Background: Although sadistic child sexual abuse (SCSA) is rare, existing typologies of child sexual abuse have identified the existence of sadistic subtypes in child molestation, as well in the sexual homicide of children. Nonetheless, no study has sought to determine whether there is heterogeneity in the manifestation of sexual sadism between sadistic child abusers.
Objective: The present study seeks to examine how SCSA manifests differently between offenders, and whether these differences are associated with specific victim, offender, and offense characteristics.
Participants and Setting: The current sample includes adult males (N = 101) who were involved in a child sexual assault and scored at least a 4 on the Severe Sexual Sadism Scale (SeSas). All offenses take place in France between 1990 and 2018.
Methods: Latent class analysis was used to analyze the sample for heterogeneity. Bivariate analyses were conducted to identify external variables associated with each of the latent classes.
Results: Three distinct latent classes were found: the sadistic kidnapper; the sadistic torturer; and the sadistic ritualist. External validity testing also revealed distinctive characteristics associated with each class.
Conclusions: SCSA involves a heterogenous population with distinctive sadistic behavioral manifestations that vary in severity and relate to differences in crime-commission processes and offender characteristics. These findings offer important insights for crime prevention and correctional practice.
Keywords: Latent class; sexual sadism; behavioral indicators; SeSas; child victims
It is well-documented in the extant literature that child sexual abusers are a heterogeneous population (see Lim et al., 2021 for a review). Although contact sexual offending against a child, particularly of a violent nature, remains less common than other forms of child sexual abuse (CSA) (Lim et al., 2021), the widespread harm that these types of crimes can cause to the victim (e.g., physical, medical, psychological, and social), as well as the enormous cost to society, is unequivocal (Hall & Hall, 2007; Fang et al., 2012). Sexual sadism represents the most extreme form of sexual violence, involving acts such as coercion, torture, humiliation, and the infliction of pain for sexual pleasure (Chopin & Beauregard, 2022; Dietz et al., 1990; Nitschke et al., 2013; Longpré et al., 2018). The ability to identify whether a sexual crime against a child involves sexual sadism is crucial from not only an intervention standpoint but also for public safety. For example, studies have shown that there is an association between sexual sadism and the most serious crimes, such as sexual homicide (e.g., Brittain, 1970; Healey et al., 2013; Mokros, 2018; Ressler et al., 1986; Ressler et al., 1988). Moreover, deviant sexual fantasies, which are central to the diagnosis of sexual sadism (5th ed.; DSM-5; American Psychiatric Association, 2013), are a risk factor for sexual reoffending (Brankey et al., 2021) and sexual preoccupation (Hanson et al., 2007).
More recently, researchers have also shown that sexual sadism is not homogenous, but rather, can manifest into different behavioral subtypes that relate to distinct individual and crime characteristics (Chopin & Beauregard, 2022). This approach suggests that sexual sadism can vary in the degree of severity (Longpré et al., 2018; 2019; 2020), in the behavioral manifestations, as well as in sadistic sexual fantasies (Chopin & Beauregard, 2021a; 2021b). Although sadistic child sexual assault (SCSA) is rare, existing typologies of child sexual abusers have also identified the existence of sadistic subtypes in child molestation (Knight & Prentky, 1989), as well in the sexual homicide of children (Chopin & Beauregard, 2019). Nonetheless, there is a lack of research that specifically examines the heterogeneity in the manifestation of sexual sadism among individuals who commit SCSA. It is of the utmost importance to identify heterogen6 eity within CSA populations in order to propose clinical approaches adapted to the specificity of these individuals, and ultimately, to help prevent future CSA.
Sexual sadism has been included in the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5; American Psychiatric Association, 2013) since the mid-twentieth century. It is defined as “pleasure and sexual arousal that is rooted in fantasized or actual infliction of psychological (including humiliation) or physical suffering on a victim. Since the DSM-III (American Psychiatric Association, 1980), the diagnosis requires that fantasies or behaviors must be severe, recurrent, and last for a minimum period of 6 months. In addition, the behavior must be directed toward non-consenting partners, or the sexual urges or fantasies must cause marked distress or interpersonal difficulty. Nonetheless, a major problem in the study of sexual sadism involves the criteria that are used to define it. Agreement around the number of required behavior criteria and the severity of the cut-off point has been a major concern and subject of contentious debate (see Longpré et al., 2018 for a review). For example, several studies have noted inconsistencies in the criteria used, particularly regarding the criteria deemed to be the “core” characteristics or essential markers for an accurate assessment of sexual sadism (e.g., Marshall et al., 2002a; Marshall et al., 2002b).
As a result of these issues, several authors have argued that sadism would be better conceptualized using a dimensional approach. This perspective suggests that sexual sadism should be defined and assessed based on severity (low, moderate, high), rather than the presence or absence of the disorder, to offer an alternative or complimentary tool to the existing categorical system used in diagnostic manuals (Berner et al., 2003; Longpré et al., 2018; Nitschke et al., 2013). In line with a dimensional perspective, researchers have developed psychometric assessment instruments for sexual sadism composed of behavioral indicators that can be rated based on crime-scene information (e.g., MTCSS; Longpré et al., 2019; SADSEX-SH; Myers et al., 2019; SeSaS; Nitschke et al., 2009). One of the most widely validated and reliable tools is the Severe Sexual Sadism Scale (SeSaS; Nitshke et al., 2009). This 11-item behavioral scale uses a cut-off score of 4 to differentiate severe sadists from non-sadists and has been shown to have the potential to improve the validity and reliability of the diagnosis of sexual sadism (Gonçalves et al., 2020). Studies have also shown that sadistic crime scene behaviors are part of a continuum of severity, with certain acts associated with more or less serious crimes. For example, both victim mutilation and foreign object insertion are associated with the most serious form of sexual violence (i.e., sadistic sexual homicide; Longpré et al., 2020; Mokros et al., 2012; Nitschke et al., 2009; Reale et al., 2017).
More recently, researchers have also sought to contribute to the empirical validation of the dimensionality of sexual sadism by using crime scene indicators to explore the behavioral manifestations of sexual sadism. These studies have demonstrated that there are important differences not only in the manifestations of sexual sadism, but also in how these differences relate to specific victim, offender, and offense characteristics. For example, Reale et al. (2017) examined the variations in sexual sadism in a sample of homicide offenders. Using a crime scene rating scale for sadism in sexual homicide (SAD-SEX-SH-R; Myers et al., 2019) they found that behavioral manifestations of sadism varied in degree (severe, mixed, and non-sadists). Moreover, the severe sadistic subgroup was also related to the use of detection avoidance strategies (e.g., selecting low-risk locations and destroying and removing evidence).
