Psychopaths Know Right from Wrong but Don't care

Maaike Cima; Franca Tonnaer; Marc D. Hauser


Soc Cogn Affect Neurosci. 2010;5(1):59-67. 

In This Article



Participants (all male adults) provided informed consent in accord with the policies of the Ethical Commission of the Faculty of Psychology and Neuroscience, Maastricht University, The Netherlands. Healthy controls (n = 35; mean age = 30.29 years, SD = 9.99) were recruited from the south of the Netherlands. The psychopath (n = 14; mean age = 36.66 years; SD = 6.55) and non-psychopath (n = 23; mean age = 40.95 years; SD = 9.77) offenders were sampled from the Forensic Psychiatric Centre de Rooyse Wissel (FPCdRW) in Venray, the Netherlands. Of the 37 delinquents, IQ scores were available for a subgroup of 20 (7 psychopaths and 13 non-psychopathic offenders) participants. Though mean IQ scores for the psychopaths (M = 81.6, SD = 8.66) was slightly lower than for the non-psychopathic offenders (M = 92.5; SD = 19.37), there was no group difference [t(18) = 1.41; P = 0.18; d = 0.66].

Clinical Diagnosis

Psychopathy was assessed by a clinician presenting the Psychopathic Checklist-Revised [PCL-R (Hare, 1991)] test. The PCL-R is a reliable and valid instrument, designed to measure psychopathic traits such as antisocial behaviour, shallowness, impulsivity, callousness, criminal history, and lack of moral emotions, based on evidence obtained from medical and juridical records and documents, as well as extensive interviews with the forensic patients. Based on a study of Grann et al. (1998), a PCL-R cutoff score of 26 was used to divide the current sample into psychopaths (PCL-R ≥ 26) and non-psychopaths (PCL-R < 26). Total PCL-R scores were available for all 37 offenders. However, of the 14 psychopaths, 2 were described in the crime record as having high PCL-R scores, without mentioning the exact scores. Therefore, the relationship within the psychopathic group between PCL-R scores and type of crime (Figures 3 and 4) were only available for 12 psychopaths. Finally, regarding PCL-R factor scores, Factor 1 and Factor 2 scores were only available for 15 subjects.

Figure 3.

(A) Scatterplot of psychopathy factor 1 scores for offenders [psychopaths and non-psychopathic delinquents[ [N = 15*] grouped by the proportion of yes judgments for all personal moral dilemmas. (B) Scatterplot of psychopathy factor 2 scores for offenders [N = 15*] grouped by the proportion of yes judgments for the personal moral dilemmas. *: For 20 individuals, crime files documented the administration of a PCL-R interview. However, these juridical file records only described PCL-R total score without mentioning both factor scores.

Figure 4.

(A) Scatterplot of psychopathy scores (PCL-R; N = 12*) grouped by type of crime, against the proportion of yes judgments for all personal moral dilemmas. (B) Scatterplot of psychopathy scores (PCL-R; N = 12*) grouped by type of crime, against the proportion of yes judgments for the high conflict, other-serving personal moral dilemmas. *: For two individuals, crime files documented the administration of a PCL-R interview. However, these juridical file records only described these subjects as being a psychopath without mentioning the exact PCL-R score.

The PCL-R has two main factors. The first factor comprises interpersonal and affective characteristics of psychopathy, including shallow affect, lack of remorse or guilt and glibness/superficial charm (Cooke and Michie, 2001; Hare, 2003). The second impulsive, antisocial and unstable lifestyle factor comprises the social deviance characteristics, and includes impulsivity, early behavioural problems, and parasitic lifestyle (Cooke and Michie, 2001; Hare, 2003).

All psychopathic offenders had a personality diagnosis (Table 1). Most of them (57.1%) had a diagnosis of cluster B (narcissistic personality disorder or antisocial personality disorder), while the minority had a personality disorder not otherwise specified. Of the non-psychopathic offenders, the majority suffered a personality disorder not otherwise specified, 21.7% had a cluster B personality disorder, and 4.3% had a cluster C personality disorder.

To strengthen the link to emotion all subjects in our test groups also participated in a well-established, physiological test of stress reactivity involving measures of cortisol [i.e., Trier Social Stress Test; (Kirschbaum et al., 1993; Kirschbaum et al., 1995)]. Results showed that psychopathic offenders, unlike the two comparison groups, showed no significant increase in cortisol in response to the stressor. Within both the non-psychopathic group and healthy controls, cortisol levels significantly increased at T1 (before administering the stressor) to T3 (after administering the stressor) as demonstrated by pairwise comparison (all t's > 2.81; all P's < 0.01 and all t's > 3.07; all P's < 0.01, respectively). In contrast, within the psychopathic group there was no significant increase of cortisol levels between T1 and T3 (all t's < 1.00; all P's > 0.34); see Cima, Popma, and Nicolson (in preparation) for a more detailed overview of these data. Thus, based on both their PCL-R scores and stress reactivity profiles, this psychopathic population showed relatively flat emotional responses, consistent with many other studies.

Participants had no history of psychosis or depression, and no current alcohol or drug dependence. In the delinquent sample these criteria were considered by the psychologist, psychiatrist and file records. In the healthy control sample, these criteria were inquired. We tested all offenders on the moral dilemmas after they had been interviewed with the PCL-R.

Summarizing, both psychopaths and non-psychopathic delinquents differed from healthy controls in that they had been convicted of crimes; and as in numerous other studies (Herpertz et al., 2001; Kirschbaum et al., 1995; Williamson et al., 1991), psychopaths differed fundamentally from non-psychopath delinquents in that they showed diminished emotional reactivity based on both the standard clinical diagnostic test [i.e. the PCL-R (Hare, 1991)] and physiological measures (Cima et al., in preparation).


We used the moral dilemmas of Greene et al. (2001, 2004), previously tested with the VMPC patients (Koenigs et al., 2007). Each dilemma was first translated into Dutch by FT, back-translated into English by a second bilingual Dutch-English speaker, and then checked by MH for accuracy. We presented seven impersonal and 14 personal moral dilemmas (see Supplementary Information). Subjects then answered "yes" or "no" to the question "Would you X?". A population of native Dutch speakers (adults, 324 females, 348 males) judged these cases on a Dutch version of the Moral Sense Test (, whereas the three test groups responded to these dilemmas with paper and pencil. Though we recognize that studies of moral judgment and responses to artificial dilemmas in particular, represent only one of several valid approaches to understanding our moral psychology, we used this approach to provide the most direct comparison with VMPC patients, as well as other recent studies of intuitive moral judgments.

To control the possibility that psychopaths simply lie about their responses to our moral dilemmas, we also administered a questionnaire [Socio-Moral Reflection; SRM-SF (Gibbs et al., 1992)], asking straightforward and explicitly whether certain familiar transgressions would be morally permissible. For instance, "How important is it to keep a promise to your friend?"; "How important is it not to steal?". Answers could be given on a 5-point scale, ranging from very unimportant to very important. Scores on the SRM-SF questionnaire result in a total score and a score of moral standing, indicating the level of moral development.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: