Patterns of Necrophilic Behaviors in Sexual Homicide: A Criminological Perspective

This study aims to empirically explore the patterns of necrophilic behaviors in sexual homicide. More specifically, the study investigates offender, victim, and crime characteristics of sexual homicides where necrophilic acts were perpetrated, to determine whether the primary motivation to kill is associated with the attainment of corpses or whether the post-mortem sexual acts were secondary deviant behaviors. The sample used in this study consists of 109 cases of extrafamilial sexual homicides where post-mortem sexual acts were committed by offenders. Latent class analysis was used to examine each step of the crime-commission process of sexual homicide offenders involved in necrophilic behaviors. Our findings suggest that there are four different patterns of necrophilia in sexual homicide: Opportunistic, experimental, preferential, and sadistic. Preferential offenders are the only ones who specifically kill their victims in order to have sex with their corpses, while for sadistic, experimental, and opportunistic offenders post-mortem sexual acts were part of a secondary deviant process. Practical implications in terms of criminal investigations and offenders’ treatment are discussed.


Introduction
Necrophilia is defined by as sexual gratification that is attained by having sex with corpses (Aggrawal, 2010;Crow, 2017). This behavior is classified by the Diagnostic and Statistical Manual of Mental Disorders 5 (American Psychiatric Association, 2013, p. 705) and is described as a "recurrent and intense sexual arousal involving corpses". Most necrophiles do not engage in sexual intercourse with the dead and are content simply with fantasying about post-mortem sexual acts (Aggrawal, 2009b). If they decide to take action, necrophile individuals use different contexts to find bodies of people who are already dead (Aggrawal, 2009a(Aggrawal, , 2009b(Aggrawal, , 2010. For instance, some individuals steal corpses from morgues, dig up bodies from cemeteries, or retain the body of their partner after their death (Aggrawal, 2009a(Aggrawal, , 2009b(Aggrawal, , 2010Boureghda et al., 2011;Wulfen, 1910). In some cases, however, homicide may precede the necrophilic act (Aggrawal, 2009a(Aggrawal, , 2009b(Aggrawal, , 2010Rosman & Resnick, 1989;Stein et al., 2010). Studies conducted in different countries (e.g., Canada, France, and South Korea) indicated that the rate of post-mortem sexual acts in sexual homicide cases varies between 8.5% and 36.6% (Beauregard & Martineau, 2013;Darjee & Baron, 2013;James et al., 2018;Sea et al., 2019).
Previous studies have considered necrophilic sexual homicide offenders (nSHOs) as a specific category in itself (Aggrawal, 2009a(Aggrawal, , 2009bHirschfeld, 1956;Krafft-Ebing, 1886;Wulfen, 1910), while Rosman and Resnick (1989) provided a distinction based on the level of attraction to corpses, suggesting the presence of two distinct types of necrophilia. For the pseudonecrophile the sexual attraction to corpses is transient, while for the genuine necrophile this attraction is persistent (Rosman & Resnick, 1989). Although interesting and informative, previous studies have been mainly based on case reports, making the generalization of these findings difficult. In order to shed additional light on the current knowledge on necrophilia associated to sexual homicide, this study aims to empirically explore the patterns of necrophilia in sexual homicide and determine whether nSHOs represent a homogeneous type of offenders or if they are driven by distinct motivations.

Characteristics of Necrophilia in Sexual Homicide
Only a few studies have examined cases of sexual homicide including necrophilic acts and most of these studies are based on case reports (e.g., Chan et al., 2017;Krafft-Ebing, 1886;Lancaster, 1978;Pettigrew, 2019aPettigrew, , 2019bPettigrew, , 2019c. Although case studies provide unique insight into the characteristics of nSHOs, these studies are difficult to generalize and may represent the most extreme or unusual cases. As a result, we decided to focus on describing the characteristics of necrophilia among studies with samples including at least 5 cases. In terms of offender demographic characteristics, nSHOs are always males and mostly single, but often living with somebody (Rosman & Resnick, 1989;Stein et al., 2010). In their sample of 16 nSHOs, Stein et al. (2010) found that the average age was 26.1 years old, most were laborers (50%; n = 8) or unemployed, and almost all had prior arrests (92.9%; n = 15). As to their sexual lifestyle, Rosman and Resnick (1989) identified that pseudonecrophilic killers (i.e., transient attraction to a corpse) are more likely to be heterosexual compared to necrophilic killers, while other studies found that some of them may suffer from sexual dysfunctions (Chopin & Beauregard, 2020b;Hirschfeld, 1944). Rosman and Resnick (1989) found that the totality of pseudonecrophilic killers present a history of sadistic acts compared to 78% of necrophilic murderers. This relationship between sexual sadism and necrophilia has been highlighted in several other publications (Holmes, 2017;Holmes & Holmes, 2002a, 2002b, 2002cPettigrew, 2019c;Purcell & Arrigo, 2006). Rosman and Resnick (1989) noted that few necrophile killers suffer from psychosis, while 83% of them present personality disorders.
Research found that victims of nSHOs were mostly female (Rosman & Resnick, 1989;Stein et al., 2010), with an average age of 30.4 years old (Stein et al., 2010). Stein et al. (2010) found that victims of necrophilia in sexual homicide cases are single, do not live alone, while approximately half of them do not know the offender. Studies focusing on specific types of victims found that elderly victims (Chopin & Beauregard, 2020b;Safarik et al., 2002) and female victims  are more likely to be targeted by nSHOs.
As to the crime characteristics, Stein et al. (2010) found that most victims are strangled, while 12.5% (n=2) are beaten. In most cases the victim's body is left at the crime location with no effort by the offender to conceal it. Stein et al. (2010) noted that victims' bodies are moved to the crime location in 12.5% (n=2) of the cases and 43.8% (n=7) victims are found with ligatures. Ante-mortem sexual behaviors are often present, mostly vaginal and anal penetration, while postmortem mutilation, biting, and necrophagia are observed in some cases (Rosman & Resnick, 1989;Stein et al., 2010). Rosman and Resnick (1989) observed that alcohol consumption prior to the crime is common and Stein et al. (2010) highlighted that 25% (n=4) of offenders took money or possessions of value belonging to the victims.