Chopin et al. (2021a) also examined the heterogeneity of sexual sadism in sexual homicide, but they focused exclusively on individuals who met the cut-off criteria for behavioral indicators of sexual sadism using the SAD-SEX-SH-R (Meyers et al., 2019). Results showed that sexual sadism manifested into four distinct patterns (1) the anal/oral sex offender class; (2) inanimate object insertion class; (3) the collection class; and (4) torture/mutilation. Although there were few situational factors that related to each class, unique offender characteristics were associated to the sadism manifestations. For example, the torture/mutilation class was the most likely to target children. Although this study was the first to examine the heterogeneity of sexual sadism within a sample of classified sexual sadists, this approach has not yet been extended to SCSA specifically.
“Author removed for blind review”, however, provides some evidence to suggest that sexual sadism is likely to manifest differently between SCSA. In this study, Multidimensional Scale Analysis (MDS) of the SeSas (Nitschke et al., 2009) was performed on two samples of sexual sadists: those with adult victims and those with child victims. Results from “Author removed for blind review” provided preliminary evidence that there are different dimensional manifestations of sexual sadism depending on whether the offense involves an adult or child victim. In particular, a complex interplay between the manifestation of sadistic sexual fantasies and pedophilic fantasies for SCSAs was evident. Despite the importance of these findings, “Author removed for blind review” focused mainly on dimensional interpretations between different factors. Thus, more research is needed that focuses exclusively on the heterogeneity of sadism within CSA and how this relates to other crime, victim, and offender characteristics to offer a more pragmatic understanding of the phenomenon.
Although there is a lack of typological research on SCSA, the extant literature has consistently shown that child sexual abusers represent a widely heterogenous group (see Lim et al., 2021 for a review). For example, using data from the Massachusetts treatment centre (MTC), Knight and Prentky (1989) provided one of the most comprehensive typologies of child molesters (CMs). Within this classification system, two different sadistic child molestation subtypes emerged. The sadistic type- high injury, is characterized by sexual arousal or pleasure from placing the victim in pain or fear. Behavioural indicators for this subtype include the use of violence to facilitate sexual arousal or ritualized, bizarre, and unusual acts as well as acts of aggressive sodomy or object insertion. In the sadistic type- muted class, an offender is classified based on the presence of object insertion without injury, reported sadistic fantasies or behavioral evidence of sadistic fantasies (e.g., bondage, humiliation, or bizarre and unusual acts), or sodomy.
Using a different approach, Lanning (2010) classified CMs on a motivational continuum (situational to preferential). On the situational side, CMs are characterized by being involved in various unlawful behaviors, consuming violent pornographic materials, being more impulsive, and engaging in either spontaneous or planned sexual crimes. On the other end, the preferential CMs are involved in specific or focused criminal behavior, consume specific types of pornographic materials, they are more compulsive, they engage in ritualized behaviors, are driven by fantasy, and are more consumed by their needs rather than the risk of the offense. The latter subtype shares several over-lapping characteristics with traditional profiles of the sadistic offender, particularly in relation to fantasy driven as well as ritualized behavior (Dietz et al., 1990; Nitschke et al., 2013; Longpré et al., 2019; Ressler et al., 1988).
Lastly, sexual sadistic behavior has also been observed as a distinct subtype in the sexual homicide of children. Chopin & Beauregard (2019) proposed the first classification of the sexual murder of children. Within the six classes identified, two could be characterized as sadistic. More specifically, the Intentional/Teen Class were characterized by acts of sadism (e.g., asphyxiation, strangulation, unusual act) and penetration of the victim. In these sexual homicides, the offender and victim were strangers, and the offender was highly socially isolated. They mostly approached their victim using a con strategy in a residential location. Additionally, in another class (Intentional/Preteen) offenders were characterized by a high prevalence of drug and alcohol use, as well as the diversity of sexual acts committed and the presence of sexual sadism (e.g., high level of violence and degrading acts). They also use a con to approach their victim, but do not typically commit their crime in a residence. Taken together, these findings not only indicate that child sexual abusers are a heterogenous population but that some of this heterogeneity is explained, at least in part, by differences in the manifestation of sexual sadism (Knight & Prentky, 1989; Chopin & Beauregard, 2019).
Prior literature has demonstrated the heterogeneity of both sexual sadism (e.g., Chopin et al., 2021a; Reale et al., 2017) and CSA (Chopin et al., Knight & Prentky, 1989; Lanning, 2010), as well as the presence of distinct types of sadistic offenders within CSA typologies (e.g., Chopin & Beauregard, 2019; Knight & Prensky, 1989). However, there is a lack of research that focuses exclusively on the heterogeneity within SCSA. Prior research on the heterogeneity of SCSA has either included mixed samples with adult and child victims (Reale et al., 2017), included non-sadistic child sexual abuse (Knight & Prentky, 1989; Lanning, 2010) or focused on sexual homicide exclusively (Chopin et al., 2021a). Therefore, the current study aims to empirically examine (1) how SCSA manifests differently between offenders, and (2) identify whether these differences are associated with specific victim, offender, and offense characteristics. The ability to differentiate between individuals who commit SCSA using behavioral indicators can be used to help improve the assessment of sexual sadism as well as help enhance the efficacy of clinical decision-making about treatment and management (Chopin et al., 2021a; 2021b; Knight & Prentky, 1989; Nitschke et al., 2013; Marshall & Hucker, 2006). For example, a better understanding of the phenomenon could help to strengthen treatment programs (e.g., cognitive behavioral therapy) based on controlling specific sadistic fantasies and behaviors (Hamilton & Rosen, 2016). This, in turn, could help to prevent future CSA from occurring.
Data used in this study comes from a larger national database from France. The 735 cases included in this database occurred between 1990 and 2018, and all involve male perpetrators (age range = 16-59) who (1) committed sexual assault and (2) scored at least a 4 on the SeSas to indicate the presence of sexual sadism (Mokros et al., 2012; Nitschke et al., 2009). For the current study, we included only the cases that involved child sexual assault (i.e.,15 years old and less, see Chopin & Beauregard, 2020; Leclerc et al., 2011) to be included in our sample (N = 101). This database includes information on offender, victim, and crime characteristics and data are derived from various sources of information. To avoid missing data, information is compiled by a team of crime analyst experts on violent crimes. Although it is still possible to have missing values as the information may not always be known, this was not the case with the variables examined in this study. For each case, the information comes from investigative reports, interview reports, medical/autopsy reports provided by pathologists, psychological reports provided by a team of forensic psychologists, and crime scene forensic reports. This study has been approved by the University Ethics Review Board.
