving CSR items are smaller because they were not granted a period of conditional release. Significant findings are denoted by the respective variable and p values in bold font. RP = recidivating psychopath offender; NRP = nonrecidivating psychopath offender; CSR = Community Support Rating; VRS = Violence Risk Scale; PCL-R = Psychopathy Checklist–Revised.

n = 34. ^bn = 42.

Hypothesis 2: Violence Risk, Criminogenic Need, and Treatment Progress

Contrary to Hypothesis 2, no significant differences were observed, however, between RPs and NRPs on the amount of treatment change/risk reduction as assessed by the VRS. However, NRPs scored significantly lower than RPs on the VRS pre- and posttreatment. VRS posttreatment scores for both RPs and NRPs were significantly lower than pretreatment scores, indicating that both groups demonstrated risk reduction as measured by the VRS when reassessed following therapy, t(64) = 7.15, p < .001 and t(57) = 7.75, p < .001, respectively. Moreover, VRS pretreatment differences were sufficiently large that NRPs, even at pretreatment, were at a lower risk for violent recidivism than RPs either before or after RPs had completed treatment (i.e., NRPs’ pretreatment VRS scores were lower than even RPs’ posttreatment scores, albeit differences failed to reach statistical significance, t(121) = −1.52, p = .13). Overall, VRS pretreatment differences were the central NRP/RP group difference (rather than VRS change scores).

Hypothesis 3: Community Support

Of the 123 offenders in the sample, 76 (61.7%) were granted conditional release. Forty-three (74.1%) of the 58 NRPs received conditional release whereas only 33 (50.8%) of the 65 RPs received conditional release, χ² = 4.24, p = .039. In addition, those who received conditional release rated significantly lower on the VRS total score, both before, t(121) = 2.8, p = .006, and after, t(121) = 2.4, p = .017, treatment. Among offenders granted conditional release, NRPs had significantly better (i.e., lower) mean CSR scores on certain items than RPs in support of Hypothesis 3. Specifically, NRPs were significantly more likely to establish themselves in the community upon release by finding employment, attending community treatment, and developing relationships with an intimate partner; the magnitude of these group differences were more than one half of a standard deviation, representing effects that would be considered approximately medium in magnitude (Table 2). Group comparisons on the remaining community utilization and community supervision items did not attain significance.

Table 2:

Means, Standard Deviations, Effect Size Magnitudes (d), and t-Test Comparisons Between RPs and NRPs on CSR Items

CSR items	RPs (n = 34)		NRPs (n = 42)		d	df	t	p
CSR items	M	SD	M	SD	d	df	t	p
1. Any relationship with a partner	0.58	0.50	0.31	0.47	.56	74	2.37	.020
2. Was the offender employed during this release?	0.58	0.50	0.26	0.44	.68	64	2.82	.006
3. Was the offender in treatment during this release?	0.28	0.47	0.07	0.26	.57	46	2.33	.024
Community establishment total	1.44	1.47	0.64	1.17	.61	74	2.64	.010
4. Work ethic	1.65	1.20	1.13	1.07	.46	74	1.63	.108
5. Attitude toward treatment	1.60	0.99	1.51	0.91	.10	74	0.38	.702
6. Intensity of therapeutic contact	1.92	1.32	1.52	1.19	.32	74	1.36	.178
Community utilization total	5.17	3.51	4.16	3.17	.30	74	1.32	.192
7. Supervision structure	1.18	1.03	0.86	0.93	.33	74	1.42	.159
8. Amount of time from date of conditional release to WED	0.21	0.54	0.12	0.40	.19	74	0.81	.420
Community supervision total	1.39	1.57	0.98	1.33	.28	74	4.41	.222
CSR total score	8.00	6.55	5.78	5.67	.37	74	1.58	.118

Note. Items 1 to 3 were rated on a dichotomous 0 (yes), 1 (no) scale. Items 4 to 7 were rated on a 4-point Likert-type scale (0, 1, 2, 3) with higher ratings indicating poorer adjustment. Item 8 was rated on a 4-point scale of 0 (10+ months), 1 (7-9 months), 2 (4-6 months), 3 (1-3 months), and 4 (less than 1 month). Significant findings are denoted by the respective variable and p value in bold font. RP = recidivating psychopath offender; NRP = nonrecidivating psychopath offender; CSR = Community Support Rating; WED = Warrant Expiry Date.

Hypothesis 4: PCL-R Profiles

NRPs scored significantly lower on the PCL-R Factor 2 than did RPs (supportive of Hypothesis 4); however, NRPs also scored significantly higher than the RPs on PCL-R Factor 1. As previously noted, NRPs and RPs did not differ on PCL-R total score (Table 1). Moreover, PCL-R Factor 1 and Factor 2 ratings were not significantly related (r = −.054, p = .554). PCL-R Factors also had opposite linear associations to VRS ratings. Factor 2 had a significant positive relationship with all VRS ratings (rs = .21 to .27, all ps < .05), whereas Factor 1 was negatively associated with VRS ratings, particularly static scores (r = −.20, p < .05), and less so, with dynamic scores (r = −.07 to −.11, ns). PCL-R total score was subsequently unrelated to VRS total ratings. Finally, correlational findings indicated that the interpersonal nature of psychopathic personality was responsible for Factor 1’s strong relationship to NRP status (Table 4). While Factor 1 was significantly negatively correlated with binary RP/NRP status (i.e., higher Factor 1 scores were associated with violent nonrecidivism among psychopaths), the strength of this association was most evident with the Interpersonal facet (r = −.26, p < .01) of Factor 1, whereas the Affective facet of Factor 1 was not related to RP/NRP status (r = −.08, ns).

Post Hoc RP–NRP Comparisons of Violent Offense History

The finding of a higher mean PCL-R Factor 1 score among NRPs prompted several post hoc analyses. Binary RP/NRP status (coded 1 = violently recidivate, 0 = did not violently recidivate) was utilized as an outcome variable in bivariate analyses to measure the relationship between NRP/RP group membership and post hoc analysis variables. Table 3 presents the results of NRP–RP group comparisons on the three violent criminal history variables included in post hoc analyses: annual rate of pretreatment violent convictions, average sentence length per pretreatment violent conviction, and type of past violent offense (see Table 3 for results of group comparisons). Rate of pretreatment violent convictions reflected the average number of violent offenses committed during periods of release in the community and it would be expected that offenders of greater risk would offend with greater rapidity and density; the rate of pretreatment violent offending was significantly higher for RPs than for NRPs. Second, NRPs received significantly longer average sentence lengths per pretreatment violent conviction than RPs. Third, proportionally, the most common type of past violent offenses for psychopaths (i.e., both RPs and NRPs) were assaults and robberies.

Table 3:

Post Hoc Analyses: RP–NRP Comparisons on Key Violence History Variables

Pretreatment violence history variables	RP (n = 65)			NRP (n = 54)			r	t or χ²	p
Pretreatment violence history variables	M	SD	%	M	SD	%	r	t or χ²	p
Rate of pretreatment violent offending	0.49	0.33	—	0.32	0.22	—	−.30	3.31	.002
Average sentence length per pretreatment violent conviction	2.6	3.0	—	4.9	4.4	—	−.30	3.26	.002
Prior robberies	—	—	24	—	—	39	−.15	2.81	.094
Prior assaults	—	—	50	—	—	26	.24	6.76	.009
Robbery–assault index	−1.14	3.09	—	0.70	3.23	—	−.28	3.16	.002
Instrumental tendency	0.18	0.27	—	0.41	0.50	—	−.35	4.59	<.001

Note. r represents correlations (phi and point biserial) between pretreatment criminal history variable and recidivism status (i.e., RP/NRP group membership). Significant findings are denoted by the respective variable and p values in bold font. RP = recidivating psychopath offender; NRP = nonrecidivating psychopath offender.

These three sets of violence history analyses thus informed a series of follow-up analyses. The total number of robbery and assault convictions were predictably positively related to Rate of Violent Offending (r = .36, p < .001 and r = .68, p < .001, respectively); however, they were negatively related to one another (r = −.26, p = .005), perhaps indicating that some psychopathic offenders have a tendency toward committing one type of violent crime over another. A new variable (robbery–assault index) was thus created through subtracting the total number of assaults from the total number of robberies committed. A positive robbery–assault index score would indicate that the offender had committed a greater number of robberies than assaults, whereas a negative score would reflect that the offender had committed more assaults than robberies. As reported in Table 3, the robbery–assault indexes were significantly higher in the NRP group than in the RP group, with NRPs tending to commit more robberies than assaults, whereas RPs committed more assaults than robberies.

