Understanding the dark side of personality in sex offenders considering the level of sexual violence

Abstract

This study examined the relationship between the actual level of physical violence in sexual offenses and dark triad, empathic and impulsive personality traits of their perpetrators. Sixty-four male perpetrators of sexual offenses without any serious mental illness were included. A 5-point Likert-type coding system based on Violence Profile for Current Offense was applied to assess the severity of physical violence of each sexual offense. Personality traits of dark triad (psychopathy, Machiavellianism, narcissism), trait empathy and impulsiveness were also evaluated. Multivariate analyses indicated that non-stranger victim, secondary psychopathy, narcissism and empathy could significantly predict greater involvement of physical violence in a sexual offense. In addition, empathy was negatively correlated with all dark triad traits. Sexual violence should be conceptualized in the form of a continuum, and, considering such an approach, offenders with high secondary psychopathy and narcissism may show preference for sexual assaults that are more violent in nature.

1. Introduction

Sexual offenses are devastating, affect-laden events that are a frequent concern for men, women and children throughout the world. The experience of sexual assault is highly distressing for both the victim and the society. Victims of all ages undergo profound adverse outcomes in physical, psychological and social contexts at an individual level and especially if the victim belongs to vulnerable groups such as children (Bradford, 2006; Krug, Mercy, Dahlberg, & Zwi, 2002; Norman et al., 2012). A variety of offenses come under the heading of ‘sexual offenses’, including rape, verbal or indecent assault, indecent exposure, stalking, child sexual abuse and the creation and dissemination of child pornography (Dartnall & Jewkes, 2013; Robertiello & Terry, 2007; Thornton, 2021). Criminological preoccupations with particular kinds of sexual offenses have varied considerably over time. The essential element of any sexual offense is the absence of consent, either because of coercion, one of the parties being underage, or because of the nature of the relationship, such as child–parent relationship.

Approximately 5–20% of all offenders in correctional systems were reported to be convicted of a sexual offense (Arbanas, Marinović & Buzina, 2020; Cook, Fox, Weaver, & Graham Root, 1991; Porter et al., 2000). Since only 5% of acts of sexual violence against adults are reported to the police, and 20% of those crimes end with a conviction, the actual numbers of sexual offenses against children and adults can only be estimated. Large numbers of people, both children and adults, report having been subject to some form of sexually unwelcome behavior. Previous epidemiological and meta-analysis research indicated that about 10–30% of women and 5–15% of men were sexually abused as children (Barth, Bermetz, Heim, Trelle, & Tonia, 2013; Finkelhor, 1994; Krug et al., 2002; Stoltenborgh, van IJzendoorn, Euser, & Bakermans-Kranenburg, 2011). Moreover, 7–13% of women and about 3% of men were reported to have been sexually assaulted at some time during their lives (Abrahams et al., 2014; Krahé et al., 2015; Spitzberg, 1999). However, the lack of official records for a large proportion of sexual assaults makes more recent worldwide statistics tenebrous. A recent review of global statistics of sexual assault has yielded that past-year prevalence of sexual assault victimization in adolescence and adulthood was as high as 60% in some countries (Dworkin, Krahé, & Zinzow, 2021).

Therefore, it is necessary not only to help victims of sexual violence but also to focus on prevention of sexual violence by identifying characteristics of offenders and those at risk to offend. Furthermore, sexual recidivism rate of adult sex offenders appears to be high compared to non-sex offenders (Porter et al., 2000; Rettenberger, Briken, Turner, & Eher, 2015). Concerning the high prevalence of sexual offenses and its impact on its victims, it would be negligent of society and the medical community not to try and reach potential/not-convicted perpetrators. Indeed, sexual offending remains a complex phenomenon since multiple and complex risk factors, including individual and environmental factors that interact with each other, may have causal impact. Moreover, sex offenders are a highly heterogeneous and diversified group on many accounts, including demographic backgrounds, criminal diversity, treatment needs, psychological and developmental characteristics, and personality traits (Margari et al., 2015; Porter et al., 2000). Myriad dispositional and contextual factors are considered to have substantial influence on motivations that induce sexual offenses (Ward & Beech, 2006), including a power-motivated and sadistic nature (Knight, 1999; Robertiello & Terry, 2007).