Main model. In order to test the heterogeneity of the manifestation of sexual sadism in sexual assaults involving children, we used a total of 11 dichotomous variables corresponding to the dimensions of the Sexual Sadism Scale (SeSaS) developed by Nitschke et al., (2009). This scale was chosen to assess sexual sadism because previous studies indicated that it was strongly related to the clinical diagnosis (Longpré et al., 2018; Mokros et al., 2012; Nitschke et al., 2013; Nitschke et al., 2009). According to most studies using the SeSaS to assess sexual sadism (e.g., Chopin & Beauregard, 2021b; Chopin et al., 2022; Gonçalves et al., 2020), we used only Part 1 which is composed of 11 items: (1) individual is sexually aroused by the act, (2) individual exercises power/control/domination over the victim; (3) individual humiliates or degrades the victim, (4) individual tortures the victim or engages in acts of cruelty, (5) individual mutilates sexual parts of the victim’s body (mutilations were committed ante mortem and none was committed only after the death of the victim, but they may have led to the death of the victim and be continued after the victim's death); (6) individual engages in gratuitous violence toward the victim, (7) individual keeps trophies of the victim, (8) individual mutilates nonsexual parts of the victim’s body, (9) victim is abducted or confined, (10) evidence of ritualism in the offense, (11) insertion of objects into victim’s body orifices. Total scores can range from 0 to a maximum of 11; with a cut-off score of 4 or above indicating the probable presence of sexual sadism (Nitschke et al., 2009). Specifically, individuals involved in sadistic sexual crimes of children present an average score of 4.51 (SD=0.76). We observed a good reliability for the SeSaS items used to assess sexual sadism with a Cronbach α = 0.84.
Covariates. In order to test the external validity of the main model we used several covariates. The first variable we used as a covariate is 1) the average SeSaS score [x̄ =4.51, SD = 0.76; range 4-7]. In addition to this variable, we used three set of variables to test whether patterns of sexual sadism are associated with offender and crime characteristics as well as target selection.
Several previous studies suggested that offender characteristics are associated with certain manifestations of sadism (Brittain, 1970; Chopin & Beauregard, 2021a, 2021b, 2021c; Reale et al., 2017; Ressler et al., 1986). Moreover, some studies (Proulx, 2001; Proulx et al., 2014; Proulx et al., 2007) report that sexual sadism is related to certain developmental, psychological, and criminological variables (for example, modus operandi). In order to test this assumption with cases involving child victims we used the following variables: (2) Offender age [x̄ =31.16; SD = 11.68], (3) offender was single, (4) offender had an active social life (i.e., participates in social situations and attends events where other people, including acquaintances and strangers, gather), (5) offender had problems with alcohol and/or drugs (i.e., diagnosis or reporting of a persistent alcohol/drugs abuse disorder), (6) offender avoided social contact with other (i.e., a loner lifestyle with few social interactions), (7) offender had previous criminal convictions (i.e., general recidivism), (8) offender experienced sexual dysfunction (i.e., erectile and/or ejaculatory dysfunctions), (9) offender possessed a sexual collection (i.e., sadistic and/or pedophilic videos/pictures), (10) offender had pedophilic sexual interest (i.e., based on previous diagnoses of pedophilia or admission of specific sexual interest in children).
Target selection variables were based on previous studies suggesting that sadistic sexual offenders were more likely to select victim in vulnerable positions (Douglas et al., 2013; Hazelwood & Douglas, 1980). Moreover, recent studies highlighted an association between manifestations of sadism and victim selection characteristics (Chopin & Beauregard, 2021a, 2021c). In order to test this assumption with child victims, we used the following variables: (11) victim age [x̄ =11.26; SD=3.71], (12) victim and offenders were strangers (i.e., they did not know each other at the time of the offense), (13) victim was a female, (14) victim was specifically targeted (i.e., victim specifically targeted by the offender for his specific characteristics), (15) victim was involved in home activities (e.g., watching TV, cooking, sleeping), (16) victim was traveling from one point to another (e.g., walking or hitchhiking), (17) victim was involved in social activities (e.g., visiting someone, partying).
Finally, we used a set of eight variables to examine whether the manifestation of sadistic fantasies influenced the crime-commission process characteristics in cases involving child victims. Such differences were identified in previous studies focusing on adult victims (Chopin et al., 2022; Douglas et al., 2013; Hazelwood & Douglas, 1980). The variables we used are the following: (18) offender used con approach (i.e., as opposed to a coercive blitz approach strategy; e.g., befriended the victim, posed as an authority figure, offered assistance), (19) offender perpetrated sexual penetration (i.e., vaginal and/or anal penetration), (20) offender forced the victim to perform a fellatio, (21) offender perpetrated a digital penetration (i.e., vaginal and/or anal), (22) offender masturbated himself, (23) diversity of sexual acts committed by the offender (i.e., number of sexual committed by the offender; x̄ =2.86, SD=1.44), (24) offender destroyed forensic evidence (i.e., used specific strategies to avoid police detection), (25) offender killed the victim.
The first step of this study was to identify whether the different dimension of the SeSaS would allow to identify heterogeneity among a sample of cases involving sexual sadism against child victims. A latent class analysis (LCA) using Latent Gold V6.0 software package was conducted to identify patterns in the manifestation of sadistic fantasies. LCA is a statistical procedure that allows for the identification of heterogeneity that is not directly observable or measurable in order to detect underlying patterns in a set of data or subgroups of individuals who share important behavioral characteristics (Collins & Lanza, 2010), as it is the case with the sample of cases used in this study. The goal is to identify mutually exclusive classes (i.e., nonoverlapping) on the basis of dichotomous variables (Collins & Lanza, 2010; Lanza et al., 2007; Lanza et al., 2003). We computed seven models from one-to-seven solutions (see Table 1). The Bayesian Information Criterion (BIC), Akaike Information Criterion (AIC; Akaike, 1974) as well as the Vuong-Lo-Mendell-Rubin Adjusted likelihood ratio test (LMRT; Lo, Mendell, Rubin, 2001) was used to evaluate the model fit and determine the number of classes to use in LCA. A lower BIC value indicates an improvement in the fit of models (Schwartz, 1978).
As a second step, we used additional variables (i.e., offender, crime, target selection characteristics) to test the external validity of the model as well as to improve its depth. Bivariate analyses (i.e., Chi-square analysis and Kruskal-Wallis test1) were used to identify significant differences between the different classes. Such a procedure is commonly used by researcher using the LCA procedure (Brownfield & Sorenson, 1987; Vaughn et al., 2008). This procedure makes it possible to distinguish whether the differences identified in the main classification model and based on a limited number of indicators are reflected in other indicators qualifying the studied phenomenon.
Table 1 describes the model fit indices for the LCA models. To assess the best latent class model, one-to-seven solutions were computed. BIC and AIC fit indicators were used to identify the best class solution. Dziak et al. (2012) noted that the AIC is more likely to identify a bigger model than BIC when the sample size is large, while BIC is more likely to select an inappropriate model when the sample size limited. Due to the limited sample size we used in this research (N=101) (see e.g., Tein et al., 2013; VanVoorhis & Morgan, 2007), BIC is constantly increasing and does not provide an accurate determination of the best class solution. For this research, the AIC is more useful, and the smallest value suggests that the trade-off between fit and parsimony was achieved. The smallest AIC (1063.88) suggested that the 3-class solution was the best fitting solution. Entropy for the 3-class solution is .94, suggesting that predictors we used to operationalize script dimensions are fair to classify the cases and that classes are almost perfectly distinct (entropy score of 1 indicate a perfect delineation of classes; see Celeux & Soromenho, 1996). The Vuong-Lo-Mendell-Rubin Adjusted likelihood ratio test indicates that the three-class model significantly improved upon the fit of the two-class model.
[Insert Table 1 here]
Table 2 describes the 3-class solution representing the three patterns of manifestations of sadism present in cases involving child victims. The largest class corresponds to the class 1 (59.41% of cases), while the smallest is class 3 (13.86 % of cases).
In class 1, the sadistic behaviors are more likely to be characterized by humiliation/degradation of victim (1.002), the sexual arousal of acts perpetrated (0.83), and the abduction/confinement of victims (0.83).
In class 2, the sadistic behaviors are manifested with the use of gratuitous violence toward or wounding of the victim (0.81), the humiliation/degradation of victim (0.89), and the torturing of the victim and engaging in acts of cruelty (0.73).
In class 3, sadistic fantasies are manifested through the humiliation/degradation of the victim (1.00), evidence of ritualism (0.91), and the insertion of foreign objects (0.92).
[Insert Table 2 here]
Offenders with an active social life (χ2 = 19.82, p = <.001, φ = 0.44) were more likely to be included in class 3. Offenders who presented problems of alcohol/drugs consumption (χ2 = 7.26, p = .027, φ = 0.27) and those who presented previous criminal convictions (χ2 = 12.52, p = .002, φ = 0.35) were more likely to be included in class 2. Offenders who were single at the time of the assault (χ2 = 9.88, p = .01, φ = 0.31) and those with sexual dysfunction (χ2 = 7.51, p = .023, φ = 0.27), with pedophilic sexual interests (χ2 = 11.15, p = .004, φ = 0.33) and with a sexual collection (χ2 = 11.66, p = .01, φ = 0.34) were more likely to be included in class 3.
[Insert Table 3 here]
As to the target selection, findings indicated that cases with younger victims (χ2 = 7.51, p = <.001, df = 2, ω2= 0.10) were more likely to be included in class 3. Cases where offenders and victims did not know each other at the time of the assault (χ2 = 10.22, p = .006, φ = 0.32), and those where victims were traveling (χ2 = 9.66, p = .008, φ = 0.31) were more likely to be included in class 2. Offenders whose victims were involved in home activities (χ2 = 16.76, p = .008, φ = 0.31) were more likely to be involved in class 3. As to the crime characteristics, offenders included in class 3 were more likely to use a con approach (χ2 = 6.23, p = .044, φ = 0.25), and to commit digital penetration (χ2 = 7.67, p = .022, φ = 0.28). Offenders included in class 2 were more likely to destroy forensic evidence (χ2 = 6.33, p = .042, φ = 0.25) and to kill their victims (χ2 = 6.95, p = .031, φ = 0.26) while those included in class 1 were more likely to target their victims (χ2 = 9.28, p = .010, φ = 0.30), force their victims to perform fellatio (χ2 = 6.69, p = .035, φ = 0.26). Finally, offenders in class 3 were more likely to perform a greater number of different sexual acts (χ2 = 7.64, p = .022, df = 2, ω2= 0.06).
[Insert Table 4 here]
Sexual sadism has been found to have different behavioral manifestations depending on the sample under analysis (e.g., Chopin & Beauregard, 2021a; 2021b; Reale et al., 2017), and prior research has shown that CSA is committed by a heterogenous population (e.g., Lim et al., 2021). Despite the seriousness of sexual sadism, the immense psychological and physical harm caused to the victim, and the danger that failing to identify a sadistic offender poses for public safety (e.g., Marshall & Hucker, 2006), there is a lack of research examining the heterogeneity of SCSA. The current study therefore sought to examine two unaddressed areas in the extant literature (1) how sexual SCSA manifests differently between offenders, and (2) identify whether these differences are associated with specific victim, offender, and offense characteristics. By focusing exclusively on the SCSA, distinctive latent subgroups related to the manifestation of sexual sadism are observed. These subgroups are mainly different from prior research examining the heterogeneity of sexual sadism in mixed victim age samples (e.g., Chopin et al. 2021a) as well as existing CSA typologies that include sadistic subgroups (e.g., Chopin & Beauregard, 2019; Knight & Prentky, 1989; Lanning, 2010), however, some significant overlapping characteristics are also observed. It is also important to note that consistent with previous research (e.g., Healey et al., 2013; Marshall & Hucker, 2006; Nitschke et al., 2009; 2013) humiliation and degradation of the victim seems to be hallmark of SCSA, as this behavior was consistently associated with all three classes.