Sexual Homicide and Necrophilia: Patterns and Motivations
Empirical studies focusing on postmortem sexual activities are scarce and for most of them, sexual homicide represents only a small proportion of cases where necrophilic acts were identified. Studies published at the end of the nineteenth century and in the first half of the twentieth century provided the first set of classification approaches. Mainly based on case reports, these studies suggested that SHOs represented a specific category of necrophiles. Krafft-Ebing (1886) was probably the first to discuss the relationship between necrophilia and sexual homicide. He provided the first classification and made the distinction between necrophiles seeking only sexual contact with the dead and those who killed and mutilated their victims (Krafft-Ebing, 1886). Wulfen (1910) identified three types of necrophile: The lust murder, the necrostuprum, and the necrophagy. In the lust murder, the offender killed to have sexual intercourse with the corpse. In the necrostuprum, offenders stole the corpse to have sexual intercourse, while for the necrophagy, offenders mutilated and ate dead body parts (Wulfen, 1910). Hirschfeld (1956), suggested that necrophiles can be divided into two categories: those who kill to obtain a corpse and those who use the corpse of a person who is already dead.
More recently, Rosman and Resnick (1989) provided a classification of necrophilic behaviors based on 88 cases from the world literature and 34 unpublished cases. The clinical classification identified made the first level of distinction between genuine necrophilia and pseudonecrophilia. The genuine necrophilia is characterized by persistent sexual attraction to corpses, while the pseudonecrophile is characterized by a transient attraction to corpses that is not the main object of their sexual fantasies (Rosman & Resnick, 1989). The genuine necrophilia cluster includes necrophilic homicide (i.e., murder to obtain a corpse for sexual purposes), regular necrophilia (i.e., the use of already dead bodies for sexual pleasure), and necrophilic fantasy (i.e., fantasizing about sexual activity with a corpse without carrying out any necrophilic acts) (Rosman & Resnick, 1989). Rosman and Resnick (1989) mentioned that the pseudonecrophile (also labeled pseudonecrophilic killer) cluster includes sadistic, opportunistic, and transitory offenders who prefer sexual contact with living partners. Rosman and Resnick (1989) noted that the most common motivation of the genuine necrophile was to possess an unresisting partner.
More recently, a ten-type theoretical classification was provided by Aggrawal (2009aAggrawal ( , 2009bAggrawal ( , 2010Aggrawal ( , 2017. The different types include the following: role players, romantic necrophiles, necrophilic fantasizers, tactile necrophile, fetishistic necrophiles, necromutilomaniacs, opportunist necrophiles, regular necrophiles, homicidal necrophile, and exclusive necrophiles (Aggrawal, 2009b). Aggrawal (2009b, p. 318) mentioned that homicidal necrophiles are the most dangerous as "they need to have sex with a dead body to the extent that they will resort to killing". Aggrawal (2010) provided a six-cluster theoretical classification specific to the homicidal necrophiles. This classification was based on acts committed on living victims (torture/mutilation and sex) and dead (mutilation and sex). The first class (IXa) is uniquely characterized by the presence of mutilation while the victim is dead. The second class (IXb) is uniquely characterized by the presence of sex activities with the victims' body. The third class (IXc) is characterized by the presence of both mutilation and postmortem sex activities (i.e., could suggest ante mortem mutilation). The fourth class (IXd) is characterized by the presence of mutilation or sex with the living victim and the presence of postmortem mutilations or post-mortem sexual activities. The fifth class (IXe) includes cases characterized by the presence of mutilation and/or sex with the living victim and the presence of postmortem mutilations and/or post-mortem sexual activities. Finally, the sixth class (IXf) combine all the possibilities (i.e., torture/mutilation and sex with living victim and postmortem mutilation and post-mortem sexual activities) (Aggrawal, 2010). Despite being the only typology to focus specifically on sexual homicide, this model has not been tested empirically and its capacity to generalize to all cases is unknown. Moreover, the information used to identify this typology is limited and fails to understand the necrophiliacs' motivations. Stein et al. (2010) provided the only empirical study to date that focuses exclusively on necrophilia and sexual homicide. Using a sample of 211 sexual homicide cases, they identified 16 with evidence of necrophilia. They suggested that in specific cases of sexual homicide, the need to have an unresisting and unrejecting partner is not necessarily applicable as most of the necrophilic offenders included in their sample were not single at the time of the offense (Stein et al., 2010). Instead, they assumed that nSHOs included in their sample were closer to the pseudonecrophilic group described by Rosman and Resnick (1989) as none of them killed in order to obtain a corpse. They found that some offenders rationalized both the homicide and the sexual postmortem acts as a consequence of being intoxicated, while others mentioned that the primarily intention was homicide but they engaged in necrophilia because the murder was not sufficiently satisfying (Stein et al., 2010).
Interestingly, recent studies have introduced the concept of Post Mortem Sexual Interference Offenders (PMSIOs) (Carter et al., 2008;Higgs et al., 2017;. This category of offender includes "homicide offenders whose offense contained at least one of the following characteristics, the perpetrator disclosed that he had sexually assaulted the victim after killing them, there was evidence from a pathologist of post mortem sexual behavior, the perpetrator had disclosed post mortem sexual behavior, there was evidence of sex with an unconscious or dead victim or the perpetrator disclosed since conviction that they had sexually assaulted the victim after killing them" (Carter et al., 2008, p. 173). PMSIOs is a large group of offenders who perpetrated postmortem sexual activities (perpetrator disclosed he had sexually assaulted the victim after killing, evidence from pathologist of post mortem sexual behavior, perpetrator disclosed post mortem sexual behavior, evidence of sex with unconscious or dead victim, perpetrator disclosed since conviction he had sexually assaulted the victims after killing them) without being necessarily "true" necrophiles (i.e., persistent sexual attraction to corpses) (Carter et al., 2008;Higgs et al., 2017). One of the underlying ideas behind the use of the concept of PMSIOs is that the use of the term necrophilia refers to a diagnosis established in accordance with the DSM 5 criteria. In several studies the concept of necrophilia is improperly used because it has not been clearly diagnosed. There could be differences between an individual diagnosed with a necrophilic paraphilia and an individual who has committed sexual acts postmortem. We decided not to use the PMSIOs terminology because the sample we used in this research presents more homogeneous cases (i.e., systematic presence of postmortem sexual penetration) than the one used by Higgs and his colleagues (i.e., included one or both of the following (i.e., included one or both of the following: (a) sexual acts or (b) sexual mutilation, see Higgs et al., 2015Higgs et al., , 2017.
In order to avoid the pitfall associated with the use of the term necrophilia, which refers to a diagnostic, we decided to use the term "necrophilic behavior". It refers to a behavior associated with necrophilia (e.g., postmortem sexual acts) but without having to meet the diagnostic criteria.