Sexually motivated offending behavior of sex offenders is heterogeneous, particularly in terms of the characteristics of sexual offenses including intensity and level of violence. Some authors have made the argument that violence is a subset of aggression; moreover, aggression earns the label ‘violence’ when the consequences are severe. Nonetheless, violent behavior is best conceptualized as being on a continuum of severity with relatively minor acts of aggression (e.g. verbal) at the low end of the spectrum and excessive physical harm (e.g. homicide) at the high end of the spectrum (Allen & Anderson, 2017a). Therefore, violence, as a construct, requires specific definition and measurement. To date, research examining the level of violence involved in sexual offenses is lacking. Indeed, many studies have classified sexual offenses in the category of violent offenses (Munkner, Haastrup, Joergensen, & Kramp, 2009; Wanklyn, Ward, Cormier, Day, & Newman, 2012). Such an approach may, however, disregard the fact that each sexual assault has an intrinsic violence profile that is distinctive for the offense, offender characteristics and motivation. In fact, only a handful of studies have examined sex offenders who are classified as violent or non-violent according to whether violence was present when they committed their offenses (Frick, 1999; Gretton, McBride, Hare, O’Shaughnessy, & Kumka, 2001; Mercado & Scalora, 2001; Porter, Brinke, & Wilson, 2009; Rada, Laws, Kellner, Stivastava, & Peake, 1983; Van Wijk, Mali, Bullens, & Vermeiren, 2007; West, 1984). Other studies have classified sexual offenders as contact or non-contact (Geddes & Andreasen, 2020; Thornton, 2021), rapist or non-rapist (Långström, Sjöstedt, & Grann, 2004; Woodworth et al., 2013) and homicidal or non-homicidal (Beauregard, DeLisi, & Hewitt, 2018; Koch, Berner, Hill, & Briken, 2011; Stefanska, Beech, & Carter, 2016). However, such empirical and heterogeneous dichotomizations have shortcomings since they prevent an understanding of the actual degree of violence involved in the referral (index) sexual offense. Thus, a deeper insight into the psychological/personality traits of the sex offenders on the basis of the severity of violence of their act has not been put forth, yet. Furthermore, previously published studies generally lack the use of standardized instruments to evaluate the severity of violence accompanying a sexual offense. Put together, each offender has distinct traits; therefore, there is a requisite to consider dispositional, contextual and disinhibiting determinants and their interrelationship with the offense to better comprehend pathways related to different levels of violence involved in sexual offending.

According to the general aggression model, which is an integrative, bio-social-cognitive, developmental approach to understanding aggression, both intrinsic and environmental modifiers affect personality,which is considered the summary of a person’s knowledge structures. Personality refers to a process in which personal and situational factors interact with each other, and such interaction influences a person’s present status of cognition, emotion and arousal, which in turn affects appraisal and decision processes that subsequently determine aggressive and nonaggressive outcomes (Allen & Anderson, 2017b; Allen, Anderson, & Bushman, 2018). Therefore, understanding personality as a function of aggression and violence seems of critical importance.

The dark triad of personality is a constellation of traits characterized by a willingness to exploit and manipulate others, callousness, disagreeableness, deceitfulness, egocentrism, lack of honesty and humility, empathy deficits and a focus on agentic goals (Paulhus & Williams, 2002). Such traits have been previously associated with sexual coercion (Pavlović, Markotić, & Bartolin, 2019; Zeigler-Hill, Besser, Morag, & Keith Campbell, 2016) and consist of Machiavellianism (proclivity to exploit others), narcissism (malevolent behavior associated with a grandiose identity) and psychopathy (Paulhus & Williams, 2002). Psychopathy is characterized as a form of personality disorder with a distinctive pattern of interpersonal (e.g. deceitfulness and manipulation), affective (e.g. lack of remorse, lack of empathy or guilt) and behavioral (e.g. irresponsibility and impulsivity) characteristics (Hare & Neumann, 2008). High level of psychopathy is indeed predictive of a host of antisocial behaviors including violent offending (Grann, Långström, Tengström, & Kullgren, 1999; Hare, Clark, Grann, & Thornton, 2000) and sexual assault (Brown, Dargis, Mattern, Tsonis, & Newman, 2015; Hanson & Harris, 2000; Seto & Barbaree, 1999; Sohn, Reyes, & Kim, 2022). The defining characteristics of psychopathy make it possible for psychopaths to perform heinous and appalling acts without feeling concern for their victims, making it a cardinal variable to consider when trying to understand deviant sexual behavior (Woodworth et al., 2013). Empathy, a pervasive lack of which is a cardinal component of the psychopathy construct, is both a cognitive process and an affective capacity and refers to understanding and sharing the context of another person’s emotional state (Blake & Gannon, 2008). Emotion recognition deficits are suggested to play a role in the low empathy found in individuals with high psychopathy (Kirsch & Becker, 2007). Psychopathy leads to a failure to develop empathy and poor moral socialization, and a greater likelihood to engage in violent and sexual offending (Hanson, 2003; Smallbone, Wheaton, & Hourigan, 2003). There is empirical support for the idea that empathy deficit is a risk factor for greater involvement in violent and sexual offending; moreover, this risk works as cognitive blocks for empathy, exacerbating a lack of concern for victims and an inability to cope with negative emotions (Barnett & Mann, 2013; Smallbone et al., 2003). A few studies have reported a relationship between lower levels of empathy and higher levels of sexual aggressiveness (Lisak & Ivan, 1995); however, as far as we are aware, global and trait empathy have not been evaluated in the context of the level of violence in sexual offenses.