To unpack these findings further, correlations between PCL-R Factor 1, and Factor 1 facets (i.e., Interpersonal and Affective) with pretreatment violent offense history variables were computed and are presented in Table 4. As reported above, the Interpersonal facet of Factor 1 was significantly negatively correlated with violent recidivism status. The same series of relationships emerged for the other pretreatment violent history variables; high scores on the Interpersonal facet were significantly associated with a tendency to commit robbery (i.e., instrumentally motivated violence) over assault. No relationships were significant for the Affective facet.

Table 4:

Correlations Between PCL-R Factor 1 and Facet Scores With Pretreatment Violent Conviction History

Pretreatment violence history variables	Factor 1 total	Factor 1 facets
Pretreatment violence history variables	Factor 1 total	Interpersonal	Affective
Rate of pretreatment violent offending	−.21*	−.24**	−.06
Average sentence length per pretreatment violent conviction	.30***	.35***	.08
Robbery–assault index	.28**	.29***	.12
NRP–RP (violent recidivism) status	−.23*	−.26**	−.08

Note. PCL-R = Psychopathy Checklist–Revised; NRP = nonrecidivating psychopath; RP = recidivating psychopath.

p < .05. **p ≤ .01. ***p ≤ .001.

To facilitate interpretation, median splits were used to identify offenders with high and low scores on the Interpersonal facet as well as through scoring high and low on the robbery–assault index. Group membership was coded dichotomously, with those above the median split on both variables labeled as having an “instrumental tendency” (coded 1), with the remainders having no such tendency (coded 0). Instrumental tendency therefore encapsulated a personality feature of being exploitative, conning, and manipulative combined with a preference for crimes involving material gain. As seen at the bottom of Table 3, instrumental tendency was significantly negatively correlated with violent recidivism status (i.e., being more common among NRPs), as were the robbery–assault index, rate of pretreatment violent offending, and average sentence length per violent conviction. In short, NRPs appeared to demonstrate a predilection for instrumental violence and possessed the prominent interpersonal features of psychopathy.

Discussion

The present study examined group differences between violently recidivating and nonrecidivating psychopathic offenders (RPs and NRPs, respectively) on a series of criminological variables. The primary objective was to identify distinguishing features of the NRP relative to the more prominently researched RP.

The present findings support the hypothesis that NRPs are typically older offenders on average compared with RPs. This finding runs counter to notions of the immutability of psychopathic violence and extends the findings of Olver and Wong (2015) who found age at release to be associated with a decrease in all recidivism outcomes, even among psychopathic offenders. The results are also consistent with the well-established age–crime curve that demonstrates a precipitous decline in all forms of criminal behavior as people age; evidently, age could be the final arbiter for even the most serious criminals, psychopaths included, although longitudinal designs (e.g., Sampson & Laub, 2003) may prove to be a more veracious test of this.

VRS pretreatment differences revealed important heterogeneity among psychopathic offenders in this sample in terms of both static and dynamic risk variables. In essence, not all psychopaths pose the same risk for violence, even prior to treatment efforts. Consistent with earlier findings of differences between the two groups on their profiles of static and dynamic VRS items (Wong & Burt, 2007), psychopathic offenders who did not have an enduring and pervasive tendency to use violence (i.e., higher static scores) had a better chance of desisting from violence after release. Moreover, both NRPs and RPs evidenced similar and significant reductions in VRS ratings from pre- to posttreatment. Change in need areas (reduction of risk) appeared nonspecific, in that a modest reduction was found across about half of the VRS dynamic items, with only one notable difference between RPs and NRPs. The major finding of the present research is that differences in risk between RPs and NRPs were evident, not due to changes associated with treatment, but rather due to VRS score differences that were apparent even at pretreatment. The present findings support a theoretical premise that violence per se is not an inherent feature of psychopathy and the propensity of psychopaths to use violence should not be conceptualized as uniform. Subsequently, as with any forensic measure, over-reliance on a single instrument to assess risk has the potential to grossly over-predict violence, even with psychopathic offenders (i.e., the NRP constitutes about one in four psychopathic offenders).

