Subcomponents of Psychopathy Have Opposing Correlations With Punishment Judgments
Jana Schaich Borg
Stanford University and Duke University
Rachel E. Kahn
University of New Orleans
Walter Sinnott-Armstrong
Duke University
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Robert Kurzban
University of Pennsylvania
Paul H. Robinson
University of Pennsylvania Law School
Kent A. Kiehl
University of New Mexico and Lovelace Biomedical and Environmental Research Institute, Albuquerque, New Mexico
Psychopathy research is plagued by an enigma: Psychopaths reliably act immorally, but they also accurately report whether an action is morally wrong. The current study revealed that cooperative suppressor effects and conflicting subsets of personality traits within the construct of psychopathy might help explain this conundrum. Among a sample of adult male offenders (N 100) who ranked deserved punishment of crimes, Psychopathy Checklist–Revised (PCL–R) total scores were not linearly correlated with deserved punishment task performance. However, these null results masked significant opposing associations between task performance and factors of psychopathy: the PCL–R Interpersonal/Affective (i.e., manipulative and callous) factor was positively associated with task performance, while the PCL–R Social Deviance (i.e., impulsive and antisocial) factor was simultaneously negatively associated with task performance. These relationships were qualified by a significant interaction where the Interpersonal/ Affective traits were positively associated with task performance when Social Deviance traits were high, but Social Deviance traits were negatively associated with task performance when Interpersonal/ Affective traits were low. This interaction helped reveal a significant nonlinear relationship between PCL–R total scores and task performance such that individuals with very low or very high PCL–R total scores performed better than those with middle-range PCL–R total scores. These results may explain the enigma of why individuals with very high psychopathic traits, but not other groups of antisocial individuals, usually have normal moral judgment in laboratory settings, but still behave immorally, especially in contexts where Social Deviance traits have strong influence. Keywords: psychopathy, moral judgment, Deserved Punishment Test
Psychopaths are infamous for their immoral behavior. Still unclear, however, is whether psychopaths’ behavior is due to personality traits that interfere with them knowing what is morally wrong or, instead, personality traits that interfere with them from caring about what is morally wrong. This unresolved issue has long been at the center of scientific debates about the neural mechanisms of moral judgment and how to treat psychopaths (Glenn, 2010), as well as legal debates about whether psychopaths
should be considered criminally (or morally) responsible for their crimes (McMillan & Malatesti, 2010).
Psychopathy
Psychopathy was originally described as a constellation of interpersonal, affective, and behavioral personality traits that included a superficial and manipulative interpersonal style, a pro-
This article was published Online First July 8, 2013. Jana Schaich Borg, Institute for Neuro-Innovation and Translational Neurosciences, Stanford University, and Department of Psychiatry, Duke University; Rachel E. Kahn, Department of Psychology, University of New Orleans; Walter Sinnott-Armstrong, Department of Philosophy and Kenan Institute for Ethics, Duke University; Robert Kurzban, Department of Psychology, University of Pennsylvania; Paul H. Robinson, University of Pennsylvania Law School; Kent A. Kiehl, Departments of Psychology and Neuroscience, University of New Mexico, and the Mind Research Network (MRN), an affiliate of Lovelace Biomedical and Environmental Research Institute (LBERI), Albuquerque, New Mexico. 667
This research was supported in part by National Institute on Drug Abuse Grant 1R01DA026505-01A1 (Principal investigator: Kent A. Kiehl). We are grateful to the staff and inmates of the New Mexico Corrections Department for their support and assistance in making this research possible. Correspondence concerning this article should be addressed to Jana Schaich Borg, who is now at the Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, Bryan Research Building Room 361, 311 Research Drive, Box 3209, Durham, NC 27710. E-mail: borg@neuro.duke.edu
668
SCHAICH BORG ET AL.
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
found lack of empathy or remorse, frequent impulsivity and irresponsibility, and socially deviant behavior or “antisociality” (Cleckley, 1941, 1976). While some psychopathic traits are similar to the criteria for antisocial personality disorder (ASPD) contained in the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM–IV; American Psychiatric Association, 1994), psychopathy and ASPD differ in at least two important ways. First, adults with psychopathic traits represent a minority or distinct subgroup of individuals with ASPD (Hare, Hart, & Harpur, 1991; Hart & Hare, 1997) who have a particularly severe and chronic pattern of antisocial behavior (Hart, Knopp, & Hare, 1988; Hemphill, Hare, & Wong, 1998; Kosson, Smith, & Newman, 1990). Second, and perhaps more important, ASPD criteria do not include the constellation of interpersonal or affective traits that characterize psychopathy (except lack of remorse or guilt), and instead focus primarily on the antisocial behavioral traits. For example, criteria for ASPD include traits of impulsivity, irresponsibility, and a lifelong pattern of violating the rights and norms of others (indicated by failure to conform to laws and repeated fighting or assaults) but do not include traits such as having a superficial and manipulative interpersonal style (American Psychiatric Association, 1994). As a consequence of these differences, approximately 75% of male prison inmates would qualify for a diagnosis of ASPD, but only 30% of those would also qualify for a diagnosis of psychopathy according to the Hare Psychopathy Checklist–Revised (PCL–R; Hare, 1991; 2003; Hart & Hare, 1996; Meloy, 1988).
Assessment of Psychopathy
Psychopathy is often clinically assessed in incarcerated adults with the PCL–R (Hare, 1991, 2003). The PCL–R is an expert-rated scale informed by extensive collateral reports and a semistructured interview. Using comprehensive descriptions provided in the manual, individuals are rated on 20 specific trait items such as superficial charm, pathological lying, lack of remorse, impulsivity, irresponsibility, and criminal versatility. These item scores are summed to create PCL–R total scores that range from 0 to 40. In the PCL–R manual, the reported mean PCL–R total scores in large samples of participants from North American prison and forensic psychiatric samples are 23.6 (SD 7.9) and 20.6 (SD 7.8), respectively (Hare, 1991). Empirical research suggests PCL–R total scores are dimensional (Edens, Marcus, Lilienfeld, & Poythress, 2006; Guay, Ruscio, Knight, & Hare, 2007; Walters, Duncan, & Mitchell-Perez, 2007; Walters, Gray, et al., 2007), but categorical cutoffs are used in practice to infer recidivism risk levels (Hare et al., 1991; see also Harris, Rice, & Quinsey, 1994). When categorical cutoffs are used, a psychopath is traditionally defined in North American participants as someone who receives a total PCL–R score of 30 or above (Hare, 1991, 2003). The PCL–R was originally developed for use with male offenders and forensic patients but has become widely used in populations of female offenders, sex offenders, and substance abusers as well (for a review see Hare, Clark, Grann, & Thornton, 2000).
Psychopathy and Moral Judgment
In this article, moral judgment will be used to refer to the event of judging some act, institution, or person as morally wrong or right. A moral verdict is the outcome of a moral judgment or the
commitment to one moral conclusion (Schaich Borg, SinnottArmstrong, Calhoun, & Kiehl, 2011). Most of the well-established psychological research relevant to moral judgment assesses either the quality of reasoning that leads up to a moral verdict (e.g., Colby & Kohlberg, 1987) or the development of the ability to differentiate moral rules from conventional rules (Turiel, 1983). More recently, attempts have been made to study moral verdicts as well (e.g., Cushman, Young, & Hauser, 2006), but these tests cannot be validated in the same way as more traditional psychological tests because there is no universally accepted criteria for identifying what is morally wrong across individuals and cultures and, thus, no universally accepted criteria for assessing the quality of moral verdicts. Therefore, at present, moral verdict tests are usually evaluated by their ability to elicit consistent moral conclusions across people and populations. To date, there is little evidence that psychopaths as defined by total scores on the PCL–R have deficits in either the processes leading up to a moral verdict or the ability to produce normal moral verdicts. The first moral judgment assessments were based on Lawrence Kohlberg’s moral development theory. To Kohlberg, the most important aspect of moral development was the quality of one’s moral justifications, not the outcome of one’s moral verdicts. Kohlberg observed that humans progress sequentially through six stages of moral reasoning in which concepts such as consequences to oneself, maintaining social order, or maintaining universal rights for humanity are preferentially used to justify one’s moral verdicts. The Moral Judgment Interview (Colby & Kohlberg, 1987) and the Defining Issues Test (Rest, Cooper, Coder, Masanz, & Anderson, 1974) were developed to assess the sophistication of participants’ moral justifications according to this developmental scheme. Psychopaths’ moral reasoning as measured by both of these moral development tests is normal (or potentially even above average; Link, Scherer, & Byrne, 1977; Lose, 1997; O’Kane, Fawcett, & Blackburn, 1996). A second class of moral judgment assessments were later developed based on the observation that children as young as age 2 years 6 months can differentiate between conventional wrongs (acts judged to be wrong because societal rules say they are wrong, such as wearing one’s pants inside out) and moral wrongs (acts judged to be wrong regardless of societal rules, such as hitting someone; Smetana, 1981). Variations of moral versus conventional tests were designed to assess whether children could distinguish between these types of transgressions, and using one of these variations, Blair and colleagues (Blair, 1995; Blair, Jones, Clark, & Smith, 1995) reported that individuals with psychopathy do not distinguish between moral versus conventional transgressions. However, in both reports, psychopaths’ performance was characterized by inflation of their ratings of conventional wrongs, not impairment in their ratings of moral wrongs. Further, it was subsequently shown that psychopaths can differentiate between moral and conventional norms as well as incarcerated nonpsychopathic controls if doing so is clearly congruent with the experimenter’s expectations (Aharoni, Sinnott-Armstrong, & Kiehl, 2011). This finding suggests that psychopaths’ abnormal performance in earlier studies was a result of their efforts to look “more moral” to the experimenter, not an inability to distinguish between moral and conventional transgressions. Most recently, tests have been developed to examine people’s moral verdicts. The most common way to assess moral verdicts is to
PSYCHOPATHY AND PUNISHMENT JUDGMENTS
669
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
ask people how they would resolve philosophical moral thought scenarios (Cushman et al., 2006). Thus far, most reports have indicated that psychopaths’ verdicts of how to resolve philosophical moral thought scenarios are not different from those of nonpsychopaths (Cima, Tonnaer, & Hauser, 2010; Glenn, Raine, Schug, 2009; Glenn, Raine, Schug, Young, & Hauser, 2009). This suggests that psychopaths can successfully report intuitions consistent with common moral principles. In addition, in a simpler method of assessing moral verdicts, psychopaths’ ratings of the severity of moral violations depicted in pictures are equally severe as ratings provided by nonpsychopaths (Harenski, Harenski, Shane, & Kiehl, 2010). Psychopathy has been shown to be negatively associated with valuing certain types of moral virtues (Aharoni, Antonenko, & Kiehl, 2011), but it has not yet been demonstrated that this negative correlation translates into different patterns of moral verdicts. Therefore, overall there is little evidence supporting the hypothesis that psychopaths have deficits in moral judgment during laboratory tasks in controlled environments. This lack of correlation between psychopathy and moral judgment impairment has proven puzzling for researchers in medical, psychological, or legal fields trying to understand why psychopaths still consistently perform immoral acts outside experimental contexts.
the DPT was chosen as the moral judgment assessment for the present study. We hypothesized that if psychopaths have impaired moral punishment judgments, PCL–R scores should be negatively correlated with measures indexing the similarity between individuals’ DPT scenario rank orderings and previously agreed-upon DPT scenario rank orderings.