That an offender will commit further sexual assault is a risk with explicit concern, and this risk is highly dependent on the characteristics of previous sexual offenses. Although an individual’s history of offending provides information about their future offending, it is better to characterize this in terms of risk dimensions (McNally & Fremouw, 2014; Thornton, 2021). Therefore, reducing recidivism of sex offenders is best accomplished by identifying the relationship between their dispositional and offense-related characteristics (Robertiello & Terry, 2007). It is a plausible argument that the level of violence displayed within acts of sexual offense can help in discriminating between more dangerous or less dangerous sex offenders, and hence result in better decisions about the level of community notification required. However, identifying these offenders not only should be based on the violence index but should encompass an understanding of the relationship between violence severity and dispositional/psychological factors. Surprisingly, the recent and relevant literature on sexual offending lacks clarification and has largely neglected to examine the link between violence severity and personality-related inclinations for sexual assaults. The current body of evidence provides only little evidence that increased levels of violence in sexual offenses is indicative of certain personality traits. This avenue of research is important because it may help identify dispositional risk factors associated with contextual factors, all of which may contribute to increased intensity of violence engaged during the sexual assault. Therefore, this study aimed to examine psychopathy, Machiavellianism, narcissism, empathy and impulsiveness as personality traits that may determine the level of physical violence used in sexual assaults by perpetrators. We hypothesized that higher levels of dark traits of personality including psychopathy were associated with greater levels of physical violence used in sexual offenses.

2. Method

2.1. Study design, participants and procedure

The current study was designed as a single-center cross-sectional study of male subjects who were identified as defendants of a sexual offense and sent for psychiatric evaluation to the Council of Forensic Medicine, Fourth Specialization Unit. This unit serves as a specialized medical board of expert witnesses at the request of criminal and civil courts. According to the Turkish Penal Code (Law No. 5237), security measures are imposed instead of punishment if an individual cannot comprehend the legal meaning or consequences of the act they have committed due to the existence of a mental illness at the time of the offense (cognitive component), or if the individual’s ability to control their behavior is significantly diminished (volitional component). All parties in the due process of law can request an evaluation of criminal responsibility of the defendant (Balcioglu, Dogan, Incı, & Solmaz, 2020). Eighty-five defendants who allegedly committed a sexual offense according to the legislation, were initially identified between January 2020 and December 2020 for the current study. These were individuals who were charged by public prosecution, but not yet convicted. Trials of all but two defendants were concluded by January 2022. These two subjects were excluded from the current study, in addition to three subjects who were found not guilty during the legal process. We excluded 10 participants as they had been recently or previously diagnosed with serious mental illnesses such as schizophrenia spectrum and other psychotic disorders, bipolar and related disorders and depressive disorders based on the Diagnostic and Statistical Manual of Mental Disorders–Fifth Edition (DSM–5) and commonly defined in the forensic psychiatry literature (Hodgins et al., 2007). Subjects with intellectual disability, cognitive impairment (dementia or any other neurological condition) and intoxication with psychoactive substance were also excluded. Six defendants were ruled out due to illiteracy, lack of cooperation with the psychometric instruments, insufficient data and not giving consent to participate in the study. The study was approved by the local Ethics Committee [IRB: 10.12.2019–21589509/2019/1005] and conducted according to the Helsinki declaration. Written informed consent was obtained from all participants following a thorough explanation of the aims of the study.

A semi-structured data form that included sociodemographic, historical and forensic information was applied to the final sample of 64 participants. Dark triad personality traits (psychopathy, Machiavellianism and narcissism), empathy and impulsiveness were evaluated with the Levenson Self-Report Psychopathy Scale, Dirty Dozen Scale, Empathy Quotient and Barratt Impulsiveness Scale. The physical violence profile of each sexual offense was further evaluated by the researchers according to the characteristics surrounding the index sexual offense.

2.2. Evaluation instruments

2.2.1. Semi-structured data form

We prepared a detailed questionnaire to collect information on the individual’s background characteristics including demographics as well as historical and forensic information. Data were acquired via case interviews, hard copies of files, electronic databases for forensic and medical case records, social service reports and interviews with family members. To identify specific characteristics surrounding the index sexual offense, detailed information was collected via reports from law enforcement officers, investigation and trial records, and statements from the defendants, victims and witnesses.

2.2.2. Levenson Self-Report Psychopathy Scale

The Levenson Self-Report Psychopathy Scale (LSRP) consists of 26 items, each of which is rated on a 4-point Likert-type questionnaire, in which 1 indicates ‘strongly disagree’, and 4 indicates ‘strongly agree’. The scale was designed to measure psychopathic personality traits and behavior in non-institutionalized populations (Levenson, Kiehl, & Fitzpatrick, 1995). The LSRP includes a Primary scale, which measures callous traits and manipulative behaviors toward others, and a Secondary scale, which identifies antisocial tendencies, impulsivity and deficiency in behavioral control, similar to the two-factor psychopathy model evaluated in the Psychopathy Checklist–Revised (PCL–R; Hare, 2003). The Turkish version of the scale was reported by Tulu et al. in a study involving incarcerated sex offenders (Tulu & Erden, 2013). The Cronbach’s alpha coefficients were .84 for the Primary psychopathy, .65 for the Secondary psychopathy and .77 for the total scale.