The use of a treated sample in this study provided a unique opportunity to investigate the potential effectiveness of correctional programming with psychopathic offenders; however, the use of a minimum period of time (i.e., four months) of treatment participation before meeting sample inclusion criteria also led to some important limitations. For instance, past research has demonstrated a strong relationship between treatment noncompletion and risk for recidivism and has found that psychopaths are especially prone to attrition (Olver, Stockdale, & Wormith, 2011). The present study is not well designed for the examination of treatment attrition in that inclusion into the study relied on offenders to have already maintained a minimum period of attendance. As such, the present finding that length in treatment did not discriminate NRPs from RPs is not surprising.

Significantly more NRPs were granted conditional release, both in terms of overall number, and in terms of proportion. This may suggest that the Parole Board of Canada is able to discriminate lower risk offenders, even within high-risk groups. It is plausible that the RP’s tendency to be more involved in institutional violence, more pervasive history of criminal and interpersonal violence, and to ultimately be less compliant to community plans and supports may have informed conditional release decisions. Several findings point to the notion that NRPs were better prepared for their release. The present results have an interesting juxtaposition with past findings demonstrating mixed support for the role of community factors in predicting recidivism among psychopathic offenders (Hart et al., 1988; Zamble & Palmer, 1995). There are a host of methodological and sample differences across studies (e.g., the CSR is unique to this study) that account for differences in findings. Overall, it seems likely that community factors are a set of robust but diverse and complex predictors that merit further research.

Although the chronic antisocial lifestyle features of the PCL-R, broadly speaking, (i.e., Factor 2) had the expected relationship with the RP/NRP groups, the finding that Factor 1 had an inverse relationship (i.e., was strongly and positively associated with NRP status) was unexpected. We believe the weak correlation between Factors 1 and 2 in the overall sample is a consequence of this being an extreme and high-scoring sample of men with a more restricted range than would be seen in a general offender population. Given the heterogeneity in PCL-R profiles, when total scores are high, Factor 1 and Factor 2 are not both necessarily high; different combinations of Factors 1 and 2 contribute to total scores. (In a broader sample with the full range of scores, one would expect higher Factor 1 to be positively associated with higher Factor 2.) Indeed, analysis of PCL-R Factor 1 items revealed that the Interpersonal facet accounted for Factor 1’s strong linear relationship with the NRP; offenders rated higher on manipulation, pathological lying, superficial charm, and grandiosity were more likely to be NRPs. In essence, NRPs tended to possess traits that emphasized the ability to exploit and manipulate others for their own personal gain.

Post hoc analyses of pretreatment violent offending patterns revealed that although RPs received violent convictions at a significantly faster rate than did NRPs, they also received significantly shorter sentences per violent conviction. The latter finding reflected a higher prevalence of robbery convictions (which carried an average of four times the sentence length) among NRPs, whereas RPs had been primarily convicted of assault. A significant positive relationship between PCL-R Factor 1 interpersonal traits and a “preference” for robbery suggested that both variables may “tap into” a common underlying construct. In particular, psychopaths with a strong manipulative and exploitative personality style may also evidence a similar type of instrumental approach to criminal behavior.

A composite variable termed “Instrumental Tendencies” provided the best means to discriminate the NRP from the RP and showed good replication in cross validation. For psychopathic offenders with instrumental tendencies, violence is simply a means to an end. Within the context of treatment, the psychopath’s reputed instrumental approach to violence has been claimed to be an asset in their eventual ability to simply “give up” violence when better strategies are recognized and implemented. In essence, clinicians are encouraged to utilize the psychopath’s own self-serving and egocentric attributes to motivate change (Wong et al., 2012; Wong & Hare, 2005). Presumably, psychopaths who primarily utilize instrumental aggression would abandon such approaches if a better or more efficient strategy was introduced, or if it became obvious to him that violence was not an effective or viable strategy, as may be the case for NRP’s. Conversely, an entrenched interpersonal pattern to resolve conflict by use of aggression and related impulsive anger may be more difficult to modify or relinquish. As noted in the earlier study, RPs had a more pervasive and earlier pattern/reliance on violence, used more interpersonal aggression, and continued the use of violence even in institutional settings (Wong & Burt, 2007).

The examination of the CSR scores within the community release sample underscored the potential of NRPs to effectively engage and utilize positive resources and subsequently abandon their reliance on old methods and strategies. Ultimately, and once in the community, it seems that the NRP’s ability to develop a stable romantic relationship, obtain a job, and to attend outpatient correctional treatment were important correlates of their success (nonrecidivism status). By contrast, the RP seemed unable or unwilling to set even a basic prosocial foundation, even during supervised conditional release.