Subcomponents of Psychopathy May Correlate With External Variables in Opposing Directions
When examining the association between psychopathy and moral judgment, it is important to consider whether the constellation of traits that form the construct of psychopathy have the potential to be differentially associated with moral judgment. Psychopathic personality traits measured by the PCL–R can be meaningfully divided into an Interpersonal/Affective factor (Factor 1) and a Social Deviance factor (Factor 2; Hare, 2003; Harpur, Hare, & Hakstian, 1989). The Interpersonal/Affective factor includes traits corresponding with psychopaths’ social dominance, manipulation, and efficacy (as well as separate traits associated with lack of general affect, including empathy). The Social Deviance factor includes traits corresponding to psychopaths’ impulsivity, irresponsibility, and antisocial behavior. The Interpersonal/Affective and Social Deviance factors of the PCL–R are modestly correlated (Hare, 2003), so when their shared variance correlates with personality traits or behavior in the same way (either in the same direction or not at all), the associations between the unique aspects of the PCL–R factors and variables of interest are often obscured or “suppressed” unless their shared variance is removed through regression or partial correlation. This phenomenon, called cooperative suppression, is well documented (Paulhus, Robins, Trzesniewski, & Tracy, 2004) and has previously been shown to conceal relationships between PCL–R factor scores and other personality measures (Blonigen et al., 2010; Hicks & Patrick, 2006). Technically, cooperative suppression in PCL–R factors can be identified as cases when the sum of partial regression coefficients between PCL–R factors and an external variable are greater than the sum of the simple, or zero-order, regression coefficients between each PCL–R factor in isolation and the same external variable (Cohen & Cohen, 1975). When cooperative suppression effects are taken into account, PCL–R factor scores have been shown to correlate with other types of personality traits or behaviors in opposite directions. For example, the Interpersonal/Affective and Social Deviance factors of the PCL–R have been shown to have mutually repulsive effects in the prediction of self-reports of emotional distress and fearfulness (Hicks & Patrick, 2006; Patrick, 1994), anger and hostility (Hicks & Patrick, 2006; Reidy, Shelley-Tremblay, & Lilienfeld, 2011), positive affect (Patrick, 1994), achievement (Verona, Patrick, & Joiner, 2001), stress reactivity (Verona et al., 2001), interpersonal aggression (Kennealy, Hicks, & Patrick, 2007), and attempted suicide history (Verona, Hicks, & Patrick, 2005). Thus, the Interpersonal/Affective and Social Deviance factors of the PCL–R often have contrasting associations with external criteria, despite their shared ability to represent parts of the overall construct of psychopathy, but these effects are not detectable unless statistical models take their shared variance into account. These cooperative suppression effects could potentially explain why previous studies have reported difficulty in identifying relationships between psychopathy and moral judgment. The unique aspects of the
Deserved Punishment Test
One kind of moral judgment that has not been assessed in psychopaths is judgment about the amount of punishment deserved by those committing differing types of wrongdoing. Robinson and Kurzban’s (2007) Deserved Punishment Test (DPT) was designed to assess whether society has shared intuitions of justice that lead people to agree on an ordinal ranking of offenders according to their blameworthiness. The DPT asks participants to read 24 short scenarios describing a situation where “John” performs an act that may or may not be considered a criminal offense (such as accidentally taking somebody else’s umbrella, stealing a microwave from somebody’s house, or intentionally shooting someone) and rank order them according to how much punishment, if any, John deserved for each act. This task was intentionally designed to be a ranking task because ratings of the relative seriousness of crimes are more consistent across cultures than the ratings of the absolute seriousness of crimes (Robinson & Kurzban, 2007). Robinson & Kurzban (2007) demonstrated that participants from community samples have “almost perfect” agreement (Landis & Koch, 1977) about what the “normal” ranking of punishment deserved by each scenario should be in the DPT. This agreement is so strong that the results of the DPT have been used to argue that shared intuitions of justice could be useful for guiding policy and legal theory relevant to legal sentencing (Carlsmith & Darley, 2008; Robinson & Darley, 2007). The DPT has three critical characteristics that make it well suited to study the relationship between psychopathy and moral judgment: (a) It asks for judgments that have no clear correct or incorrect answers, minimizing the potentially dominating influence of impression management, (b) it allows for a wide range of performance, and (c) performance can be indexed by a continuous variable representing how similar a participant’s ranking order is to the modal ranking order, rather than a constrained interval or ordinal variable. These characteristics make the test both more resilient to intentional manipulation and more sensitive to subtle differences than moral judgment tests previously used in incarcerated populations. For these reasons,
670
SCHAICH BORG ET AL.
two PCL–R factors may have opposing associations with moral judgment that cancel each other out when PCL–R total scores are used to assess psychopathy.
Current Study
In the current study, we assessed whether the Interpersonal/ Affective factor (Factor 1) and the Social Deviance factor (Factor 2) of psychopathy have common or distinct correlations with moral judgment as measured by DPT performance. To do this, we implemented regression models that allowed us to investigate whether the PCL–R factors have contrasting associations with moral judgment that are masked when their shared variance is not taken into account. If true, the opposing relationships between PCL–R factors and moral judgment could cancel each other out when both sets of traits are present in high-scoring psychopaths, explaining why their moral judgment appears normal. To maximize our chances of being able to detect conflicting correlations between subsets of personality traits, we administered the DPT to a large sample of incarcerated individuals with a wide range of PCL–R scores, including many who scored above the clinical criteria for psychopathy. In addition, previous research has shown that anxiety can interact with psychopathy (Lilienfeld & Penna, 2001) and that individuals with high levels of psychopathic traits may manipulate their answers to impress experimenters (Blair, 1995; Book, Holden, Starzyk, Wasylkiw, & Edwards, 2006; Wilson & McCarthy, 2011). Therefore, we also assessed whether trait anxiety or impression management influenced the association between PCL–R and DPT performance. Overall we predicted that (a) PCL–R total scores would not correlate with DPT scenario ratings, consistent with previous studies reporting a lack of relationship between psychopathy and moral judgment; (b) the variance in DPT performance accounted for by the addition of both PCL–R factor scores simultaneously in a regression model would be more than expected if the variance accounted for by each factor in isolation was simply added (indicating a cooperative suppression effect); and (c) when both factor scores were entered simultaneously in a regression model, PCL–R factor scores would correlate with DPT scenario ratings in opposing directions. If these predictions were confirmed, the results of the present study could help explain the lack of relationship between psychopathy and moral judgment reported in the past and could help reconcile intuitions that psychopaths’ immoral behavior should somehow be related to impairments in moral judgment.
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
level of education, 22.3% had no high school degree, 40.4% completed high school or obtained a General Educational Development (GED) degree, 25.5% had some college education, 8.5% had a 2-year college degree, 1.1% had a 4-year college degree, and 2.2% had a graduate degree. Participants were excluded if they were older than 60 years; had a history of seizures, epilepsy, or psychosis as assessed by the Structured Clinical Interview for DSM–IV (First, Spitzer, Gibbon, & Williams, 1997); had experienced a clinical loss of consciousness exceeding 30 min; or had other major medical problems. These exclusions were made as part of SWANC, in accordance with standard clinical neuroimaging protocols. In addition, all participants were required to show fluency in English at or above a fourth-grade reading level and have an estimated IQ (see description of IQ measure in the Materials section) of at least 70 (M 98.3, SD 14.4). Consistent with the average wage in the correctional system, participants were paid $1.00/hr for their time. The study was approved by the Human Research Review Committee at the University of New Mexico, and all participants provided written informed consent.
Materials
Moral judgment. The deserved punishment test (DPT; Robinson & Kurzban, 2007) consists of 24 short scenarios describing a situation in which “John” performs an act that may or may not be considered a criminal offense (see Appendix A for transcription of each scenario). Robinson & Kurzban (2007) collected two separate data sets in community population samples to help illustrate how consistent relative assessments of the punishment deserved for crimes can be. One sample of 64 participants was given the test in person and another sample of 246 participants was given the test online. Both samples gave highly consistent rankings of the punishments deserved in the 24 DPT scenarios (Robinson & Kurzban, 2007, summarized in Table 1 of the present study). Participants only reversed the modal pairwise rankings of the DPT scenarios 4% and 8.2% of the time, respectively (Robinson & Kurzban, 2007). Further, the Kendall’s coefficient of concordance—a statistic that measures agreement among rank orderings of the same set of stimuli (across the entire ranking, not just pairwise comparisons)—was 0.95 for the cohort of 64 participants and 0.88 for the cohort of 246 participants, where a coefficient of 1.0 indicates perfect agreement and 0.0 indicates no agreement (Robinson & Kurzban, 2007). This level of agreement was similar across the different demographic groups represented in the larger sample of 246 participants.1
1 The demographics of Robinson and Kurzban’s (2007) nonincarcerated populations were as follows: (a) University population (N 64): gender: 36% male, 64% female; race: 91% White, 9% non-White; age: 2% age 20 years or younger, 11% ages 21–30, 5% ages 31– 40, 11% ages 41–50, 22% ages 51– 60, 47% ages 61 or older; highest level of education: 5% some college, 2% 2-year college degree, 38% 4-year college degree, 39% master’s degree, 17% doctorate or professional degree; and (b) Internet population (N 246): gender: male 50%, female 50%; race: 6% Hispanic, 70% White, 1% Native American, 5% mixed, and 4% other; age: 6% age 20 years or younger, 15% ages 21–30, 21% ages 31– 40, 18% ages 41–50, 17% ages 51– 60, 23% ages 61 or older; highest level of education: 4% no high school degree, 18% high school graduate, 47% some college, 7% 2-year college degree, 14% 4-year college degree, and 10% graduate degree.
Method Participants
One hundred adult male inmates were recruited from the Western New Mexico Correctional Facility, a minimum–medium security institution in Grants, New Mexico, as part of the larger Southwest Advanced Neuroimaging Cohort–Adult sample (SWANC–A) from 2007 to 2010. Participants’ ages ranged from 18 to 56 years (M 33.7; SD 9.4). Ethnic/racial composition was 52% Hispanic or Latino, 36% White, 7% Black or African American, 3% American Indian or Alaska Native, and 2% of other or mixed ethnicity. Participants were classified based on their self-classification into these categories. When asked their highest
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Note. Bold entries marked with an asterisk indicate notable differences. S
Replicating the procedures implemented with Robinson and Kurzban’s population sample given the test in person, we pasted each scenario onto one index card for a total of 24 cards. Participants were asked to read each scenario and rank order the cards to reflect the amount of punishment, if any, John deserved for each act. Once completed, participants were asked to review and compare each scenario to the one next to it and confirm their answers. The approved rank orderings were then recorded. Community participants have “almost perfect” agreement (Landis & Koch, 1977) about what the “normal” ranking of punishment deserved by each scenario should be (Robinson & Kurzban, 2007; provided for comparison in Table 1). The rankings of the present incarcerated population were compared with this previously established “normal” ranking using the statistical tests described in detail later. (For further details about the theory behind the deserved punishment test, see Robinson & Kurzban, 2007.) In the present study, one inmate scored over three standard deviations below the mean on the DPT task and was suspected of not paying attention to the task and was therefore excluded from all analyses. Psychopathy. All participants were evaluated for psychopathy using the PCL–R (Hare, 1991, 2003). This assessment includes a semistructured interview that provides in-depth information about participants’ school, family, work, and criminal history, as well as their interpersonal and emotional skills. All participants signed a release form allowing the researcher to review their institutional files (criminal history, social worker reports, family history, education, employment, and so forth), which was used as sources of collateral information for scoring the PCL–R. The
PCL–R consists of 20 items, each of which is scored on a 3-point scale: 0 (doesn’t apply), 1 (applies somewhat), and 2 (definitely applies) to the individual. Total scores range from 0 to 40 (Hare, 1991). In the present population, PCL–R total scores ranged from 3.2–36.8 (SD 7.8), with 25 participants scoring a 30 or above. Double ratings were conducted on approximately 17% of the sample, and reliability was high: interclass correlation coefficient (ICC) .96 for total scores, .92 for the Interpersonal/Affective factor, and .97 for the Social Deviance factor. General intelligence. A two-subtest form of the Wechsler Adult Intelligence Scale (WAIS–III; Wechsler, 1997) was used to collect a full-scale IQ estimate. The two subtests, Vocabulary and Matrix Reasoning, have been shown to provide a reliable estimate of full-scale IQ (Ryan, Lopez, & Werth, 1999). One individual did not complete the IQ assessment. Anxiety. Trait anxiety was assessed using the Spielberger State–Trait Anxiety Scale (STAI; Spielberger, Gorsuch, & Lushene, 1970). The STAI is a 40-item self-report measure containing a 20-item section that assesses trait anxiety (how much anxiety a person generally feels). Participants rate each statement on a 4-point scale ranging from almost never to almost always (e.g., “I feel nervous and restless” or “I am ‘cool, calm, and collected’” [reverse-scored]). Within our sample, reliability for the trait anxiety scale of the STAI was Cronbach’s alpha .93. Five individuals did not complete the STAI, and an additional four had missing or incomplete data (total missing, n 9). Social desirability and impression management. The Balanced Inventory of Desired Responding (BIDR– 6; Paulhus, 1991)
672
SCHAICH BORG ET AL.
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
is a 40-item self-report scale designed to assess whether individuals are likely to distort their answers on tests or questionnaires in an effort to portray themselves positively. The BIDR is composed of two scales: a Self-Deceptive Enhancement (SDE) scale and an Impression Management (IM) scale, each composed of 20 items (e.g., “I never swear” or “I always obey laws, even if I’m unlikely to get caught”). For the purposes of the present study, only the IM scale was used. The IM scale assesses the extent to which an individual deliberately tries to present himself or herself in a favorable light to others. All items on the BIDR are rated on a 7-point scale (1 not true, 4 somewhat true, and 7 very true). Higher scores on the BIDR suggest an individual has greater socially desirable responding, or is attempting to present himself or herself in a more positive light. In past research, the BIDR has been shown to successfully measure socially desirable responding in offender populations similar to the one used in the present study (Kroner & Weekes, 1996). In addition, internal consistency for the two scales has been adequate ( .80) in both general and forensic samples (Kroner & Weekes, 1996; Paulhus, 1991). In our sample, this scale demonstrated adequate internal consistency as well (Cronbach’s .78). Two individuals did not complete the BIDR, and IM scores were not calculated for an additional three participants due to missing or incomplete data (total missing, n 5). Demographic variables. All participants completed a demographic questionnaire that included questions regarding ethnicity, age, education level, income level prior to incarceration, number of years incarcerated, number of times incarcerated, and prison security level. Education level was coded using a 7-point scale ranging from 1 (some high school) to 3 (some college) to 7 (doctoral/professional degree). Similarly, income level was coded on a 6-point scale ranging from 1 (less than $20,000/year) to 6 (more than $100,000/year). Age, number of years incarcerated, and number of times incarcerated were coded as continuous variables. Prison security level ranged from Level 2 (minimum security) to Level 4 (medium security).
tistic used to assess agreement among raters, was used to assess agreement in DPT scenario rankings across participants. A Kendall’s tau statistic was calculated for each participant to measure the similarity (or distance) between each participant’s rank ordering and the “normal” or modal rank order agreed upon by the nonincarcerated university population from Robinson and Kurzban, 2007 (provided in Table 1). The nonincarcerated university population was chosen as the reference population because, like the present population, they completed the task with physical cards in person, whereas Robinson and Kurzban’s second community population completed the task over the Internet.2 The Kendall’s tau statistic is a metric that counts the number of pairwise disagreements between two lists and gives a value from 1 to 1, where values of 1 represent a perfect negative relationship between the two lists (or complete disagreement), values of 1 represent a perfect positive relationship between the two lists (or complete agreement), and values of 0 represent no relationship at all (Lapata, 2006). Therefore, if a participant ranked the DPT scenarios in the exact same order as Robinson and Kurzban’s nonincarcerated university population, that participant would have a Kendall’s tau value of 1. If a participant ranked the DPT scenarios in the exact opposite order as Robinson and Kurzban’s nonincarcerated university population, that participant would have a Kendall’s tau value of 1. If there was no relationship between a participant’s DPT scenario rankings and the rankings provided by Robinson and Kurzban’s university nonincarcerated population, that participant would have a Kendall’s tau value of 0. We conducted hierarchical regression analyses to test the effects of psychopathy on DPT performance using Kendall’s tau as the dependent variable. In all of these analyses, Kendall’s tau was logit transformed (log[x/1 x]) to improve heteroscedasticity.3 Finally, significant interactions between PCL–R factors were analyzed using the regression equation from the full sample to calculate the predicted tau values if Factors 1 and 2 were high (1 SD above the mean) or low (1 SD below the mean; Holmbeck, 2002). The slopes of the simple effects from these post hoc analyses were then used to assess statistical significance.
Procedure
This study was conducted as part of a larger neuroimaging research project that began in 2007. In total, the participants completed 8 –15 hr of research over the course of five to six visits. The assessments pertinent to the present research took between 3– 4 hr, and participants were paid for their time as they progressed through the research study. As recommended by the institutional review board, participants were told they could skip any questions that they did not feel comfortable answering.
Results Contribution of Demographic Variables
To test the contribution of demographic and relevant cognitive variables to PCL–R total and factor scores and determine whether they needed to be considered in analyses investigating the rela2 The modal ranking of the university population reversed the order of one set of scenarios compared with the Internet population, indicated by an asterisk and bold font in Table 1, but the basic results of the present study do not change if the modal rank provided by Robinson and Kurzban’s Internet population is used as the reference rank order instead of the modal rank order provided by Robinson and Kurzban’s university population. 3 The logit transform was chosen because it is the standard transformation for proportional variables bounded by 0 and 1 (Lesaffre, Rizopoulos, & Tsonaka, 2007). The transformation did reduce heteroscedasticity in the Kendall’s tau distribution and was therefore used for analyses reported in the main text, but as a cautionary measure, all the analyses were repeated on nontransformed values of Kendall’s tau. The results of these models were comparable to the results discussed in the main text and are provided in Appendix B.
Data Analysis
We analyzed data with SPSS Version 19. Pearson correlations and analyses of variance (ANOVAs) were conducted to assess the association between PCL–R scores and other potentially confounding variables. The DPT was intentionally designed to assess relative (as opposed to absolute) judgments of moral punishment, and performance on the DPT is defined by the order in which participants rank the 24 DPT scenarios. Therefore, DPT performance was assessed using statistical tests comparing ordinal rankings. Kendall’s coefficient of concordance, a nonparametric sta-
PSYCHOPATHY AND PUNISHMENT JUDGMENTS
673
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
tionship between psychopathy and moral judgment, we computed two-tailed Pearson correlations between all PCL–R scores (total, Interpersonal/Affective factor, Social Deviance factor, and all four PCL–R facet scores), all demographic variables (age, IQ, education level, income level, number of years incarcerated, number of times incarcerated, and prison security level), and task performance (see Table 2). Age was negatively correlated with PCL–R total (r .22, p .03) and Social Deviance factor scores (r .30, p .01), but not significantly associated with Interpersonal/ Affective factor scores. This is consistent with previous literature showing age-related changes in psychopathic traits are usually explained by a decline in Social Deviance traits (such as impulsivity and antisocial behavior) rather than changes in Interpersonal/ Affective traits (such as callousness or superficial charm), which tend to remain stable across age spans (Hare, 2003; Harpur & Hare, 1994). To assess potential effects of ethnicity, we compared PCL–R scores for the two largest ethnic groups in our sample (White, n 36; Hispanic, n 52) using ANOVAs. Sample sizes for the other ethnic groups were too small for statistical comparison. No significant differences in PCL–R total (F(1, 86) 0.27, p .61) or factor scores (Interpersonal/Affective: F(1, 85) 0.01, p .93; Social Deviance: F(1, 86) 1.10, p .30) were found between ethnic groups. Associations between all demographic variables and Kendall’s tau, the measurement of how similar a participant’s rankings are to the rankings provided by nonincarcerated populations (see Methods), are also reported in Table 2. Kendall’s tau was positively correlated with age, IQ, and education level (all p .05). In previous studies of moral verdicts, the relationship of age, IQ, and education level to task performance has not been examined, but the present results are Table 2 Descriptive Statistics and Correlations Between Study Variables
Variable 1. PCL-R Total PCL-R factors 2. Interpersonal/Affective 3. Social Deviance PCL-R facets 4. Interpersonal 5. Affective 6. Lifestyle 7. Antisocial Control variables 8. Age 9. IQ 10. Education level 11. Income level 12. Years incarcerated 13. No. of times incarcerated 14. Prison security level 15. BIDR–IM Scale 16. STAI Trait Anxiety Dependent variable 17. Kendall’s tau 1 2 3 4 5 6 7
consistent with previous reports that all three variables correlate positively with moral reasoning as defined by Kohlberg’s moral development stages (Freeman & Giebink, 1979; Gibbs, Basinger, Grime, & Snarey, 2007). In addition, Kendall’s tau was positively correlated with years of incarceration (r .35, p .05), even when age was taken into account (r .26, p .05), but not number of times in prison (p .11). An ANOVA showed that White participants had significantly higher Kendall’s tau values on the DPT than Hispanic participants, F(1, 89) 8.95, p .01. Because age was the only variable significantly correlated with both psychopathy scores and Kendall’s tau, age was included as a covariate in all subsequent regression models using either PCL–R total or Social Deviance factor scores as independent variables and Kendall’s tau as the dependent variable by entering age in the first block of each hierarchical regression. Regression models including IQ and education levels as covariates were also tested, but adding these independent variables to the models did not affect the statistical relationship between psychopathy variables and task performance (Appendix C), so they were not used as primary analyses.
Validation of the Deserved Punishment Test in an Incarcerated Sample
We validated the use of the DPT in our incarcerated population by comparing three characteristics of DPT performance in our population with that of nonincarcerated populations: (a) the modal rankings of each DPT scenario, (b) the amount of agreement across participants about how the DPT scenarios should be ranked (indicated by Kendall’s coefficient of concordance), and (c) the distribution of Kendall’s tau values from individual participants.
Note. N 100. Kendall’s tau was inverse logit transformed before the reported correlations were calculated. The mean and standard deviation of Kendall’s tau in the table reflects the logit-transformed variable; as reported in the text; the mean and standard deviation of the nontransformed variable was 0.81 (SD 0.11), with a range from 0.47– 0.99. PCL–R Psychopathy Checklist–Revised; BIDR–IM Balanced Inventory of Desired Responding–Impression Management; STAI Spielberger State–Trait Anxiety Inventory. Correlations in bold significant at p .05. † p .08.
674
SCHAICH BORG ET AL.
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
The modal rankings provided for how much punishment each scenario deserved were almost identical between the present incarcerated population and the previously described nonincarcerated populations (Table 1). Minor differences in rankings are indicated by bold type and an asterisk in Table 1. The consistency of modal rankings across the three populations indicates that they are a reasonable reference with which to compare individuals’ moral punishment judgments. Kendall’s coefficient of concordance was used to statistically assess how much participants agreed in their rankings of how much punishment each of the 24 items in the DPT deserved. Our incarcerated sample presented very high agreement in their rankings of deserved punishment, with a Kendall’s W of .86 overall (p .001). Similarly, Robinson and Kurzban’s university population of 64 participants provided rankings with a Kendall’s W of .95 and Robinson and Kurzban’s Internet-based population of 246 participants provided rankings with a Kendall’s W of .88 (Robinson & Kurzban, 2007). This means that like nonincarcerated participants, incarcerated participants have “almost perfect” agreement (Landis & Koch, 1977) about what the “normal” ranking of punishment deserved in each scenario should be in the DPT. In other words, all three populations of incarcerated or nonincarcerated participants have shared intuitions of how offenders should be ranked according to their blameworthiness. We also calculated a Kendall’s tau statistic for each participant to measure how similar each participant’s rank ordering was to the modal ranking provided by nonincarcerated populations. The mean Kendall’s tau value from the present incarcerated population was 0.81 (SD 0.11), with a range from 0.47– 0.99 and a distribution that was positively skewed. The mean Kendall’s tau value from the previously published nonincarcerated university population was 0.93 (SD 0.06) with a range from 0.69 –1.00, and the Kendall’s tau value from the previously published nonincarcerated Internet population was 0.85 (SD 0.11) with a range from 0.42– 0.99. Both of these samples were also positively skewed (Appendix D). These comparable Kendall’s tau distributions indicate that the current sample of incarcerated adults provided a distribution of punishment rankings similar to those provided by nonincarcerated adults. Overall, the results of these three metrics from our incarcerated population support the use of the DPT as an assessment of robust shared intuitions of justice that can be used to probe the relationship between psychopathy and moral punishment judgments.
anxiety was not considered to be a confounding variable and was not included as a covariate in subsequent regression analyses.
Ruling Out Potentially Confounding Influences of Impression Management
To confirm that DPT performance was not sensitive to impression management, we calculated the two-tailed Pearson correlation between participants’ Kendall’s tau values and their scores on the BIDR IM subscale. There was no significant correlation between BIDR IM subscale scores (p .27) and Kendall’s tau. Therefore, impression management did not correlate with moral judgment and was not included as a covariate in subsequent regression analyses.
PCL–R Total Scores and DPT Performance
To test the effects of psychopathy on DPT performance, we implemented a hierarchical regression with Kendall’s tau as the dependent variable, age in the first block (because it was correlated with both PCL–R total scores and Kendall’s tau as described previously in the Contribution of Demographic Variables section),4 and PCL–R total scores in the second block. As summarized in Table 3, PCL–R total scores did not account for a significant amount of variance in DPT performance (p .84). The R2 change due to the addition of PCL–R total scores was 0. Thus, as predicted, when PCL–R total scores are used to index psychopathy, there was no significant linear association between psychopathy and moral judgment.
Isolated PCL–R Factor Scores and DPT Performance
To test the effects of each factor score in isolation on DPT performance, we implemented a hierarchical regression with Kendall’s tau as the dependent variable, age in the first block, and either the Interpersonal/Affective PCL–R factor (Factor 1) or the Social Deviance PCL–R factor (Factor 2) scores in the second block. Consistent with the zero-order correlations provided in Table 2, neither the Interpersonal/Affective PCL–R factor (p .19) nor the Social Deviance PCL–R factor (p .25) scores individually accounted for a significant amount of variance. The R2 change due to the addition of Interpersonal/Affective scores or the Social Deviance scores was 0.02 or 0.01, respectively. Thus, compatible with our predictions, when PCL–R factor scores are used in isolation to index psychopathy, psychopathy is not associated with moral judgment.
Ruling Out the Confounding Influences of Anxiety
To ensure that anxiety could not explain or alter correlations between PCL–R total scores and DPT performance, we examined the interaction among trait anxiety using the STAI (Spielberger et al., 1970), psychopathy, and DPT performance. We performed a hierarchical regression with each participant’s Kendall’s tau value as the dependent variable, age entered in Step 1 (because it was correlated with both PCL–R total scores and Kendall’s tau as described previously in the Contribution of Demographic Variables section),4 total PCL–R scores and STAI scores entered in Step 2, and a Psychopathy anxiety interaction term (PCL–R scores STAI scores) entered in Step 3. There were no significant interactions between psychopathy and trait anxiety for DPT performance, R2 .01, F(1, 86) 0.57, p .45, b 0.08. Thus,
Combined PCL–R Factor Scores and DPT Performance
To test whether cooperative suppression could reveal opposing relationships between PCL–R factors and DPT performance that would not be observable when either PCL–R factor score was assessed in isolation, we implemented a hierarchical regression with Kendall’s tau as the dependent variable, age in the first block, and both the PCL–R Interpersonal/Affective factor (Factor 1) and
4 As a cautionary measure, all the analyses reported here were repeated without age included in the model. The results of the models without age were comparable to the results discussed in the main text.
PSYCHOPATHY AND PUNISHMENT JUDGMENTS
675
Table 3 Regression Analyses With Psychopathy Checklist–Revised Total and Factor Scores as Predictors of Deserved Punishment Test Performance (as Measured by Logit-Transformed Kendall’s Tau)
Variable Psychopathy total scores Step 1 Age Step 2 PCL–R Total Psychopathy factor scores (entered independently) Step 1 Age Step 2 PCL–R Interpersonal/Affective Factor Step 1 Age Step 2 PCL–R Social Deviance Factor Psychopathy factor scores (entered simultaneously) Step 1 Age Step 2 PCL–R Interpersonal/Affective Factor PCL–R Social Deviance Factor Psychopathy factor interaction Step 1 Age Step 2 PCL–R Interpersonal/Affective Factor PCL–R Social Deviance Factor Step 3 Interpersonal/Affective Social Deviance Psychopathy total scores with quadratic term Step 1 Age Step 2 PCL–R Total PCL–R Total2 (quadratic) Psychopathy factor scores with quadratic term Step 1 Age Step 2 PCL–R Interpersonal/Affective Factor PCL–R Interpersonal/Affective2 (quadratic) Step 1 Age Step 2 PCL–R Social Deviance Factor PCL–R Social Deviance2 (quadratic) Note. SE † p .076. standard error; PCL–R p .05. p .01. B SE R2 .09 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.03 0.03 0.01 0.03 0.02 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 .31 .01 .02 .09 .32 .02 .13 .09 .27 .01 .12 .09 .25 .07 .30 .30 .09 .28 .07 .30 .22† .06 .26 .09 .31 .08 .08 .29 .09 .27 .11 .10 .32 .09 .28 .02 .06 .11
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Psychopathy Checklist–Revised.
the PCL–R Social Deviance Factor (Factor 2) scores in the second block. As predicted, the R2 change due to the addition of both factor scores simultaneously was 0.07, more than expected if the variance accounted for by each variable in isolation was simply added. This indicates that cooperative suppression was present in the relationship between PCL–R factor scores and DPT performance. Also predicted, PCL–R Factor 1 and Factor 2 scores correlated with DPT performance in opposite directions (Table 3). Interpersonal/Affective factor (Factor 1) scores predicted larger Kendall’s tau values, meaning they predicted participants would provide punishment rankings that were more similar to the modal punishment rankings ( 0.30, p .01). PCL–R Social Deviance factor (Factor 2) scores simultaneously predicted smaller Kend-
all’s tau values, meaning they predicted participants would provide punishment rankings that were less similar to the modal punishment rankings ( 0.30, p .02). Given that neither factor significantly correlated with tau on their own, these results suggest that the PCL–R factors only predict DPT performance when their shared variance is removed through regression analysis. In other words, the characteristics of the Interpersonal/Affective factor that predict “normal” DPT performance are the characteristics that are unique and that do not overlap with the Social Deviance factor. Likewise, the characteristics of the Social Deviance factor that predict deviant DPT performance are the characteristics that are unique and that do not overlap with the Interpersonal/Affective factor.
676
SCHAICH BORG ET AL.
Regression Analysis: Interaction Between PCL–R Factor Scores and DPT Performance
To test whether an interaction between the Interpersonal/Affective factor (Factor 1) and the Social Deviance factor (Factor 2) could account for variance in DPT performance, we repeated the previously described regression with the two centered factor scores in the second block, and a multiplicative interaction term between the two factors in a third block. The interaction was significant ( 0.26, p .01). Collectively, the interaction term, the Interpersonal/Affective factor, and the Social Deviance factor accounted for 13.1% of the variance (p .01) in Kendall’s tau, with the individual factors accounting for 7.0% of the variance and the interaction term accounting for an additional 6.1% of the variance. To plot this interaction, we calculated the simple slopes of the effects of the Interpersonal/Affective factor for high (1 SD above the mean) and low (1 SD below the mean) Social Deviance factor scores according to the procedures described by Holmbeck (2002; see Figure 1). The effect of the Interpersonal/Affective factor was significant when Social Deviance factor scores were high ( 0.54, p .001), but not when Social Deviance factor scores were low (p .71). In contrast, the effect of the Social Deviance factor was significant when Interpersonal/Affective factor scores were low ( 0.46, p .001), but not when Interpersonal/Affective factor scores were high (p .90). To help interpret the results of the interaction, we plotted Interpersonal/Affective factor scores as a function of Social Deviance factor scores (Figure 2). This plot illustrated that the factor scores correlated positively in this study population (r .60) as reported in Table 2, but in addition clarified that Social Deviance factor scores started to plateau near their maximum possible value of 20 when Interpersonal/Affective factor scores were only slightly past their mean of 7.01 (out of a possible score of 20). Thus, part of the interaction effects identified in the regression analyses could be due to the fact that Social Deviance factor scores had already plateaued near their ceiling when Interpersonal/Affective factor
scores were high ( 1 SD in Figure 1), obscuring any effects of the Social Deviance factor but permitting the observation of the effects of the Interpersonal/Affective factor.
Implications of the Interaction Between PCL–R Factor Scores and DPT Performance for PCL–R Total Scores
PCL–R total scores are used more often than PCL–R factor scores to assess psychopathy in both research and clinical settings. Thus, in order to make our results more directly comparable to past research and applicable to clinical settings, we translated the results of the significant PCL–R factor interaction into corresponding PCL–R total scores. This was done by calculating the mean PCL–R total scores for participants with low (1 SD below the mean) Interpersonal/Affective (Factor 1) scores or high (1 SD above the mean) Interpersonal/Affective (Factor 1) scores. The mean PCL–R total score for participants with low Interpersonal/ Affective (Factor 1) scores was 17.5, with a range of 3.2–29. The mean PCL–R total score for participants with high Interpersonal/ Affective (Factor 1) scores was 28.9, with a range of 16 –36.8. Given that the Social Deviance Factor only predicted worse DPT performance when Interpersonal/Affective (Factor 1) scores were low, this curiously suggests that participants with mean PCL–R total scores of 17.5 may have worse DPT performance than those with mean PCL–R total scores of 28.9. In other words, participants with PCL–R total scores near the clinical cutoff for psychopathy (PCL–R total scores of 30 or above) may perform better than those with high levels of social deviance traits, but who have PCL–R total scores far below the clinical cutoff for psychopathy. To test these nonlinear predictions explicitly, we performed a hierarchical regression with Kendall’s tau as the dependent variable, age in the first block (recall that age was correlated with both PCL–R total scores and Kendall’s tau), and both a centered linear and a centered quadratic PCL–R total predictor in the second block. The relationship between the quadratic PCL–R Total predictor and tau was significant ( 1.46, p .01), and the linear and quadratic PCL–R total terms collectively accounted for 7.9% of the variance in DPT performance. This quadratic relationship between PCL–R total scores and Kendall’s tau is illustrated in the scatter plot provided in Figure 3. Thus, PCL–R total scores do correlate with DPT performance, but they do so in a nonlinear fashion. Participants with very low PCL–R total scores (who have simultaneously low Interpersonal/Affective and Social Deviance factor scores) have normal DPT performance. Participants with very high PCL–R total scores (who have simultaneously high Interpersonal/Affective and Social Deviance factor scores) have normal DPT performance as well. In contrast, participants with midlevel PCL–R total scores (who have higher Social Deviance scores than Interpersonal/Affective scores in the present population) have impaired DPT performance. Interestingly, when hierarchical regressions were performed with centered quadratic PCL–R Interpersonal/Affective or Social Deviance factor terms, only the Interpersonal/Affective quadratic term was significant (Table 3; Figure 4). Thus, the quadratic relationship between PCL–R total scores and DPT performance is likely driven by a combination of the nonlinear relationship between Interpersonal/Affective factor scores and DPT performance and the nonlinear relationship between Interpersonal/Affective and Social Deviance factor scores.
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Figure 1. Interaction between Deserved Punishment Test (DPT) performance (as measured by logit-transformed Kendall’s tau) and Psychopathy Checklist–Revised (PCL–R) Interpersonal/Affective (Factor 1) scores at high and low levels of PCL–R Social Deviance (Factor 2) scores. SD standard deviation.
PSYCHOPATHY AND PUNISHMENT JUDGMENTS
677
only approached significance (p .09). Therefore, the two-factor model of psychopathy described performance in the current study better than the four-facet model of psychopathy.
Discussion
Thus far, it has proven difficult to demonstrate deficits in moral judgment in individuals with psychopathy, despite their clearly immoral behavior. This study provides evidence that one reason for this historical difficulty may be that Interpersonal/Affective and Social Deviance psychopathic traits can predict moral judgment performance in opposing directions, ultimately canceling each other out and obscuring the relationship when PCL–R total scores are used as the only index of psychopathy. Consistent with previous studies showing no relationship between psychopathy and moral judgment, the present study also showed PCL–R total scores did not linearly predict DPT performance. However, by having a large sample size with a full range of PCL–R scores, choosing a task that minimized impression management and allowed for a wide range of performance, and testing for cooperative suppression effects, we demonstrated that PCL–R factor scores do predict moral judgment. When entered into a regression together, Interpersonal/Affective (Factor 1) PCL–R scores predict that incarcerated participants will rate the relative amount of punishment crimes deserve more like other participants, while the Social Deviance (Factor 2) PCL–R scores predict that incarcerated participants will rate the relative amount of punishment crimes deserve less like other participants. The factors’ opposing effects are driven by each factor’s unique qualities, because the zero-order correlations between each factor and DPT performance were not significant, likely because of a lack of correlation between their strong shared variance and Kendall’s tau. The associations between the Interpersonal/Affective factor and the Social Deviance factor were qualified by a significant multiplicative interaction. Specifically, Interpersonal/Affective traits were positively associated with task performance when Social Deviance traits were high, but Social Deviance traits were nega-
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Figure 2. Scatter plot of Psychopathy Checklist–Revised (PCL–R) Interpersonal/Affective (Factor 1) scores relative to PCL–R Social Deviance (Factor 2) scores.
Four Factor Model of Psychopathy and DPT Performance
Some models of psychopathy describe PCL–R scores as being composed of four facets (Hare, 2003). The PCL–R Interpersonal/ Affective factor can be divided into two facets: Facet 1 (Interpersonal, e.g., lying, manipulation, and superficial charm) and Facet 2 (Affective, e.g., shallow affect and lack of empathy). The PCL–R Social Deviance factor can be divided into Facet 3 (Lifestyle, e.g., impulsivity, irresponsibility, parasitic lifestyle, and lack of goals) and Facet 4 (Antisocial, e.g., juvenile delinquency, adult antisocial behavior, criminal versatility, and lack of behavioral control; Hare, 2003). To compare the four-facet model to the two-factor model and determine whether specific facets could account for the cooperative suppression between the Interpersonal/Affective factor and the Social Deviance factor, we implemented a hierarchical regression with Kendall’s tau as the dependent variable, age in the first block, and all four facets in the second block. The R2 change due to the addition of the facet scores simultaneously was 0.10 (p .02). The Interpersonal facet (Facet 1) predicted normal DPT performance ( 0.31, p .01) while the Antisocial facet (Facet 4) predicted more deviant DPT performance ( 0.23, p .04). Neither the Affective facet (Facet 2; p .60) nor the Lifestyle facet (Facet 3; p .29) significantly predicted DPT performance. Therefore, the positive relationship between the Interpersonal/Affective factor and DPT performance is likely due more to the interpersonal traits, rather than the affective traits, of this factor, and the negative relationship between the Social Deviance factor and DPT performance is likely due more to the antisocial traits, rather than the lifestyle traits, of this particular factor. A model including age in the first block, centered Interpersonal facet, centered Antisocial facet, and a multiplicative interaction between the centered Interpersonal and Antisocial facets did not account for as much variance in DPT performance as the interaction model using factor scores ( R2 .11). Further, the interaction between the Interpersonal facet and the Antisocial facet
Figure 3. Scatter plot of Psychopathy Checklist–Revised (PCL–R) total scores relative to deserved punishment test performance (as measured by logit-transformed Kendall’s tau).
678
SCHAICH BORG ET AL.
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Figure 4. Scatter plots of Psychopathy Checklist–Revised (PCL–R) factor scores (Interpersonal/Affective and Social Deviance) relative to deserved punishment test performance (as measured by logit-transformed Kendall’s tau).
tively associated with task performance when Interpersonal/Affective traits were low. The consequences of this interaction in the present population were that participants who had either very low PCL–R total scores or very high PCL–R total scores had more normal DPT performance than participants who had middle range PCL–R total scores. Put differently, Social Deviance psychopathic traits predicted deviant moral punishment judgment performance, but Interpersonal/Affective psychopathic traits counteracted these negative associations when they were sufficiently present. Although the nonlinear properties of the present incarcerated population needs to be replicated in other incarcerated populations, this interaction between opposing influences provides one mechanism by which the construct of psychopathy might manifest as a taxon, despite its composition of dimensional traits (Walters, Duncan, & Mitchell-Perez, 2007; Walters, Gray, et al., 2007). The mean PCL–R total scores for participants who had an Interpersonal/Affective (Factor 1) score above the population mean (where the positive relationship between Interpersonal/Affective scores and DPT performance is significant) in our sample was 28.9. This average is very close to the clinical PCL–R cutoff of 30 used to diagnose psychopathy, and thus suggests that people who have PCL–R total scores of above 30 may indeed perform differently on the DPT than those who have high psychopathic traits but score below 30. However, our data suggest their performance would be a result of qualified opposing linear interactions between specific subsets of psychopathic traits and moral judgment, not a result of simply having “enough” general psychopathic traits. It would be useful in future studies to assess whether cooperative suppression and PCL–R factor distributions can lead to similar taxonomic effects on other behaviors, perhaps explaining why individuals with very high psychopathy scores can have such distinct psychiatric profiles despite their diagnosis being grounded in the identification of dimensional, continuous traits. When interpreting the interaction effects in the present study, it is important to note that the present population sample did not include many participants with low Social Deviance (Factor 2) scores, especially low Social Deviance scores with simultaneous high Interpersonal/Affective (Factor 1) scores (Figure 2). Further, Social Deviance scores correlated with Interpersonal/Affective
scores when Interpersonal/Affective scores were low, but not when Interpersonal/Affective scores were high (because Social Deviance scores had already reached their ceiling). These population characteristics may have made it difficult to statistically detect the effects of the Interpersonal/Affective scores when Interpersonal/ Affective scores were low or detect the effects of the Social Deviance factor when Interpersonal/Affective scores were high. Thus, it would be informative in future studies to determine whether the interaction between PCL–R factors and moral judgment performance reported here persists in samples with many participants who have low Social Deviance and high Interpersonal/ Affective scores. That said, the characteristics of the present sample may be similar to the characteristics of other incarcerated populations because the antisocial behavioral characteristics of the Social Deviance factor are likely to be much more strongly represented in incarcerated samples than nonincarcerated samples. Most studies of psychopathy do not provide extensive details of their reported correlations between PCL–R factor scores. However, one study of male sex offenders providing adequate detail reported a relationship between Interpersonal/Affective and Social Deviance scores similar to the relationship reported here, including a lack of participants scoring high in Interpersonal/Affective traits but low in Social Deviance traits (Patrick, Bradley, & Lang, 1993; see Figure 2). Also interesting, in another study of primarily noninstitutionalized and noncriminal participants, researchers reported a lack of correlation between Interpersonal/Affective and Social Deviance scores in higher scoring participants who had PCL–R total scores of 15–27, despite significant correlations between Interpersonal/Affective and Social Deviance scores across their entire sample (DeMatteo, Heilbrun, & Marczyk, 2006). Therefore, it is possible that the distribution of factor scores reported here represents fundamental characteristics of the relationship between the PCL–R factors and thus may be present in other study populations as well. One important implication of the present study’s results is that the different subsets of personality traits in psychopathy may lead to different constellations of behaviors depending on which traits dominate in a given individual or in a particular life situation. The results of the present study suggest that psychopaths have impaired
PSYCHOPATHY AND PUNISHMENT JUDGMENTS
679
moral judgment and thus may have more antisocial behavior in contexts where the Social Deviance (Factor 2) traits would be more influential (such as when behavioral controls are required, i.e., when anger is invoked, when opportunities for strong and exciting sensations are presented, or when strong immediate rewards are offered). Conversely, individuals with psychopathy may have fully intact moral judgment and restrained antisocial behavior in contexts where the Interpersonal/Affective (Factor 1) traits would be more influential (such as when no powerfully provoking stimulus is in the immediate environment or when manipulation of other people is the ideal strategy). Further research is needed to test and develop this hypothesis. If true, the results may have implications for treatment programs aimed at modifying psychopathic traits, which historically have had debatable and inconsistent success in adult populations (Salekin, Worley, & Grimes, 2010). Specifically, our results may suggest that different strategies should be taught for situations in which Interpersonal/Affective traits are likely to dominate (including strategies that exploit moral reasoning) compared with situations in which Social Deviance traits are likely to dominate (which might require strategies that manage behavior rather than try to alter judgment). The deserved punishment test (DPT) was chosen for this study because it is a validated moral judgment test that would maximize statistical ability to detect associations with psychopathy. Additionally, the content of the task was very relevant to an incarcerated population, and the brain regions involved in punishment judgments (Buckholtz et al., 2008) overlap with the brain regions believed to be involved in psychopathy (Kiehl, 2006). However, it should be recognized that deciding how much punishment a moral wrong deserves is only one kind of moral judgment among others. In addition, how much punishment a moral wrong is seen to deserve can potentially be influenced by nonmoral factors such as arbitrary legal rules or evolved motives to cooperate (Cushman, 2008; Darley & Pittman, 2003; DeScioli & Kurzban, 2009, 2012; Kurzban, DeScioli, & O’Brien, 2007), and these nonmoral factors can result in situations where punishment is deemed necessary even though no moral wrong was committed. Thus, it is important for future studies to develop other appropriate moral judgment tasks that allow a wide range of performance to replicate our PCL–R factor results and assess whether PCL–R factor scores have opposing effects on other types of moral decision making as well. The data from this study help begin to unravel the enigma of psychopaths’ simultaneous reports of normal moral judgment and immoral behavior. First, in a population of individuals scoring above the recommended clinical diagnostic threshold for psychopathy, there is now evidence supporting the powerful intuition that psychopathic traits influence moral judgment, and equally important, this evidence provides plausible explanations for why these influences have been so difficult to uncover in the past. However, in other ways, these data underscore the enigma of apparently intact moral judgment in psychopathy. Most important, this study adds to the mounting evidence that PCL–R total scores do not correlate linearly with moral judgment. Further, even the nonlinear PCL–R total results suggest that many variables other than psychopathy affect how people make moral judgments or behave. Supporting this idea, it is important to acknowledge that PCL–R factor scores accounted for only 13.1% of the variance in DPT performance (when age is placed in the first block). This finding
reinforces the notion that impairments in moral judgment are not likely to be the strongest explanation for psychopaths’ antisocial behavior. Nonetheless, the results of this study demonstrate that features of psychopathy do influence moral judgment, which provides evidence to bridge our intuitive notions of the relationship between moral judgment and moral behavior with the dramatic clinical features of psychopathy.
References
Aharoni, E., Antonenko, O., & Kiehl, K. A. (2011). Disparities in the moral intuitions of criminal offenders: The role of psychopathy. Journal of Research in Personality, 45, 322–327. doi:10.1016/j.jrp.2011.02.005 Aharoni, E., Sinnott-Armstrong, W., & Kiehl, K. A. (2012). Can psychopathic offenders discern moral wrongs? A new look at the moral/ conventional distinction. Journal of Abnormal Psychology, 121, 484 – 497. doi:10.1037/a0024796 American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Blair, R. J. R. (1995). A cognitive developmental approach to morality: Investigating the psychopath. Cognition, 57, 1–29. doi:10.1016/00100277(95)00676-P Blair, R. J. R., Jones, L., Clark, F., & Smith, M. (1995). Is the psychopath “morally insane”? Personality and Individual Differences, 19, 741–752. doi:10.1016/0191-8869(95)00087-M Blonigen, D. M., Patrick, C. J., Douglas, K. S., Poythress, N. G., Skeem, J. L., Lilienfeld, S. O., . . . Krueger, R. F. (2010). Multimethod assessment of psychopathy in relation to factors of internalizing and externalizing from the Personality Assessment Inventory: The impact of method variance and suppressor effects. Psychological Assessment, 22, 96 –107. doi:10.1037/a0017240 Book, A. S., Holden, R. R., Starzyk, K. B., Wasylkiw, L., & Edwards, M. J. (2006). Psychopathic traits and experimentally induced deception in self-report assessment. Personality and Individual Differences, 41, 601– 608. doi:10.1016/j.paid.2006.02.011 Buckholtz, J. W., Asplund, C. L., Dux, P. E., Zald, D. H., Gore, J. C., Jones, O. D., & Marois, R. (2008). The neural correlates of third-party punishment. Neuron, 60, 930 –940. doi:10.1016/j.neuron.2008.10.016 Carlsmith, K. M., & Darley, J. M. (2008). Psychological aspects of retributive justice. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 40, pp. 193–236). San Diego, CA: Elsevier. doi: 10.1016/S0065-2601(07)00004-4 Cima, M., Tonnaer, F., & Hauser, M. D. (2010). Psychopaths know right from wrong but don’t care. Social Cognitive and Affective Neuroscience, 5, 59 – 67. doi:10.1093/scan/nsp051 Cleckley, H. (1941). The mask of sanity: An attempt to reinterpret the so-called psychopathic personality (1st ed.). St. Louis, MO: Mosby. Cleckley, H. (1976). The mask of sanity (5th ed.) St. Louis, MO: Mosby. Cohen, J., & Cohen, P. (1975). Applied multiple regression/correlation analysis for the behavioral sciences. Hillsdale, NJ: Erlbaum. Colby, A., & Kohlberg, L. (1987). The measurement of moral judgment: Theoretical foundations and research validations and standard issues scoring manual (Vols. 1 and 2). Cambridge, MA: Cambridge University Press. Cushman, F. (2008). Crime and punishment: Distinguishing the roles of causal and intentional analyses in moral judgment. Cognition, 108, 353–380. doi:10.1016/j.cognition.2008.03.006 Cushman, F., Young, L., & Hauser, M. (2006). The role of conscious reasoning and intuition in moral judgment: Testing three principles of harm. Psychological Science, 17, 1082–1089. doi:10.1111/j.1467-9280 .2006.01834.x Darley, J. M., & Pittman, T. S. (2003). The psychology of compensatory and retributive justice. Personality and Social Psycholology Review, 7, 324 –336. doi:10.1207/S15327957PSPR0704_05