Class 1: The Sadistic Kidnapper. In this subgroup sexual sadism is mainly manifested through humiliation and degradation for sexual purposes. This is consistent with studies that have described a core feature of sexual sadism as being sexual aroused by violent or humiliating behavior (e.g., Abel, 1989, Groth & Birnhaum, 1979). In terms of victim characteristics, offenders in this class were more likely to target their victims. Interestingly, Healey et al. (2013) observed that there may be two different types of sexual sadists, the sadistic murderer and the sadistic sexual aggressor. The sadistic sexual aggressor is more likely to humiliate and seek out specific victim characteristics compared to sadistic murders, but equally likely to kidnap or confine their victim. Healey et al. (2013) suggested that sadistic sexual aggressors may get more pleasure from verbal aggression that involves humiliation to obtain sexual gratification, whereas sexual murderers might need extreme physical violence (such as mutilation) to obtain sexual gratification. This seems to be consistent with offenders in this profile, which notably, do not typically engage in the more severe forms of violent sadistic behaviors, such as torture, sexual mutilation, or foreign object insertion. Offenders in this class also have the lowest scores on the SeSaS relative to other classes. This suggests that individuals included in the Sadistic Kidnapper Class represent a less extreme form of severe sadism compared to other classes, which is consistent with a dimensional interpretation of sexual sadism along a continuum of severity (Longpré et al., 2018). Taken together, offenders in this class are least likely to kill their victims, relative to the other classes.
Class 2: The Sadistic Torturer. Offenders in this subgroup are characterized by sexual sadism that is manifested through torture, humiliation/degradation, cruelty, and gratuitous violence or wounding of the victim. Relative to the other classes, offenders in this subgroup had the highest average score on the SeSaS, were the most likely to kill their victims and destroy and remove forensic evidence. As Healey et al. (2013) observed, individuals who commit sadistic sexual murders might need extreme physical violence to obtain sexual gratification, which is consistent with the behavioral manifestations of sexual sadism in Sadistic Torturer Class. Moreover, this subgroup overlaps considerably with the severe sadistic subgroup identified by Reale et al. (2017), suggesting that individuals included in the Sadistic Torturer Class fall onto the more extreme end of severe sadism. This is congruent with previous studies suggesting that sadistic manifestations of torture and mutilation were related to the most serious offenses (Chopin et al., 2022). In terms of offender characteristics, this subgroup was the most likely to have previous criminal convictions and have problems with alcohol or drug consumption. This is similar to Chopin & Beauregard’s (2019) Intentional/Preteen class of sexual homicide which was characterized by sexual sadism and a high prevalence of alcohol and drug use. The presence of alcohol consumption problems is a commonly observed characteristic of sexual sadists in most studies (e.g., Dietz et al., 1990; Hazelwood et al., 1992; Meloy, 2000; Warren et al., 1996).
Cases included in the Sadistic Torturer Class were also characterized by the fact that the victims were most likely strangers to the offenders and traveled to or from a location prior to the crime. This is the same modus operandi observed in studies of sadistic homicide offenders on adult victims (Reale et al., 2017) and child victims (Chopin & Beauregard, 2022). This suggests that offenders in this subgroup target their victims based on vulnerability and opportunity. This behavior is consistent with the notion of "premeditated opportunism" (Rossmo, 2000), in which the offender is prepared to commit their crime, but likely did not choose a specific time or victim beforehand. Rather, they plan in their fantasies so that they are ready to act when the convergence of a suitable victim and a low-risk situation presents itself. Additionally, nearly half of the offenders in this subgroup exhibit a pedophilic preference for children, which is consistent with Lanning's (2010, p.33) continuum of motivation for pedophilia. Offenders are preferential characterized by being involved in specific illicit behaviors, being more compulsive, and committing scripted sex crimes related to sexual fantasies (Lanning, 2010). Overall, offenders in this subgroup appear to share characteristics with preferential CMs (Lanning et al., 2010), and Chopin et al.'s (2019) Intentional/Preteen Class sadistic sexual homicide perpetrators who actualize their world of sadistic sexual fantasies in the commission of their sexual crime, which are characterized by high levels of sexual sadism.
Class 3: The Sadistic Ritualist. In this class, sadistic behaviors are manifested through the humiliation/degradation of the victim, evidence of ritualism, and the insertion of foreign objects. Additionally, offenders in this class engaged in digital penetration and a high number of sexual acts. These crime scene behaviors are consistent with behaviors that are typically considered to be core features of sexual sadism (e.g., Beauregard et al., 2020; Dietz et al., 1990). More specifically, foreign object insertion is a behavior that is associated serious crimes, including sadistic sexual homicide (e.g., Longpré et al., 2020; Reale et al., 2017) and is perpetrated as a way to torture the victim (Dietz et al., 1999). Ritualism is also a distinctive characteristic of sadism (e.g., Hucker, 2009; Marshall & Hucker, 2006) and is typically used to re-experience the crime and relive sadistic fantasies (Dietz et al., 1990; Hazelwood et al., 1992). Similar behaviors were also observed by Knight and Prensky (1989), who identified a sadistic type-high injury class in their sample of CMs. These offenders were similarly characterized by the infliction of pain on their victim, engaging in ritualized acts as well as foreign object insertion.
Despite these similarities, there are also several distinctive victim and offender characteristics observed in the current class, compared to the sadistic subgroups identified in Knight and Prentky’s (1989) classification of CMs. More specifically, included in the sadistic ritualistic class were more likely to have pedophilic sexual interests and younger victims. Notably, this presents a like profile to CMs who are “preferentially motivated” and engage in ritualized behaviors, are driven by fantasy, and are consumed more by their needs (i.e., pedophilic urges) rather than the risk of their offense (Lanning, 2010). Additionally, offenders in this class were more likely to target their victims while they were involved in home activities, using a con approach. This is the same crime-commission process that Chopin & Beauregard (2019) identified in the Intentional/Teen Class of sexual homicide against children. Moreover, the overlapping paraphilias (i.e., pedophilia and sexual sadism) suggest that offenders in Class 3 represent a significant recidivism risk. For example, compared to non-pedophiles, pedophiles tend to have more victims, respond more poorly to treatment, and are more likely to reoffend (Seto, 2008). Furthermore, having a variety of paraphilic interests have a significant impact on recidivism rates among offenders (Hanson et al., 2007). Finally, as we have seen, the offenders in this class are the most distinctive in terms of sexuality: among the three classes, they have the highest rates of pedophilia, sexual collection, sexual dysfunction, and diversity in sexual acts committed during the offence. Taken together, this subgroup therefore represents a particularly important target for clinical intervention.