2.2.3. Dirty Dozen Scale

The Dirty Dozen Scale (DDS) consists of 12 items; each item is rated using a 5-point Likert scale ranging from 1 to 5 (Jonason & Webster, 2010). DDS is composed of three subdimensions: Machiavellianism, psychopathy and narcissism, considered as the dark triad traits of personality (Paulhus & Williams, 2002). Validity and reliability of the Turkish version of DDS were reported by Özsoy et al. in a sample of university students, in which Cronbach’s alpha coefficients were .81 for Machiavellianism, .67 for psychopathy and .80 for narcissism (Özsoy, Rauthmann, Jonason, & Ardıç, 2017). In the current study, the psychopathy subdimension was made redundant since psychopathy was evaluated with the LSRP. The instrument demonstrated an acceptable level of internal consistency for the entire sample in the current study with a Cronbach’s alpha score of .74 for Machiavellianism and .81 for narcissism.

2.2.4. The Barratt Impulsiveness Scale, Version 11 (BIS–11)

The Barratt Impulsiveness Scale, Version 11 (BIS–11) is a self-report scale that is used to assess trait impulsiveness. It consists of 30 items and has three subscales: attention, motor and non-planning impulsiveness. BIS–11 is structured to assess long-term patterns of behavior and is used to assess trait levels of impulsiveness across a variety of populations (Stanford et al., 2009). The higher the total BIS–11 score, the higher the patient’s level of impulsiveness. The first version of BIS–11 was developed by Barratt in 1959, who investigated the relationship between anxiety and impulsiveness, but its factorial structure was reported later in 1995 (Patton, Stanford, & Barratt, 1995). The Turkish validity and reliability study of BIS–11 was performed by Güleç et al. (2008), in which the Cronbach’s alpha coefficient was found to be .80. In the current study, the internal consistency of the total scale was acceptable, with a Cronbach’s alpha score of .81.

2.2.5. Empathy Quotient (EQ)

Empathy Quotient (EQ) is a 60-item self-report scale that is designed to examine global empathy competencies. The scale was developed by Simon Baron-Cohen in 2004. Responses are given on a 4-point scale ranging from 1 (strongly agree) to 4 (strongly disagree). Each of the items gets a score of 1 point if the respondent records mildly empathic behavior, or 2 points if strongly. These scores are summed for a total empathy score. The EQ is considered to be the most comprehensive, reliable and valid empathy scale to date. The EQ scores well with a 12-month test–retest reliability with r = .97 and a Cronbach’s alpha measured validity of .92 (Baron-Cohen & Wheelwright, 2004). In the current study, the Cronbach’s alpha coefficient for EQ was found to be .82.

2.2.6. Evaluation of physical violence profile of the index sexual offense

To assess and classify the seriousness of the index sexual offense, we used a coding system based on Violence Profile for Current Offense (VPCO), which was developed by Taylor (1985) and adapted to the Turkish recently (Balcioglu & Oncu, 2021). This tool was devised to avoid vague and sharply demarcated classification of offenses that is sometimes used in legal description as a matter of chance. Rather, the tool assesses incremental levels of physical violence displayed in each offense, through an evaluation of its behavioral components according to information gleaned from all available sources (Gunn & Bonn, 1971; Gunn & Robertson, 1976). It is a single-item instrument ranking the severity of violence accompanying each offense. A score of ‘0’ is assigned as completely non-violent, while ‘1’ is assigned as minimal violence, including verbal aggression, carrying a weapon that was not usedand minimal and accidental damage to property. A score of ‘2’ is assigned to moderate violence, including actual bodily harm, sexual offense under force and use of an offensive weapon but without causing injury and damage to property when this was the main intent. A score of ‘3’ refers to moderately serious violence, including grievous bodily harm, extensive damage to property and a potential threat to life. A score of ‘4’ is assigned to serious violence where the victim died, life was actually endangered or the victim was detained in hospital for more than 24 hours. Taylor’s categories also included crime against property; moreover, previous studies have modified this instrument according to characteristics of their offender sample (Stompe, Ortwein-Swoboda, & Schanda, 2004). In the current study, we adapted the tool according to the level of physical contact and inflicted harm in each sexual offense. Taylor’s item ‘Completely non-violent (0)’ sexual offense was equated to indecent exposure, voyeurism or online offenses with verbal sexual assault. ‘Minimal violence (1)’ was assigned to threat with sexual motivation and minimal physical contact such as unwanted touching or fondling. ‘Moderate violence (2)’ was associated with nonconsensual penetration or other sexual intercourse, display of a weapon that was not used and physical force leading to no or minimal bodily injury. ‘Moderately serious violence (3)’ was assigned to coercive sexual intercourse causing grievous physical injury requiring medical attention. ‘Serious violence (4)’ was assigned to sexual assault that resulted in the victim(s)’ death or endangered the life of the victim(s) and required a hospital care of more than 24 hours. Using the comprehensive data collected from the participants directly and from multiple other sources, two raters, who were blind to each other’s ratings, scored each offense according to this classification. The intra-class correlation was .88 (p < .001), and any inconsistencies in the ratings were resolved through consensus.

2.3. Statistical analysis

Statistical analyses were performed using the IBM Statistical Package for the Social Sciences (SPSS) for Mac OS, Version 23.0 software (Armonk, NY: IBM Corp.). Descriptive data including background information and offense-related characteristics were analyzed. The association between these data and violence severity of the index offense was tested with Spearman’s rank-order correlation. Pearson’s correlation test was used to evaluate associations between violence of the index offense, dark triad traits, empathy and impulsiveness levels. Multivariate ordinal regression analysis was used to determine the potential background and offense-related characteristics and personality dimensions that could predict the violence level of the index offense. Statistical significance was accepted as p <.05.

3. Results

The 64 perpetrators were classified according to the level of physical violence of the index offense as completely non-violent (n = 12), minimally violent (n = 17), moderately violent (n = 25), moderately seriously violent (n = 8) and seriously violent (n = 2). All participants were male, and their mean age was 37.08 ± 11.89 years (range = 19–71), while education in years was 8.64 ± 4.25. Among the participants, 71.9% were unmarried/single, and 57.8% were unemployed. The victim was a stranger in 39.1%, and victim was underage in 57.1% of the offenses. According to Spearman’s correlation test, being unemployed (r = −.25, p < .05), and perpetrating a stranger victim (r = −.50, p < .001) were inversely correlated with greater level of violence in the index offense. Perpetrating an underage victim (r = .28, p < .05) and committing the offense with a co-offender (r = .27, p < .05) were directly correlated with greater level of violence in the index offense. The remaining background and offense-related characteristics are presented in Table 1.

Table 1.

Background and offense-related characteristics of the sample.

Background characteristics	All offenders (n = 64)
	Mean ± SD/n(%)	r
Age (range 19–71 years)	37.08 ± 11.89	.01
Education (range 0–16 years)	8.64 ± 4.25	−.16
Unmarried	46 (71.9)	−.04
Unemployed/irregular work	37 (57.8)	−.25*
Alcohol use disorder	38 (59.4)	.14
Substance use disorder	27 (42.2)	.06
Age at first offense (range 13–71 years)	30.84 ± 13.37	.04
Total convictions (range 1–20)	2.69 ± 3.15	−.04
Type of offenses (range 1–6)	1.72 ± 1.13	−.05
Previous interpersonal offense	27 (42.2)	.02
Previous violent offense	20 (31.3)	.20
Previous sexual offense	11 (17.2)	.05
Characteristics of the index offense
Stranger victim	25 (39.1)	−.50**
Underage victim	37 (57.8)	.28*
Co-offender	2 (3.1)	.27*
Intoxicated state	11 (16.9)	.09
Violence index
Completely non-violent	12 (18.8)
Minimal violence	17 (26.6)
Moderate violence	25 (39.1)
Moderately serious violence	8 (12.5)
Serious violence	2 (3.1)

Note: r = Spearman’s rank-order correlation coefficient (with violence index).

*p<.05. **p<.001.

Correlation analyses (Table 2) indicated that violence index was correlated only with secondary psychopathy (r = .35, p < .05). EQ was negatively correlated with other psychometric scale scores (r = −.27 to −.45). BIS–11 was positively correlated with dark triad traits (r = .27–.45). Machiavellianism was positively correlated with both primary (r = .42, p < .05) and secondary (r = .41, p < .05), but not total psychopathy (r = .18, p > .05). Narcissism was positively correlated with total psychopathy (r = .26, p < .05) and Machiavellianism (r = .37, p < .05).

Table 2.

Correlations between violence index and personality traits.

Note: DDS = Dirty Dozen Scale; Mach = Machiavellianism; EQ = Empathy Quotient; BIS–11 = Barratt Impulsiveness Scale. Number in squares refers to the Pearson’s r correlation coefficient; p < .05 for all squares that are not blank. *p < .05, **p < .01. Red colors refer to negative correlation, blue colors refer to positive correlation.

In the multivariate ordinal regression analyses (Table 3) non-stranger victim (β = 2.448, p < .001), secondary psychopathy (β =  0.061, p < .05), narcissism (β = 0.145, p < .05) and EQ (β = 0.047, p < .05) could significantly predict that a sexual offense would involve increased levels of violence, χ² (138.197) = 41.185, p < .001 (Nagelkerke pseudo R² =.505). The variance inflation factor (VIF) for each predictor was below 5, which is a threshold value without critical collinearity issues (Hair, Risher, Sarstedt, & Ringle, 2019).

Table 3.

Ordinal multivariate regression analysis of offense-related characteristics and personality traits for the level of violence involved.

	β	S E	Wald	p	[95% CI]
Age	0.008	0.024	0.105	.746	[−0.039,0.055]
Victim (non-stranger)	2.448	0.645	14.415	<.001*	[1.184,3.712]
Victim (underage)	0.792	0.596	1.768	.184	[−0.375,1.959]
Intoxicated state	0.430	0.696	0.382	.537	[−0.935,1.795]
Psychopathy total	0.003	0.032	0.007	.933	[−0.060,0.065]
Primary	0.006	0.054	0.011	.916	[−0.101,0.112]
Secondary	0.061	0.029	4.300	.038*	[0.003,0.118]
DDS Machiavellianism	0.164	0.102	2.613	.106	[−0.035,0.364]
DDS Narcissism	0.145	0.062	5.505	.019*	[0.024,0.266]
EQ	0.047	0.024	3.994	.046*	[0.001,0.094]
BIS–11	−0.007	0.020	0.129	.720	[−0.047,0.032]

Note: Results from ordinal regression, model summary; χ²(138.197) = 41.185, p<.001, Nagelkerke pseudo R² of .505.DDS = Dirty Dozen Scale; EQ = Empathy Quotient; BIS–11 = Barratt Impulsiveness Scale; CI = confidence interval.

*p < .05 (bold values).

4. Discussion

The current study is the first to examine increments in physical violence engaged during sexual offenses and their association with dispositional as well as contextual factors. We aimed to highlight personality traits and their relationship with the level of physical violence employed during a sexual offense in a sample of sex offenders without any serious mental illness. The findings from our study reveal that greater secondary psychopathy, narcissism and empathy and having a non-stranger victim were closely linked to increments in physical violence used during sexual offense. Moreover, as expected, empathic competencies were negatively correlated with the dark triad traits of personality.

More than half of the sample were unemployed in the current study, supporting data from previous research (Arbanas, Marinović & Buzina, 2022; Valença, Meyer, Freire, Mendlowicz, & Nardi, 2015). The available data on the relationship between employment status and the severity of violence involved in an offense is inconclusive. While one study comparing violent and non-violent offenders did not report any significant differences in terms of employment status (Caravaca-Sánchez, Fear, & Vaughn, 2019), Mercado and Scalora (2001) reported that sex offenders committing intensely violent offenses were more likely to be employed than perpetrators who committed less violent offenses (Mercado & Scalora, 2001). Therefore, there is a need for a closer understanding of the association between occupational functionality of the offender and violence level of the conducted offense, including sexual offenses.

In the current study, less than two fifths of the sexual assault victims were strangers. Nation-wide and population-level prevalence studies have reported that only one fourth of sexual victimizations were committed by strangers (Arbanas, Marinović & Buzina, 2022; Breiding, 2015; Peterson, Liu, Merrick, Basile, & Simon, 2021). Our findings and the current evidence indicate that, contrary to popular belief, the majority of victims are sexually assaulted by someone they know. Multivariate analysis in the current study indicated that a non-stranger victim was associated with more violent sexual assault. Similarly, about 90% of the victims of the most serious and violent sexual offenses were reported to know the perpetrator; however, this rate fell to below 50% for less serious or minor sexual offenses (Howitt, 2018).

In the current sample, the victim was underage in about 58% of the sexual offenses. Although we found that child victims tend to undergo sexual offenses with higher violence, a multivariate analysis suggested that the victim being underage was not significantly associated with the level of violence used during a sexual offense. Child molesters are believed to display relatively low level of violence and aggressive behavior compared to rapists or other adult sex offenders (Browne, Hines, & Tully, 2018; Mercado & Scalora, 2001). Understanding how the violence severity relates to child molestation may help in predicting recidivism as there is evidence to suggest that among sex offenders, the level of force correlates with the risk of re-offending (Schopp, Pearce, & Scalora, 1998).

Secondary psychopathy and narcissism were associated with greater level of violence employed during sex offense in the current study. Dark triad traits, particularly psychopathy and narcissism, were reported to be related to perpetration of sexual aggression (Mouilso & Calhoun, 2012). Psychopathic traits are considered to be an integral part and a key construct of sexual violence (Cale, Lussier, McCuish, & Corrado, 2015). High psychopathy, established with either clinician-based or self-report tools, was reported to be predictive of more severe sexual assault and sexual aggression (Barbaree, Seto, Serin, Amos, & Preston, 1994; DeGue, DiLillo, & Scalora, 2010; Krupp, Sewall, Lalumière, Sheriff, & Harris, 2012). Moreover, severe psychopathic traits are reliable predictors of general, violent and sexual recidivism (Declercq, Willemsen, Audenaert, & Verhaeghe, 2012; Hare & Neumann, 2008; Kirsch & Becker, 2007). Denials, rationalizations, distortions, and minimization are the cognitive distortions related to psychopathy that are the norm with sex offenders (Geddes & Andreasen, 2020). Due to such distortions, sex offenders do not know how to anticipate humane primary needs and goods through prosocial means and hence follow a criminal path to attain them (Ward, 2002). The current line of evidence suggests that higher levels of global psychopathy was associated with more violent sexual offenses than non-violent sexual offenses; on the other hand, the difference was more prominent in the antisocial/lifestyle dimension of psychopathy for sexual murderers than for other sex offenders (Firestone, Bradford, Greenberg, & Larose, 1998; Koch et al., 2011; Langevin, 2003). The absolute level of violence could be primarily explained by the antisocial/secondary psychopathy factor (Hare & Neumann, 2009). However, in general, no firm conclusions have yet been drawn regarding which dimensions of psychopathy are closely linked to extreme physical force used during sexual offense (Gretton et al., 2001; Porter et al., 2009). On the other hand, in line with our findings, previous research suggests that secondary psychopathy encompassed an antisocial pathway that was characterized by an escalation of the violent offending trajectory (Cale et al., 2015; Mills, Anderson, & Kroner, 2004), and narcissism was associated with increased severity of sexual offenses (Parks & Bard, 2006; Pavlović et al., 2019). Offenders with greater narcissism may be willing to resort to sexual coercive behavior when they are denied sexual access to individuals they desire, explaining the nexus between narcissism and sexual coercion (Zeigler-Hill et al., 2016). Sexual sadism has also been associated with secondary psychopathic traits (Barbaree et al., 1994; Robertson & Knight, 2014). Secondary psychopathy reflects an unstable and antisocial lifestyle and includes features that are predominantly behavioral in nature. This pattern encompasses sensation-seeking and unsocial behaviors that can lead to violent offending (Kirsch & Becker, 2007).

Impulsiveness is a defining feature of psychopathy and entails rapid, spontaneous and maladaptive behavior that is related to various offenses and aggression (Neumann, Schmitt, Carter, Embley, & Hare, 2012). Findings on whether high impulsiveness was more attributable to an increased risk of sexual offending rather than general offending are currently inconclusive (Baltieri & de Andrade, 2008a; Joyal, Beaulieu-Plante, & de Chantérac, 2014; Parry & Lindsay, 2003). Trait impulsiveness was particularly associated with antisocial/secondary dimension of psychopathy in the current study, corroborating previous research (De Tribolet-Hardy, Vohs, Mokros, & Habermeyer, 2014; Morgan, Gray, & Snowden, 2011). However, it is yet to be confirmed whether trait impulsiveness is directly linked with level of violence engaged during sexual offense.

Lack of empathy can be considered as a hallmark of psychopathy since empathy deficits in psychopathic individuals may contribute to their offending behavior (Nitschke, Istrefi, Osterheider, & Mokros, 2012). Lack of sensitivity or awareness of other people’s emotions may impair an individual’s ability to understand the effects of their behavior on others (Gery, Miljkovitch, Berthoz, & Soussignan, 2009; Kirsch & Becker, 2007). The current study suggests that dark triad traits including psychopathy and its subdimensions were significantly associated with low empathy, as expected. On the other hand, empathy, as a complex construct, has not been clearly attested to be related to the severity of sexual offenses. Indeed, some authors have argued that the role of empathy in sexual offending may be best understood in terms of contextual factors (Smallbone et al., 2003). Moreover, empathy deficit should not be conceptualized as the sole reason of sexual offending but is better comprehended as a symptom of one or more of the stable dispositions that have been established as risk factors for offending (Barnett & Mann, 2013; Tierney & McCabe, 2001). Decreased empathy skills may lead to deficits in the ability to empathize with others, resulting in an increased likelihood for perpetrating offending and violent behavior (Kirsch & Becker, 2007). Previous studies suggest that sex offenders with a greater lack of empathic concern appear to use more physical engagement and aggression than those without such deficits (Elliott, Beech, & Mandeville-Norden, 2013; Lisak & Ivan, 1995; Porter et al., 2000). On the contrary, Nitschke et al. reported that emotional empathy was similar between sadistic and non-sadistic sex offenders (Nitschke et al., 2012). Furthermore, other studies have reported no differences in general empathy between rapists and other sex offenders (Hayashino, Wurtele, & Klebe, 1995; Langevin, Wright, & Handy, 1988). Our findings indicate that lower empathy was associated with lower levels of violence employed during sexual offending behavior. In this regard, empathy, which needs to be further examined in relation to the severity of specific sexual assault, is a personality construct that is highly influenced by internal and external situational factors. Empathy is linked to the offending pattern rather than determining the characteristics of a sex offense. Thus, the relationship between empathy and sexual offending is not as clear-cut as it is generally believed to be.

Motivation and violence involved in sexual offenses may depend not only on personality traits but also on specific situational and contextual factors (Mann, Hanson, & Thornton, 2010). Some factors, such as preference for use of a weapon (Felson & Messner, 1996), intoxication in the offender (Beauregard & Mieczkowski, 2012; Bechtel & Holstege, 2007) or presence of a co-offender (Aebi, Vogt, Plattner, Steinhausen, & Bessler, 2012), increases the perpetrator’s aggression and severity of the physical injury to the victim in sexual offenses. Other factors, such as the victim coming from a criminogenic environment, decrease the likelihood of involvement of violence and lethal outcomes (Mieczkowski & Beauregard, 2010). However, studies that reported the above arguments suffered from small sample sizes. These studies also show that violent and non-violent (lethal vs. non-lethal) sex offenders present more similarities than differences in their offender characteristics (Proulx, Beauregard, Cusson, & Nicole, 2007). In the current study, the proportion of offenses in which a co-offender was present was similar to previous reports showing that more than one offender was present in only 3–11% of cases (Arbanas, Marinović & Buzina, 2022; Valença et al., 2015). The use of alcohol and other drugs at the time of the offense has been reported in up to onethird of all offenders charged with sexual assault (Arbanas, Marinović & Buzina, 2020; Baltieri & de Andrade, 2008b). This rate was lower in the current sample, perhaps due to demographic factors. Moreover, we found that a state of intoxication was not directly related to the level of violence displayed. Koch et al. (2011) argued that alcohol or substance use may initially increase sexual arousal and disrupt inhibition. However, later on there is an increased probability of a frustrated offender killing the victim during a sexual assault (Koch et al., 2011). We posit that this implication may not be valid when considering violence as a continuum.

Since the current study focused on dimensional personality traits as dispositional factors related to sexual offending and their relationship with the level of violence employed, we did not consider whether any paraphilic disorders existed in the subjects. Indeed, a diagnosis of paraphilia is not sufficient to understand the pathways to sexual offending. Recent studies examining traits related to sexual offenses have omitted diagnoses of paraphilia (Bismpas et al., 2020; Brewer, Lyons, Perry, & O’Brien, 2021; Gillespie, Mitchell, Beech, & Rotshtein, 2021; Navas, Maneiro, Cutrín, Gómez-Fraguela, & Sobral, 2022; Reynolds, Maurer, Cook, Harenski, & Kiehl, 2020). This is due to an asymmetric relationship between paraphilia and sexual offending and a reluctance to medicalize sexual offenses (Geddes & Andreasen, 2020). Since most sex offenders are male, although a small number of women may commit similar crimes (Colson, Boyer, Baumstarck, & Loundou, 2013), we could not include female subjects in the current study; this precluded an examination of gender differences on the association between personality traits and violence involved in a sex offense. Aggressive subtypes (reactive/proactive) were also not considered. Reliance on self-report questionnaires for capturing personality traits particularly among forensic populations requires cautious interpretation. There may be a difficulty in validating dark triad domains due to a considerable construct and item overlap. File-based review of crime-scene data may have precluded us from obtaining the real crime-scene-related information. The coding system of violence index used may be inherently subjective. The small-sized sample of offenders used in the current study should not be considered representative of the total sex offender population.

5. Conclusions

Despite possible limitations, the current research has important implications in improving our understanding of sexual violence. Taken together, our findings suggest that it is possible to offer a refined conceptualization of sexual violence in the form of a continuum. Considering such a continuum, offenders with high secondary psychopathy and narcissism may show a preference for sexual assault that is more violent in nature. We also reckon that dispositional factors such as personality traits and their combination with circumstantial factors may influence the level of physical violence engaged during a sexual assault. Moreover, the results from our study imply that a violent sexual assault is a complex event governed by complex dynamics that shift the odds of using severe violence in predictable and understandable ways, beyond the offenders’ characteristics. It is possible that particular personality traits of an offender may influence complex associations and increase the likelihood of intense violence used during a sexual assault. Therefore, our study shows the value of parsing the most severe offenses and their perpetrators into a conceptual continuum. The heterogeneity in the offending population is lost when offenders are evaluated on the basis of additional violence inherent to their instant sexual offense. Therefore, we believe that risk assessment tools should be tailored to specific physical violence profiles of the previous sexual offenses in offender groups. A more informed risk stratification through assessing relevant personality traits and violence profiling of previous sexual offenses is of clinical and legal interest since such a differential assessment of sex offenders would enable determining suitable interventions for this population. It is important to scrutinize the dark side of personality; establishment of an integrated framework consisting of these core clusters of individual dispositions that are associated with the level of physical violence may aid in the development of risk identification tools, improve legal decision making and can be the focus of intervention in sex offender treatment programs. We believe that this study will stimulate further inquiry that can extend beyond a research perspective shaped by legal or vague classifications of violence. Instead, it is necessary to address psychological variations in crime scene actions and refine the actual heterogeneity of sex offenders more effectively.

Ethical standards

Declaration of conflicts of interest

Yasin Hasan Balcioglu has declared no conflicts of interest

Mehmet Dogan has declared no conflicts of interest

Ipek Inci has declared no conflicts of interest

Abdulkadir Tabo has declared no conflicts of interest

Mustafa Solmaz has declared no conflicts of interest

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee [Ministry of Justice of Turkey, Council of Forensic Medicine, IRB: 10.12.2019 – 21589509/2019/1005] and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent

Written informed consent was obtained from all individual participants included in the study, following a thorough explanation of the study procedure.