The present study builds on preliminary NRP data, which were limited to the comparison of differing areas criminogenic need; however, this study also contains some of the same limitations. In particular, the sample is self-selected in that offenders had to volunteer for high-intensity long-term treatment and to remain in it. As such, the present group likely represents an important subset of psychopathic offenders who are more motivated, more compliant, and less volatile than are usually represented in the literature. This likely accounts for the higher percentage of NRPs found in this sample relative to past research, and also suggests that even with such a highly selected sample just over half of psychopathic offenders will still fail (i.e., violently recidivate). However, in our view, this limitation is offset by having an exclusively psychopathic sample which is a rare component in the extant psychopathy literature. The present research approach allows for important differences and the potential heterogeneity of psychopathic offenders to be explored. Moreover, it is our view that the NRP represents a particularly important and overlooked area within the psychopathy literature, with promising potential to provide directions for violence reduction in high-risk offenders.

Moreover, a further potential limitation is that there are other possible candidate variables and correlates that may differentiate between RPs and NRPs. The present study presented a selection of potential domains informed by the extant literature that had potential to discriminate these two groups; this is not exhaustive, however. For instance, research has shown cognitive ability to be inversely associated with criminal recidivism (Gendreau, Little, & Goggin, 1996), and in some studies, to even be positively associated with higher Factor 1 scores (Vitacco, Neumann, & Wodushek, 2008). Research in the domain of “successful psychopathy” has also identified neurocognitive correlates, such as higher executive functioning, that discern successful (i.e., not criminally convicted) from unsuccessful (i.e., criminally convicted) psychopaths (Ishikawa, Raine, Lencz, Bihrle, & Lacasse, 2001). In all, it is possible that higher levels of functioning in certain neurocognitive domains (e.g., global cognitive ability, executive functions) could be a further feature that sets RPs and NRPs apart, and further exploration of this possibility is a viable candidate for further research.

The present findings support the exploration of heterogeneity of psychopaths and may overlap with other recent research efforts attempting to explore subtypes of psychopathy. For instance, it is possible that different proportions of primary and secondary psychopathic offenders are found among NRPs. Although recent work on a large Canadian correctional sample found no differences in base rates of violent and general recidivism between these two psychopathy subtypes (Olver, Sewall, Sarty, Lewis, & Wong, 2015), recent research has suggested that there may be different areas of treatment emphasis and need (Olver et al., 2015; Skeem, Johansson, Andershed, Kerr, & Eno Louden, 2007). The present study supports this notion.

In conclusion, the present findings suggest that increasing age, criminal history, treatment factors, and community support variables are a conduit toward better acquainting ourselves with the NRP and by extension, are associated with reduced risk of violence among psychopathic offenders. We would thus argue that efforts to reduce risk with psychopathic offenders are not without hope. To the contrary, a combination of three variables provided a promising starting point to identify possible NRPs.

Footnotes

Appendix

The views, opinions, and assumptions expressed in this paper are those of the authors and do not necessarily reflect the views or official positions of the British Columbia Ministry of Children and Family Development, the University of Saskatchewan, the University of Nottingham, or Swinburne University of Technology. The authors thank Brenda Maire, Abigail Abada (nee Mallillin), and Karen Parhar for their assistance with data collection.

Notes

Grant N. Burt, PhD, is a registered psychologist in the province of British Columbia. He currently works at an outpatient forensic community mental health clinic where he provides assessment and treatment services to court adjudicated young offenders.

Mark E. Olver, PhD, is professor in the Department of Psychology, University of Saskatchewan, and a registered doctoral psychologist in the province of Saskatchewan. He conducts research and provides training on the assessment and treatment of psychopathic offenders and other high-risk offender groups.

Stephen C. P. Wong, PhD, Institute of Mental Health University of Nottingham and Department of Psychology, University of Saskatchewan and Swinburne Institute of Technology. He has provided consultation and training services internationally on the assessment of violence risk and the treatment of high-risk psychopathic offenders. He is the lead author of the Violence Risk Scale tools.

References

Andrews

D. A.

Zinger

Hoge

R. D.

Bonta

Gendreau

Cullen

F. T.

(1990). Does correctional treatment work? A clinically relevant and psychologically informed meta-analysis. Criminology, 28, 369-464.

Cohen

(1992). A power primer. Psychological Bulletin, 112, 155-159. doi:10.1037/0033-2909.112.1.155

Cox

Clark

J. C.

Edens

J. F.

Smith

S. T.

Magyar

(2013). Jury panel member perceptions of interpersonal-affective traits of psychopathy predict support for execution in a capital murder trial simulation. Behavioral Sciences & the Law, 31, 411-428. doi:10.1002/bsl.2073

Dolan

Fullam

(2007). The validity of the Violence Risk Scale second version (VRS-2) in a British forensic inpatient sample. Journal of Forensic Psychiatry & Psychology, 18, 381-393. doi:10.1080/14789940701489390

Douglas

K. S.

Yeomans

Boer

D. P.

(2005). Comparative validity analysis of multiple measures of violence risk in a sample of criminal offenders. Criminal Justice and Behavior, 32, 479-510. doi:10.1177/0093854805278411

Edens

J. F.

Boccaccini

M. T.

Johnson

D. W.

(2010). Inter-rater reliability of the PCL-R total and factor scores among psychopathic sex offenders: Are personality features more prone to disagreement than behavioral features? Behavioral Sciences and the Law, 28, 106-119. doi:10.1002/bsl.918

Gendreau

Little

Goggin

(1996). A meta-analysis of the predictors of adult offender recidivism: What works! Criminology, 34, 575-595. doi:10.1111/j.1745-9125.1996.tb01220.x

Guy

L. S.

Edens

J. F.

Anthony

Douglas

K. S.

(2005). Does psychopathy predict institutional misconduct among adults? A meta-analytic investigation. Journal of Consulting and Clinical Psychology, 73, 1056-1064. doi:10.1037/0022-006X.73.6.1056

Hare

R. D.

(1991). The Hare Psychopathy Checklist-Revised. Toronto, Ontario, Canada: Multi-Health Systems.

10.

Hare

R. D.

(1996). Psychopathy, a clinical construct whose time has come. Criminal Justice and Behavior, 23, 25-54. doi:10.1177/0093854896023001004

11.

Hare

R. D.

(2003). The Hare Psychopathy Checklist–Revised (2nd ed.). Toronto, Ontario, Canada: Multi-Health Systems.

12.

Hare

R. D.

McPherson

L. M.

(1984). Violent and aggressive behavior by criminal psychopaths. International Journal of Law and Psychiatry, 7, 35-50. doi:10.1016/0160-2527(84)90005-0

13.

Hare

R. D.

McPherson

L. M.

Forth

A. E.

(1988). Male psychopaths and their criminal careers. Journal of Consulting and Clinical Psychology, 56, 710-714. doi:10.1037/0022-006X.56.5.710

14.

Harpur

T. J.

Hare

R. D.

(1994). Assessment of psychopathy as a function of age. Journal of Abnormal Psychology, 103, 604-609. doi:10.1037/0021-843X.103.4.604

15.

Harris

G. T.

Rice

M. E.

Cormier

C. A.

(1991). Psychopathy and violent recidivism. Law and Human Behavior, 15, 625-637. doi:10.1007/BF01065856

16.

Hart

S. D.

Kropp

P. R.

Hare

R. D.

(1988). Performance of psychopaths following conditional release from prison. Journal of Consulting and Clinical Psychology, 56, 227-232. doi:10.1037/0022-006X.56.2.227

17.

Ishikawa

S. S.

Raine

Lencz

Bihrle

Lacasse

(2001). Autonomic stress reactivity and executive functions in successful and unsuccessful criminal psychopaths from the community. Journal of Abnormal Psychology, 110, 423-432. doi:10.1037/0021-843X.110.3.423

18.

Langton

C. M.

Barbaree

H. E.

Harkins

Peacock

E. J.

(2006). Sex offenders’ response to treatment and its association with recidivism as a function of psychopathy. Sexual Abuse: A Journal of Research and Treatment, 18, 99-120. doi:10.1177/107906320601800107

19.

Leistico

A. R.

Salekin

R. T.

DeCoster

Rogers

(2008). A large-scale meta-analysis relating the Hare measures of psychopathy to antisocial conduct. Law and Human Behavior, 32, 28-45. doi:10.1007/s10979-007-9096-6

20.