The ability to differentiate between SCSA using behavioral indicators can be used to help enhance the efficacy of clinical decision-making about treatment and management (e.g., Chopin et al., 2021a; 2021b). Through the identification of three distinct classes among SCSA, findings from the current study also reinforce the need for treatment needs and intervention strategies to help prevent future offenses. For example, the sadistic ritualist class presents with concurrent paraphilias (i.e., pedophilia and sexual sadism) that are both risk factors for recidivism (Seto, 2008; Hanson et al., 2007). Offenders in this class may represent an especially high-risk and dangerous subgroup that requires intensive clinical intervention to prevent from further offending and offense escalation. Moreover, it is likely that the presence of severe sadism and pedophilia will present a clinical challenge due to major deficits in the ability to form consensual adult sexual relationships. This subgroup would likely require tailored treatment approaches that target their deviant sexual fantasies and behaviors and decrease the level of distress associated with their urges and fantasies (Hamilton & Rosen, 2016). They may also benefit from supervision orders aimed at reducing their contact with children specifically.
The sadistic torturer class is characterized by their opportunistic and situationally-drive offending. Considering that this class was also associated with previous criminal convictions as well as significant drug and alcohol problems, these offenders would likely require tailored intervention strategies that target a variety of criminogenic risk factors, in addition to treatment for sexual sadistic urges and fantasies. The ability for clinicians to recognize these offenders at an early stage of their criminal career is imperative considering that this subgroup was the most likely to commit sexual homicide. Lastly, the sadistic kidnapping class also presents with distinctive treatment and intervention needs. In particular, the lack of overtly violent behavioral manifestations of sexual sadism in this subgroup suggest that they are more sexually aroused by acts of humiliation and degradation of the victim. Clinicians may be able to use this information to help tailor intervention approaches that target these specific sexual fantasies and urges to help improve treatment outcomes. Future research should continue to investigate differences in the manifestation of SCSA on cross-cultural samples, as well as determine whether there are differences between solved and unsolved SCSA to further validate findings from the present study.
Although the current study has important practical applications, it is not without methodological limitations. Consistent with all studies using police data, biases may occur in terms of data validity and reliability (Aebi, 2006). It is also important to keep in mind that only cases reported to authorities have been considered. However, considering the level of violence used against a stranger child victim, and the fact that some of these cases resulted in homicidal outcomes, it is likely that the dark figure is much lower for these cases than would be expected in other types of sexual assaults. Additionally, it is important to note that our study only focuses on solved cases, and it is possible that there are different patterns in unsolved cases of SCSA. Thus, findings from the present study may not be generalized to unsolved offenses. We also used the SeSaS scale to identify sadistic cases on the basis of crime scene behaviors, however, it is possible that some sadistic cases were unidentified by this tool.
The issue of whether an offender meets criteria for sexual sadism has serious implications for decision-makers in the criminal justice system (Marshall & Hucker, 2006). For example, failure by clinicians to identify a true sadist might result in the offender’s release from custody when he remains a significant threat to the community. Moreover, considering the extreme levels of violence associated with these offenses, and the widespread and long-term physical and psychological harm that child sexual abuse causes both to victims and society (Hall & Hall., 2007; Fang et al., 2012; Lim et al., 2021), SCSA represent a major public safety concern. Of considerable issue, however, is that current diagnostic tools (e.g., DSM-5) have been criticized for lacking validity as well as having poor interrater reliability (e.g., Marshall & Hucker, 2006; Nitschke et al., 2013). The use of behavioral crime scene indicators can therefore help to improve the diagnosis and assessment of sexual sadism (Longpré et al., 2018; Nitschke et al., 2009; Marshall & Hucker, 2006). In line with this approach, the current study provides additional evidence to demonstrate the utility of crime scene indicators to not only assess sexual sadism, but also to identify differences in the manifestation of sexual sadism between SCSAs. Through this process, we identified three distinctive classes of SCSA that are consistent with the notion that sexual sadism can vary in degree of severity (Longpré et al., 2018) and in behavioral manifestations of sadistic fantasies (Chopin & Beauregard, 2021a; 2021b). Thus, findings indicate that SCSAs are a heterogenous population, which has important implications for the assessment of sexual sadism as well as for intervention and management strategies for CSA.
Aebi, M.F. (2006). How to measure deliquency? Armand Colin.
Abel, G. G. (1989). Paraphilias. In Kaplan, H. I., Sadock, B. J. (Eds.) Comprehensive textbook of psychiatry (Vol. 1, 5th ed., pp. 1069-1089). Williams & Williams.
American Psychiatric Association. (1980). Diagnostic and Statistical Manual of Mental Disorders (3rd ed.). American Psychiatric Publishing.
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). American Psychiatric Publishing.
Beauregard, E., & Proulx, J. (2002). Profiles in the offending process of nonserial sexual murderers. International Journal of Offender Therapy and Comparative Criminology, 46(4), 386–399. https://doi.org/10.1177/0306624X02464002
Berner, W., Berger, P., & Hill, A. (2003). Sexual sadism. International Journal of Therapy and Comparative Criminology, 47(4), 383–395. https://doi.org/10.1177/030662402236739
Brankley, A. E., Babchishin, K. M., & Hanson, R. K. (2021). STABLE-2007 demonstrates
predictive and incremental validity in assessing risk-relevant propensities for sexual offending: A meta-analysis. Sexual Abuse, 33(1), 34–62. https://doi.org/10.1177/1079063219871572
Brittain, R. P. (1970). The sadistic murderer. Medicine, Science and the Law, 10(4), 198–207. https://doi. org/10.1177/002580247001000402
Brownfield, D., & Sorenson, A. M. (1987). A latent structure analysis of delinquency. Journal of Quantitative Criminology, 3(2), 103-124. https://doi.org/10.1007/BF01064211
Celeux, G., & Soromenho, G. (1996). An entropy criterion for assessing the number of clusters in a mixture model. Journal of Classification, 13(2), 195-212.
Chopin, J., & Beauregard, E. (2019). Sexual homicide of children: A new classification. International Journal of Offender Therapy and Comparative Criminology, 63(9), 1681–1704. https://doi.org/10.1177/0306624X19834419
Chopin, J. & Beauregard, E. (2021a). Manifestation of sadism in sexual homicide: a criminological contribution. Psychology, Crime & Law. https://doi.org/10.1080/1068316X.2021.2018442
Chopin, J., & Beauregard, E. (2021b). Sadistic juvenile offenders: Identifying clinical intervention dimensions based on sadism manifestations and criminogenic factors. The Journal of Forensic Psychiatry & Psychology, 32(6), 792–807. https://doi.org/10.1080/14789949.2021.1909105
Chopin, J., & Beauregard, E. (2021c). Typology of sadistic sexual offenders: A latent class analysis of sadism manifestations and their correlates. Criminologie, Advance online publication.