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
680
SCHAICH BORG ET AL. ation with criminal conduct. In D. M. Stoff, J. Breiling, & J. D. Maser (Eds.), Handbook of antisocial behavior (pp. 22–35). New York, NY: Wiley. Hart, S. D., Knopp, P. R., & Hare, R. D. (1988). Performance of male psychopaths following conditional release from prison. Journal of Consulting and Clinical Psychology, 56, 227–232. doi:10.1037/0022-006X .56.2.227 Hemphill, J. F., Hare, R. D., & Wong, S. (1998). Psychopathy and recidivism: A review. Legal and Criminological Psychology, 3, 139 – 170. doi:10.1111/j.2044-8333.1998.tb00355.x Hicks, B. M., & Patrick, C. J. (2006). Psychopathy and negative emotionality: Analyses of suppressor effects reveal distinct relations with emotional distress, fearfulness, and anger-hostility. Journal of Abnormal Psychology, 115, 276 –287. doi:10.1037/0021-843X.115.2.276 Holmbeck, G. N. (2002). Post hoc probing of significant moderational and mediational effects in studies of pediatric populations. Journal of Pediatric Psychology, 27, 87–96. doi:10.1093/jpepsy/27.1.87 Kennealy, P. J., Hicks, B. M., & Patrick, C. J. (2007). Validity of factors of the Psychopathy Checklist–Revised in female prisoners: Discriminant relations with antisocial behavior, substance abuse, and personality. Assessment, 14, 323–340. doi:10.1177/1073191107305882 Kiehl, K. A. (2006). A cognitive neuroscience perspective on psychopathy: Evidence for paralimbic system dysfunction. Psychiatry Research, 142, 107–128. doi:10.1016/j.psychres.2005.09.013 Kosson, D. S., Smith, S. S., & Newman, J. P. (1990). Evaluating the construct validity of psychopathy in Black and White inmates: Three preliminary studies. Journal of Abnormal Psychology, 99, 250 –259. doi:10.1037/0021-843X.99.3.250 Kroner, D. G., & Weekes, J. R. (1996). Balanced inventory of desirable responding: Factor structure, reliability, and validity with an offender sample. Personality and Individual Differences, 21, 323–333. doi: 10.1016/0191-8869(96)00079-7 Kurzban, R., DeScioli, P., & O’Brien, E. (2007). Audience effects on moralistic punishment. Evolution and Human Behavior, 28, 75– 84. doi:10.1016/j.evolhumbehav.2006.06.001 Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33, 159 –174. doi:10.2307/ 2529310 Lapata, M. (2006). Automatic evaluation of information Ordering: Kendall’s tau. Computational Linguistics, 32, 471– 484. doi:10.1162/coli .2006.32.4.471 Lesaffre, E., Rizopoulos, D., & Tsonaka, R. (2007). The logistic transform for bounded outcome scores. Biostatistics, 8, 72– 85. Lilienfeld, S. O., & Penna, S. (2001). Anxiety sensitivity: Relations to psychopathy, DSM–IV personality disorder features, and personality traits. Journal of Anxiety Disorders, 15, 367–393. doi:10.1016/S08876185(01)00070-6 Link, N. F., Scherer, S. E., & Byrne, P. N. (1977). Moral judgement and moral conduct in the psychopath. Canadian Psychiatric Association Journal/Revue de l’Association des psychiatres du Canada, 22, 341– 346. Lose, C. A. (1997). Level of moral reasoning and psychopathy within a group of federal inmates. Dissertation Abstracts International: Section B: The Sciences and Engineering, 57(7-B), 4716. McMillan, J., & Malatesti, L. (2010). Responsibility and psychopathy. In L. Malatesti & J. McMillan (Eds.), Responsibility and psychopathy: Interfacing law, psychiatry, and philosophy (pp. 185–198). New York, NY: Oxford University Press. doi:10.1093/med/9780199551637.003 .0010 Meloy, J. R. (1988). Psychopathic mind: Origin, dynamics, and treatment. Northvale, NJ: Aronson. O’Kane, A., Fawcett, D., & Blackburn, R. (1996). Psychopathy and moral reasoning: Comparison of two classifications. Personality and Individual Differences, 20, 505–514. doi:10.1016/0191-8869(95)00203-0
DeMatteo, D., Heilbrun, K., & Marczyk, G. (2006). An empirical investigation of psychopathy in a noninstitutionalized and noncriminal sample. Behavioral Sciences & the Law, 24, 133–146. doi:10.1002/bsl.667 DeScioli, P., & Kurzban, R. (2009). Mysteries of morality. Cognition, 112, 281–299. doi:10.1016/j.cognition.2009.05.008 DeScioli, P., & Kurzban, R. (2012). A solution to the mysteries of morality. Psychological Bulletin. Advance online publication. doi:10.1037/ a0029065 Edens, J. F., Marcus, D. K., Lilienfeld, S. O., & Poythress, N. G. (2006). Psychopathic, not psychopath: Taxometric evidence for the dimensional structure of psychopathy. Journal of Abnormal Psychology, 115, 131– 144. doi:10.1037/0021-843X.115.1.131 First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1997). Structured Clinical Interview for DSM-IV Axis I Disorders–Clinical Version (SCID–IV). Washington, DC: American Psychiatric Press. Freeman, S. J. M., & Giebink, J. W. (1979). Moral judgment as a function of age, sex, and stimulus. The Journal of Psychology: Interdisciplinary and Applied, 102, 43– 47. doi:10.1080/00223980.1979.9915093 Gibbs, J. C., Basinger, K. S., Grime, R. L., & Snarey, J. R. (2007). Moral judgment development across cultures: Revisiting Kohlberg’s universality claims. Developmental Review, 27, 443–500. doi:10.1016/j.dr.2007 .04.001 Glenn, A. L. (2010). How can studying psychopaths help us understand the neural mechanisms of moral judgment. Cell Science Reviews, 6, 30 –35. Glenn, A. L., Raine, A., & Schug, R. A. (2009). The neural correlates of moral decision making in psychopathy. Molecular Psychiatry, 14, 5– 6. doi:10.1038/mp.2008.104 Glenn, A. L., Raine, A., Schug, R., Young, L., & Hauser, M. (2009). Increased DLPFC activity during moral decision making in psychopathy. Molecular Psychiatry, 14, 909 –911. doi:10.1038/mp.2009.76 Guay, J. P., Ruscio, J., Knight, R. A., & Hare, R. D. (2007). A taxometric analysis of the latent structure of psychopathy: Evidence for dimensionality. Journal of Abnormal Psychology, 116, 701–716. doi:10.1037/ 0021-843X.116.4.701 Hare, R. D. (1991). The Hare Psychopathy Checklist–Revised (PCL–R). Toronto, ON, Canada: Multi-Health Systems. Hare, R. D. (2003). Manual for the Revised Hare Psychopathy Checklist (2nd ed.). Toronto, ON, Canada: Multi-Health Systems. Hare, R. D., Clark, D., Grann, M., & Thornton, D. (2000). Psychopathy and the predictive validity of the PCL–R: An international perspective. Behavioral Sciences & the Law, 18, 623– 645. doi:10.1002/10990798(200010)18:5 623::AID–BSL409 3.0.CO;2-W Hare, R. D., Hart, S. D., & Harpur, T. J. (1991). Psychopathy and the DSM–IV criteria for antisocial personality disorder. Journal of Abnormal Psychology, 100, 391–398. doi:10.1037/0021-843X.100.3.391 Harenski, C. L., Harenski, K. A., Shane, M. S., & Kiehl, K. A. (2010). Aberrant neural processing of moral violations in criminal psychopaths. Journal of Abnormal Psychology, 119, 863– 874. doi:10.1037/a0020979 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 Harpur, T. J., Hare, R. D., & Hakstian, A. R. (1989). Two-factor conceptualization of psychopathy: Construct validity and assessment implications. Psychological Assessment, 1, 6 –17. doi:10.1037/1040-3590.1.1.6 Harris, G. T., Rice, M. E., & Quinsey, V. L. (1994). Psychopathy as a taxon: Evidence that psychopaths are a discrete class. Journal of Consulting and Clinical Psychology, 62, 387–397. doi:10.1037/0022-006X .62.2.387 Hart, S. D., & Hare, R. D. (1996). Psychopathy and antisocial personality disorder. Current Opinion in Psychiatry, 9, 129 –132. doi:10.1097/ 00001504-199603000-00007 Hart, S. D., & Hare, R. D. (1997). Psychopathy: Assessment and associ-
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
PSYCHOPATHY AND PUNISHMENT JUDGMENTS Patrick, C. J. (1994). Emotion and psychopathy: Startling new insights. Psychophysiology, 31, 319 –330. doi:10.1111/j.1469-8986.1994 .tb02440.x Patrick, C. J., Bradley, M. M., & Lang, P. J. (1993). Emotion in the criminal psychopath: Startle reflex modulation. Journal of Abnormal Psychology, 102, 82–92. doi:10.1037/0021-843X.102.1.82 Paulhus, D. L. (1991). Measurement and control of response bias. In J. P. Robinson, P. R. Shaver & L. S. Wrightsman (Eds.), Measures of personality and social psychological attitudes (pp. 17–59). New York, NY: Academic Press. Paulhus, D. L., Robins, R. W., Trzesniewski, K. H., & Tracy, J. L. (2004). Two replicable suppressor situations in personality research. Multivariate Behavioral Research, 39, 303–328. doi:10.1207/ s15327906mbr3902_7 Reidy, D. E., Shelley-Tremblay, J. F., & Lilienfeld, S. O. (2011). Psychopathy, reactive aggression, and precarious proclamations: A review of behavioral, cognitive, and biological research. Aggression and Violent Behavior, 16, 512–524. doi:10.1016/j.avb.2011.06.002 Rest, J., Cooper, D., Coder, R., Masanz, J., & Anderson, D. (1974). Judging the important issues in moral dilemmas: An objective test of development. Developmental Psychology, 10, 491–501. doi:10.1037/ h0036598 Robinson, P. H., & Darley, J. M. (2007). Intuitions of justice: Implications for criminal law and justice policy [Scholarship at Penn Law Paper 150]. Retrieved from http://lsr.nellco.org/upenn_wps/150 Robinson, P. H., & Kurzban, R. (2007). Concordance and conflict in intuitions of justice. Minnesota Law Review, 91, 1829 –1907. Ryan, J. J., Lopez, S. J., & Werth, T. R. (1999). Development and preliminary validation of a Satz–Mogel short form of the WAIS–III in a sample of persons with substance abuse disorders. International Journal of Neuroscience, 98, 131–140. doi:10.3109/00207459908994796 Salekin, R. T., Worley, C. B., & Grimes, R. D. (2010). Treatment of psychopathy: A review and brief introduction to the mental models
681
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
approach. Behavioral Sciences & the Law, 28, 235–266. doi:10.1002/ bsl.928 Schaich Borg, J., Sinnott-Armstrong, W., Calhoun, V. D., & Kiehl, K. A. (2011). Neural basis of moral verdict and moral deliberation. Social Neuroscience, 6, 398 – 413. doi:10.1080/17470919.2011.559363 Smetana, J. G. (1981). Preschool children’s conceptions of moral and social rules. Child Development, 52, 1333–1336. doi:10.2307/1129527 Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the State–Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press. Turiel, E. (1983). The development of social knowledge: Morality and convention. New York, NY: Cambridge University Press. Verona, E., Hicks, B. M., & Patrick, C. J. (2005). Psychopathy and suicidality in female offenders: Mediating influences of personality and abuse. Journal of Consulting and Clinical Psychology, 73, 1065–1073. doi:10.1037/0022-006X.73.6.1065 Verona, E., Patrick, C. J., & Joiner, T. E. (2001). Psychopathy, antisocial personality, and suicide risk. Journal of Abnormal Psychology, 110, 462– 470. doi:10.1037/0021-843X.110.3.462 Walters, G. D., Duncan, S. A., & Mitchell-Perez, K. (2007). The latent structure of psychopathy: A taxometric investigation of the Psychopathy Checklist–Revised in a heterogeneous sample of male prison inmates. Assessment, 14, 270 –278. doi:10.1177/1073191107299594 Walters, G. D., Gray, N. S., Jackson, R. L., Sewell, K. W., Rogers, R., Taylor, J., & Snowden, R. J. (2007). A taxometric analysis of the Psychopathy Checklist: Screening Version (PCL:SV): Further evidence of dimensionality. Psychological Assessment, 19, 330 –339. doi: 10.1037/1040-3590.19.3.330 Wechsler, D. (1997). WAIS–III administration and scoring manual. San Antonio, TX: Psychological Corp. Wilson, M. S., & McCarthy, K. (2011). Greed is good? Student disciplinary choice and self-reported psychopathy. Personality and Individual Differences, 51, 873– 876. doi:10.1016/j.paid.2011.07.028
(Appendices follow)
682
SCHAICH BORG ET AL.