Lewis

Olver

M. E.

Wong

S. C. P.

(2013). The Violence Risk Scale: Predictive validity and linking changes in risk with violent recidivism in a sample of high risk offenders with psychopathic traits. Assessment, 20, 150-164. doi:10.1177/1073191112441242

21.

Lloyd

Mair

Hough

(1994). Explaining reconviction rates: A critical analysis. London, England: Home Office.

22.

Looman

Abracen

Serin

Marquis

(2005). Psychopathy, treatment change, and recidivism in high-risk, high-need sexual offenders. Journal of Interpersonal Violence, 20, 549-568. doi:10.1177/0886260504271583

23.

Lösel

(1998). Treatment and management of psychopaths. In Cooke

D. J.

Forth

A. E.

Hare

R. D.

(Eds.), Psychopath: Theory, research and implications for society (pp. 303-354). Dordrecht, The Netherlands: Kluwer.

24.

Murrie

D. C.

Boccaccini

M. T.

Turner

D. B.

Meeks

Woods

Tussey

(2009). Rater (dis)agreement on risk assessment measures in sexually violent predator proceedings. Psychology, Public Policy, and Law, 15, 19-53. doi:10.1037/a0014897

25.

Ogloff

J. D.

Wong

S. C. P.

Greenwood

(1990). Treating criminal psychopaths in a therapeutic community program. Behavioral Sciences & the Law, 8, 181-190. doi:10.1002/bsl.2370080210

26.

Olver

M. E.

Lewis

Wong

S. C. P.

(2013). Risk reduction treatment of high-risk psychopathic offenders: The relationship of psychopathy and treatment change to violent recidivism. Personality Disorders: Theory, Research, and Treatment, 4, 160-167. doi:10.1037/a0029769

27.

Olver

M. E.

Sewall

L. A.

Sarty

G. E.

Lewis

Wong

S. C. P.

(2015). A cluster analytic investigation and external validation of psychopathic offender subtypes in a multisite sample of Canadian federal offenders. Journal of Abnormal Psychology, 124, 355-371. doi:10.1037/abn0000038

28.

Olver

M. E.

Stockdale

K. C.

Wormith

J. S.

(2011). A meta-analysis of predictors of offender treatment attrition and its relationship to recidivism. Journal of Consulting and Clinical Psychology, 79, 6-21. doi:10.1037/a0022200

29.

Olver

M. E.

Wong

(2011). Predictors of sex offender treatment dropout: Psychopathy, sex offender risk, and responsivity implications. Psychology, Crime & Law, 17, 457-471. doi:10.1080/10683160903318876

30.

Olver

M. E.

Wong

S. C. P.

(2009). Therapeutic responses of psychopathic sexual offenders: Treatment attrition, therapeutic change, and long term recidivism. Journal of Consulting and Clinical Psychology, 77, 328-336. doi:10.1037/a0015001

31.

Olver

M. E.

Wong

S. C. P.

(2015). Short and long-term recidivism prediction of the PCL-R and the effects of age: A 24-year follow-up. Personality Disorders: Theory, Research, and Treatment, 6, 97-105. doi:10.1037/per0000095

32.

Prochaska

J. O.

DiClemente

C. C.

Norcross

J. C.

(1992). In search of how people change: Applications to the addictive behaviors. American Psychologist, 47, 1102-1114. doi:10.1037/0003-066X.47.9.1102

33.

Quinsey

V. L.

Rice

M. E.

Harris

G. T.

(1995). Actuarial prediction of sexual recidivism. Journal of Interpersonal Violence, 10, 85-105. doi:10.1177/088626095010001006

34.

Rice

M. E.

Harris

G. T.

(2005). Comparing effect sizes in follow-up studies: ROC area, Cohen’s d, and r. Law and Human Behavior, 29, 615-620. doi:10.1007/s10979-005-6832-7

35.

Rice

M. E.

Harris

G. T.

Cormier

C. A.

(1992). An evaluation of a maximum security therapeutic community for psychopaths and other mentally disordered offenders. Law and Human Behavior, 16, 399-412. doi:10.1007/BF02352266

36.

Salekin

R. T.

(2002). Psychopathy and therapeutic pessimism: Clinical lore or clinical reality? Clinical Psychology Review, 22, 79-112. doi:10.1016/S0272-7358(01)00083-6

37.