Chopin, J., & Beauregard, E. (2022). Sexual sadism: Its role in the crime-commission process of sexual homicide of children. Journal of Interpersonal Violence, 37(1–2), NP350– NP374. https://doi.org/10.1177/0886260520916844
Chopin, J., Beauregard, E., & Dietz, P. E. (2022). Death in sadistic sexual crimes: A neural network analysis of factors associated with a lethal outcome. Criminal Justice and Behavior, Advance online publication. https://doi.org/10.1177/00938548211066911
Collins, L. M., & Lanza, S. T. (2010). Latent class and latent transition analysis: With applications in the social, behavioral, and health sciences. Wiley.
Dietz, P. E., Hazelwood, R. R., & Warren, J. (1990). The sexually sadistic criminal and his offenses. Journal of the American Academy of Psychiatry and the Law Online, 18(2), 163–178.
Dziak, J. J., Coffman, D. L., Lanza, S. T., & Runze, L. (2012). Sensitivity and specificity of information criteria. Technical Report Series #12-119, College of Health and Human Development, The Pennsylvania State University.
Gonçalves, L. C., Rossegger, A., Gerth, J., Singh, J. P., & Endrass, J. (2020). Sexual sadism among sex offenders in Switzerland. Sexual Abuse, 32(1), 79–100. https://doi.org/10.1177/1079063218800473
Groth, A. N., Birnbaum, H. J. (1979). Men who rape: The psychology of the offender. Basic Books.
Fang, X., Brown, D. S., Florence, C. S., & Mercy, J. A. (2012). The economic burden of child maltreatment in the United States and implications for prevention. Child Abuse & Neglect, 36(2), 156–165. https://doi:10.1016/j.chiabu.2011.10.006
Hamilton, D. V., & Rosen, J. (2016). Sexual sadism disorder. In R. Balon (Ed.), Practical guide to paraphilia and paraphilic disorders (pp. 123–139). Springer International Publishing. https://doi.org/10.1007/978-3-319-42650-1_9
Hanson, K., Harris, A. J. R., Scott, T. L., & Helmus, L. (2007). Assessing the risk of sexual offenders on community supervision. The Dynamic Supervision Project. Ottawa, Ontario, Canada: Public Safety.
Healey, J., Lussier, P., & Beauregard, E. (2013). Sexual sadism in the context of rape and sexual homicide: An examination of crime scene indicators. International Journal of Offender Therapy and Comparative Criminology, 57(4), 402–424. https://doi.org/10.1177/0306624X1243753
Hall, R.C.W., & Hall, R.CW. (2007). A profile of pedophilia: definition, characteristics of offenders, recidivism, treatment outcomes, and forensic issues. Mayo Clinic Proceedings, 82(4), 457–471. https://doi.org/10.4065/82.4.457
Hamilton, & Rosen, J. (2016). Sexual Sadism Disorder. In Practical guide to paraphilia and paraphilic disorders (pp. 123–139). Springer International Publishing. https://doi.org/10.1007/978-3-319-42650-1_9
Hazelwood, R. R., Dietz, P. E., Warren, J. (1992). The criminal sexual sadist. FBI Law Enforcement Bulletin, 61(2), 12-20.
Hazelwood, R. R., & Douglas, J. (1980). The lust murderer. FBI Law Enforcement Bulletin, 49, 1–5.
Kingston, D. A., Seto, M. C., Firestone, P., & Bradford, J. M. (2010). Comparing indicators of sexual sadism as predictors of recidivism among adult male sexual offenders. Journal of Consulting and Clinical Psychology, 78(4), 574–584. doi:10.1037/a0019734
Knight, Carter, D. L., & Prentky, R. A. (1989). A system for the classification of child molesters. Journal of Interpersonal Violence, 4(1), 3–23. https://doi.org/10.1177/088626089004001001
Lanning, K. V. (2010). Child molesters: A behavioral analysis for professional investigating the sexual exploitation of children (5th ed.). National Center for Missing & Exploited Children.
Lanza, S. T., Collins, L. M., Lemmon, D. R., & Schafer, J. L. (2007). PROC LCA: A SAS procedure for latent class analysis. Structural Equation Modeling: A Multidisciplinary Journal, 14(4), 671-694. https://doi.org/10.1080/10705510701575602
Lanza, S. T., Flaherty, B. P., & Collins, L. M. (2003). Latent class and latent transition analysis. In J. A. J. Schinka & W. F. Velicer (Eds.), Handbook of psychology: Vol. 2 Research methods in psychology (pp. 663-685). Wiley.
Leclerc, B., Wortley, R., & Smallbone, S. (2011). Getting into the script of adult child sex offenders and mapping out situational prevention measures. Journal of Research in Crime and Delinquency, 48(2), 209-237. https://doi.org/10.1177/0022427810391540
Lim, Y. Y., Wahab, S., Kumar, J., Ibrahim, F., & Kamaluddin, M. R. (2021). Typologies and psychological profiles of child sexual abusers: An Extensive Review. Children, 8(5), 333. MDPI AG. http://dx.doi.org/10.3390/children8050333
Longpré, N., Guay, J. P., Knight, R. A., & Benbouriche, M. (2018). Sadistic offender or sexual sadism? Taxometric evidence for a dimensional structure of sexual sadism. Archives of Sexual Behavior, 47(2), 403–416. https://doi.org/10.1007/s10508-017-1068-4
Longpré, N., Guay, J.-P., & Knight, R. A. (2019). MTC sadism scale: Toward a dimensional assessment of severe sexual sadism with behavioral markers. Assessment, 26(1), 70–84. https://doi.org/10. 1177/1073191117737377
Longpré, N., Sims-Knight, J. E., Neumann, C., Guay, J.-P., & Knight, R. A. (2020). Is paraphilic coercion a different construct from sadism or the lower end of an agonistic continuum? Journal of Criminal Justice, 71, 1–11. https://doi.org/10.1016/j.jcrimjus.2020.101743
Longpré, N., Proulx, J., & Brouillette-Alarie, S. (2018). Convergent validity of three measures of sexual sadism: Value of a dimensional measure. Sexual Abuse, 30(2), 192- 208. https://doi.org/10.1177/1079063216649592
Marshall, W. L., & Hucker, S. J. (2006). Issues in the diagnosis of sexual sadism. Sexual Offender Treatment, 1(2), 1–4.
Marshall, W. L., Kennedy, P., & Yates, P. (2002a). Issues concerning the reliability and validity of the diagnosis of sexual sadism applied in prison settings. Sexual Abuse, 14(4), 301– 311. https://doi.org/10.1177/107906320201400402
Marshall, W. L., Kennedy, P., Yates, P., & Serran, G. (2002b). Diagnosing sexual sadism in sexual offenders: Reliability across diagnosticians. International Journal of Offender Therapy and Comparative Criminology, 46(6), 668–677.