Appendix A Deserved Punishment Test (DPT) Scenarios (S; Originally Published in Robinson & Kurzban, 2007) S1 Defending Attack
John is knocked down from behind by a man with a knife who moves to stab him. As the man lunges for him, John stabs him with a piece of glass he finds on the ground, which is the only thing he can do to save himself from being killed. The man later dies of his injuries.
S8 Clock Radio From Car
As he is walking to a party in a friend’s neighborhood, John sees a clock radio on the backseat of a car parked on the street. Later that night, on his return from the party, he checks the car and finds it unlocked, so he takes the clock radio from the back seat.
S9 Electric Drill From Garage S2 Coercive Threat to Child
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
A man grabs John’s child and puts a sharp knife to her throat. He tells John that he will kill the child if John does not steal an expensive digital camera from a nearby shop or he attempts to contact police. Because the man can see everything he does, John does as he is told in order to save his child. (What punishment for John?)
John does not have all the tools he needs for his workshop but knows of a family two streets over who sometimes leave unlocked the door to the detached garage next to their house. When he next sees his chance, he enters the detached garage through the unlocked door and takes a medium-size electric drill, intending to keep it forever.
S10 Microwave From House
While a family is on vacation, John jimmies the back door to their house and steps into their kitchen. On the counter, he sees their microwave, which he carries away.
S3 Umbrella Mistake
John takes another person’s umbrella assuming it to be his own because it is has the same unusual color pattern as his own, a fact that the police confirm.
S11 Smashing TV
While a family is away for the day, John breaks in through a bedroom window and rummages through the house looking for valuables. He can only find an 18-inch television, which angers him. When he gets it outside, he realizes that it is an older model than he wants, so he smashes it onto the driveway, breaking it into pieces.
S4 Wolf Hallucination
Another person slips a drug into John’s food, which causes him to hallucinate that he is being attacked by a wolf. When John strikes out in defense, he does not realize that he is in fact striking a person, a fact confirmed by all of the psychiatrists appointed by the state, who confirm that John had no ability to prevent the hallucination.
S12 Slap and Bruising at Record Store
A record store patron is wearing a cap that mocks John’s favorite band. John follows him from the store, confronts him, then slaps him in the face hard, causing him to stumble. The man’s face develops a harsh black and yellow bruise that does not go away for some time.
S5 Whole Pies From Buffet
The owner has posted rules at his all-you-can-eat buffet that expressly prohibit taking food away; patrons can only take what they eat at the buffet. The owner has set the price of the buffet accordingly. John purchases dinner at the buffet, but when he leaves he takes with him two whole pies to give to a friend.
S13 Head–Butt at Stadium
While attending a football game, John becomes angry as he overhears an opposing fan’s disparaging remarks about John’s team. At the end of the game, John sticks his face in the man’s face and head-butts him, causing a black eye and a gash that requires two stitches to close.
S6 Logo T–Shirt From Store
John notices in a small family owned music store a T–shirt with the logo of his favorite band. While the store clerk is preoccupied with inventory, John places the $15 T-shirt in his coat and walks out, with no intention of paying for it.
S14 Stitches After Soccer Game
Angry after overhearing another parent’s remarks during a soccer match in which John’s son is playing, John approaches the man after the game, grabs his coffee mug, knocks him down, and then kicks him several times while he is on the ground, knocking him out for several minutes and causing cuts that require five stitches.
S7 Short Change Cheat
John is a cab driver who picks up a high school student. Because the customer seems confused about the money transaction, John decides he can trick her and gives her $20 less change than he knows she is owed.
(Appendices continue)
PSYCHOPATHY AND PUNISHMENT JUDGMENTS
683
S15 Necklace Snatch at Mall
As a woman searches her purse for car keys in a mall parking lot, John runs up and grabs her gold necklace, but it does not break. He yanks the woman to the ground by her necklace, where she gashes her head, requiring stitches. John runs off without the necklace.
S20 Stabbing
John is offended by a woman’s mocking remark and decides to hurt her badly. At work the next day, when no one else is around, he picks up a letter opener from his desk and stabs her. She later dies from the wound.
S16 Attempted Robbery at Gas Station
John demands money from a man buying gas at a gas station. When the man refuses, John punches the man several times in the face, breaking his jaw and causing several cuts that each require stitches. He then runs off without getting any money.
S21 Ambush Shooting
John knows the address of a woman who has highly offended him. As he had planned the day before, he waits there for the woman to return from work, and when she appears, John shoots her to death.
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
S17 Clubbing During Robbery
To force a man to give up his wallet during a robbery attempt, John beats the man with a club until he relinquishes his wallet, which contains $350. The man must be hospitalized for 2 days.
S22 Abduction Shooting
A woman at work reveals John’s misdeeds to his employer, thereby getting him fired. John devises a plan to get even with her. The next week he forces the woman into his car at knife point and drives her to a secluded area where he shoots her to death.
S18 Mauling by Pit Bulls
Two vicious pit bulls that John keeps for illegal dog fighting have just learned to escape and have attacked a person who came to John’s house. The police tell John he must destroy the dogs, which he agrees to do but does not intend to do. The next day, the dogs escape again and maul to death a man delivering a package.
S23 Burning Mother for Inheritance
John works out a plan to kill his 60-year-old invalid mother for the inheritance. He drags her to her bed, puts her in, and lights her oxygen mask with a cigarette, hoping to make it look like an accident. The elderly woman screams as her clothes catch fire, and she burns to death. John just watches her burn.
S19 Infant Death in Car
John is driving to see a man about buying an illegal gun but must babysit his friend’s toddler son. It occurs to him that it is too hot to safely leave the toddler in the car, but he decides to leave him anyway and to return soon. He gets talking with the seller, however, and forgets about the toddler, who passes out and dies.
S24 Ransom, Rape, Torture and Strangling
John kidnaps an 8-year-old girl for ransom, rapes her, and then records the child’s screams as he burns her with a cigarette lighter, sending the recording to her parents to induce them to pay his ransom demand. Even though they pay as directed, John strangles the child to death to avoid leaving a witness.
(Appendices continue)
684
SCHAICH BORG ET AL.
Appendix B Regression Analyses With Psychopathy Checklist–Revised Total and Factor Scores as Predictors of Deserved Punishment Test Performance (Measured by Raw, Nontransformed Kendall’s Tau Values)
Variable Psychopathy total scores Step 1 Age Step 2 PCL–R Total Psychopathy factor scores (entered independently) Step 1 Age Step 2 PCL–R Interpersonal/Affective Factor Step 1 Age Step 2 PCL–R Social Deviance Factor Psychopathy factor scores (entered simultaneously) Step 1 Age Step 2 PCL–R Interpersonal/Affective Factor PCL–R Social Deviance Factor Psychopathy factor interaction Step 1 Age Step 2 PCL–R Interpersonal/Affective Factor PCL–R Social Deviance Factor Step 3 Interpersonal/Affective Social Deviance Psychopathy total scores with quadratic term Step 1 Age Step 2 PCL–R Total PCL–R Total2 (quadratic) Psychopathy factor scores with quadratic term Step 1 Age Step 2 PCL–R Interpersonal/Affective Factor PCL–R Interpersonal/Affective2 (quadratic) Step 1 Age Step 2 PCL–R Social Deviance Factor PCL–R Social Deviance2 (quadratic) Note. SE † p .07.
§
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
standard error; PCL–R p .06. p .05.
Psychopathy Checklist–Revised. p .01.
(Appendices continue)
PSYCHOPATHY AND PUNISHMENT JUDGMENTS
685
Appendix C Regression Analyses With IQ, Education Level, Age, and Psychopathy Checklist–Revised Total and Factor Scores as Predictors of Deserved Punishment Test Performance (Measured by Logit-Transformed Kendall’s Tau)
Variable Psychopathy total scores Step 1 IQ Education level Step 2 Age Step 3 PCL–R Total Psychopathy factor scores (entered independently) Step 1 IQ Education level Step 2 Age Step 3 PCL–R Interpersonal/Affective Factor Step 1 IQ Education level Step 2 Age Step 3 PCL–R Social Deviance Factor Psychopathy factor scores (entered simultaneously) Step 1 IQ Education level Step 2 Age Step 3 PCL–R Interpersonal/Affective Factor PCL–R Social Deviance Factor Psychopathy factor interaction Step 1 IQ Education level Step 2 Age Step 3 PCL–R Interpersonal/Affective Factor PCL–R Social Deviance Factor Step 4 Interpersonal/Affective Social Deviance Psychopathy total scores with quadratic term Step 1 IQ Education level Step 2 Age
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Step 3 PCL–R Total PCL–R Total2 (quadratic) Psychopathy factor scores with quadratic term Step 1 IQ Education level Step 2 Age Step 3 PCL–R Interpersonal/Affective Factor PCL–R Interpersonal/Affective2 (quadratic) Step 1 IQ Education level Step 2 Age Step 3 PCL–R Social Deviance Factor PCL–R Social Deviance2 (quadratic) Note. SE §§ p .08.
†
Appendix D Histograms Illustrating the Distribution of Kendall’s Tau Values From the Present Incarcerated Population and Previously Published Nonincarcerated Populations
University Non-incarcerated Population (N = 64) Frequency
30 20 10 0 0 0.7 0.1 0.2 0.3 0.4 0.5 0.6 0.8 0.9 0.9 0.9 1 1 1
This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Kendall's Tau Internet Non-incarcerated Population (N = 246) Frequency
60 40 20 0 0.1 0.2 0.4 0.5 0.6 0.7 0.7 0.3 0.8 0.8 0
Kendall's Tau Incarcerated Population (N = 100)
20 15 10 5 0 0 0.1 0.3 0.2
Frequency
Kendall's Tau
0.6
0.4
0.5
Received June 12, 2012 Revision received March 18, 2013 Accepted May 23, 2013