Sampson

R. J.

Laub

J. H.

(2003). Life-course desisters? Trajectories of crime among delinquent boys followed to age 70. Criminology, 41, 555-592. doi:10.1111/j.1745-9125.2003.tb00997.x

38.

Serin

R. C.

Amos

N. L.

(1995). The role of psychopathy in the assessment of dangerousness. International Journal of Law and Psychiatry, 18, 231-238. doi:10.1016/0160-2527(95)00008-6

39.

Simourd

D. J.

Hoge

R. D.

(2000). Criminal psychopathy: A risk-and-need perspective. Criminal Justice and Behavior, 27, 256-272. doi:10.1177/0093854800027002007

40.

Skeem

J. L.

Johansson

Andershed

Kerr

Eno Louden

(2007). Two subtypes of psychopathic violent offenders that parallel primary and secondary variants. Journal of Abnormal Psychology, 116, 395-409. doi:10.1037/0021-843X.116.2.395

41.

Suedfeld

Landon

P. B.

(1978). Approaches to treatment. In Hare

R. D.

Schalling

(Eds.), Psychopathic behavior: Approaches to research (pp. 347-376). Chichester, UK: John Wiley.

42.

Villeneuve

D. B.

Quinsey

V. L.

(1995). Predictors of general and violent recidivism among mentally disordered inmates. Criminal Justice and Behavior, 22, 397-410. doi:10.1177/0093854895022004004

43.

Vitacco

M. J.

Neumann

C. S.

Wodushek

(2008). Differential relationships between the dimensions of psychopathy and intelligence with adult jail inmates. Criminal Justice and Behavior, 35, 48-55. doi:10.1177/0093854807309806

44.

Walters

G. D.

Knight

R. A.

Grann

Dahle

K. P.

(2008). Incremental validity of the Psychopathy Checklist facet scores: Predicting release outcome in six samples. Journal of Abnormal Psychology, 117, 396-405. doi:10.1037/0021-843X.117.2.396

45.

Wintrup

Coles

Hart

Webster

C. D.

(1994). The predictive validity of the PCL-R in high-risk mentally disordered offenders [Abstract]. Canadian Psychology, 35, 47.

46.

Wong

Gordon

A. E.

(1999-2003). The Violence Risk Scale. Unpublished Manuscript.

47.

Wong

S. C. P.

(1984). The criminal and institutional behaviors of psychopaths (User Report No. 1984-87). Ontario, Ottawa: Ministry of the Solicitor General of Canada.

48.

Wong

S. C. P.

(1988). Is Hare’s Psychopathy Checklist reliable without the interview? Psychological Reports, 62, 931-934.

49.

Wong

S. C. P.

Burt

G. N.

(2007). The heterogeneity of incarcerated psychopaths: Differences in risk, need, recidivism and management approaches. In Hervé

Yuille

J. C.

(Eds.), The psychopath: Theory, research, and practice (pp. 461-484). New York, NY: Taylor & Francis.

50.

Wong

S. C. P.

Gordon

(2006). The validity and reliability of the Violence Risk Scale: A treatment-friendly violence risk assessment tool. Psychology, Public Policy, and Law, 12, 279-309. doi:10.1037/1076-8971.12.3.279

51.

Wong

S. C. P.

Gordon

Lewis

Olver

M. E.

(2012). The effectiveness of violence reduction treatment for psychopathic offenders: Empirical evidence and a treatment model. International Journal of Forensic Mental Health, 11, 336-349. doi:10.1080/14999013.2012.746760

52.

Wong

S. C. P.

Hare

R. D.

(2005). Program guidelines for the treatment of psychopathy. Toronto, Ontario, Canada: Multi-Health Systems.

53.

Woodworth

Porter

(2002). In cold blood: Characteristics of criminal homicide as a function of psychopathy. Journal of Abnormal Psychology, 111, 436-445. doi:10.1037//0021-843X.111.3.436

54.

Yang

Wong

S. C. P.

Coid

(2010). The efficacy of violence prediction: A meta-analytic comparison of nine risk assessment instruments. Psychological Bulletin, 136, 740-767. doi:10.1037/a0020473

55.

Zamble

Palmer

(1995). Prediction of recidivism using psychopathy and other meaningful variables. Issues in Criminological and Legal Psychology, 24, 153-156.