Meloy, J. R. (2000). The nature and dynamics of sexual homicide: An integrative review. Aggression and Violent Behavior, 5(1), 1–22. https://doi.org/10.1016/ S1359- 1789(99)00006-3
Mokros, A., Schilling, F., Eher, R., & Nitschke, J. (2012). The severe sexual sadism scale: cross- validation and scale properties. Psychological assessment, 24(3), 764–769. https://doi.org/10.1037/a0026419
Mokros, A. (2018). Sexual sadism and sexual homicide: An overview. In J. Proulx, E. Beauregard, A. J. Carter, A. Mokros, R. Darjee, & J. James (Eds.), Routledge international handbook of sexual homicide studies (pp. 253–268). Routledge.
Myers, W. C., Beauregard, E., & Menard, W. (2019). An updated sexual homicide crime scene rating scale for sexual sadism (SADSEX-SH). International Journal of Offender Therapy and Comparative Criminology, 63, 1766–1775.
Nitschke, J., Mokros, A., Osterheider, M., & Marshall, W. L. (2013). Sexual sadism: Current diagnostic vagueness and the benefit of behavioral definitions. International Journal of Offender Therapy and Comparative Criminology, 57(12), 1441–1453.
Nitschke, J., Osterheider, M., & Mokros, A. (2009). A cumulative scale of severe sexual sadism. Sexual Abuse, 21(3), 262–278
Proulx, J., Beauregard, E., Lussier, P., & Leclerc, B. (2014). Pathways in the offending process of extrafamilial sexual aggressors against women. Pathways to sexual aggression, 71-109.
Proulx, J., Blais, E., & Beauregard, E. (2007). Sadistic sexual offenders. In J. Proulx, M. Cusson, E. Beauregard, & A. Nicole (Eds.), Sexual murderers: A comparative analysis and new perspective (pp. 107–122). Chichester, UK: Wiley.
Proulx, J. (2001). Sexual preference and personality disorders of MTC:R3 rapist types. Paper presented at the 19th conference of the Association for the Treatment of Sexual Offenders (ATSA), San Antonio, TX.
Reale, K., Beauregard, E., & Martineau, M. (2017). Sadism in sexual homicide offenders: Identifying distinct groups. Journal of Criminal Psychology, 7(2), 120–133. https://doi.org/10.1108/JCP-11- 2016-0042
Richards, H., Jackson, L. R. (2011). Behavioral discriminators of sexual sadism and paraphilia nonconsent in a sample of civilly committed sexual offenders. International Journal of Offender Therapy and Comparative Criminology, 55, 207-227.
Ressler, R. K., Burgess, A. W., & Douglas, J. (1988). Sexual homicide: Patterns and motives. Free Press.
Ressler, R. K., Burgess, A. W., Hartman, C. R., Douglas, J. E., & McCormack, A. (1986). Murderers who rape and mutilate. Journal of Interpersonal Violence, 1(3), 273–287. https://doi.org/10.1177/ 088626086001003002
Rossmo, K. (2000). Geographic profiling. CRC Press
Schwartz, G. (1978). Estimating the dimension of a model. The Annals of Statistics, 461-464.
Seto, M. C. (2008). Pedophilia and sexual offending against children: Theory, assessment, and intervention. Washington, DC: American Psychological Association.
Tein, J.-Y., Coxe, S., & Cham, H. (2013). Statistical power to detect the correct number of classes in latent profile analysis. Structural Equation Modeling: A Multidisciplinary Journal, 20(4), 640-657.
VanVoorhis, C. R. W., & Morgan, B. L. (2007). Understanding power and rules of thumb for determining sample sizes. Tutorials in Quantitative Methods for Psychology, 3(2), 43-50. https://doi.org/10.20982/tqmp.03.2.p043
Vaughn, M. G., DeLisi, M., Beaver, K. M., & Howard, M. O. (2008). Toward a quantitative typology of burglars: A latent profile analysis of career offenders. Journal of Forensic Sciences, 53(6), 1387-1392.
Table 1. Fit indices for latent classes (N=101).
Nb of classes
Note: Boldface type indicates the selected model
Vuong-Lo-Mendel-Rubin likelihood ratio test not applicable for one-class model
Table 2. Profile of three latent classes - Mean probabilities of manifestations of sadism based on class membership (N=101)
Item 1. Offender engaged in gratuitous violence/wounding of the victim
Item 2. Offender exercised power/control/domination over the victim
Item 3. Offender humiliated/or degraded the victim
Item 4. Offender was sexually aroused by the act
Item 5. Offender tortured victim/engaged in acts of cruelty
Item 6. Evidence of ritualism in offense
Item 7. Victim was abducted/or confined
Item 8. Insertion of objects into victim's bodily orifice
Item 9. Offender mutilated sexual parts of the victim's body
Item 10. Offender mutilated nonsexual parts of the victim's body
Item 11. Offender kept trophies
SeSaS Score (average)
Table 3. External validity analysis of offender characteristics (N=101)
χ2 / Kruskal-Wallis H
Effect size φ / ω2
Offender age (average)
Offender is single
Offender had an active social life
Offender had problems with alcohol and/or drugs
Offender avoided social contact with other
Offender had previous criminal convictions
Offender had sexual dysfunction
Offender possessed a sexual collection
Offender had pedophilic sexual interest
Notes: *p ⩽ .05. **p ⩽ .01. ***p ⩽ .001. Pairwise comparisons: Each subscript letter denotes a subset of the 3-class model whose column proportions do not differ significantly from each other at the .05 level.
Table 4. External validity analysis of target selection and crime characteristics (N=101)
χ2 / Kruskal-Wallis H
Effect size φ / ω2
Victim age (average)
Victim and offenders were strangers
Victim was a female
Victim was targeted
Victim was involved in home activities
Victim was traveling (walking or hitchhiking)
Victim was partying
Offender used con approach
Offender perpetrated sexual penetration
Offender forced the victim to perform a fellatio
Offender perpetrated a digital penetration
Diversity of sexual acts committed by the offender (average)
Offender destroyed forensic evidence
Offender killed the victim
Notes: *p ⩽ .05. **p ⩽ .01. ***p ⩽ .001. Pairwise comparisons: Each subscript letter denotes a subset of the 3-class model whose column proportions do not differ significantly from each other at the .05 level.