Aim of Study
Despite the existence of a few studies looking specifically at necrophilic behaviors in sexual homicide, this unusual behavior is still not well-understood. Theoretical and clinical typologies often considered necrophilic behaviors occurring in sexual homicide as a category in itself (Aggrawal, 2009a(Aggrawal, , 2009bHirschfeld, 1956;Krafft-Ebing, 1886;Wulfen, 1910). We believe that it is important for investigative practice (e.g., crime-commission process reconstitution, offender profile identification) as well as offender management and treatment (e.g., identification of the appropriate treatment and intervention program) to improve our understanding of the offenders' motivations as well as determine whether sexual homicide offenders (SHOs) involved in necrophilic behaviors constitute a homogeneous group of offenders as suggested by most studies. The current study investigates offender, victim, and crime characteristics of sexual homicides where necrophilic acts were perpetrated to determine whether the primary motivation to kill is associated with the attainment of corpses or whether the post-mortem sexual acts were secondary deviant behaviors.

Sample
The sample used in this study was extracted from the Sexual Homicide International Database (SHIelD; see Chopin & Beauregard, 2021 for a complete description of the database methodology). This database includes solved and unsolved extrafamilial SH that occurred between 1948 and 2018 in France and Canada. To identify the sexual motivation of the homicide, the FBI definition by Ressler et al. (1988) was used. This definition states that to be considered as sexual, a homicide should include a least one of the following elements: (a) victim's attire or lack of attire, (b) exposure of the sexual parts of the victims body, (c) sexual positioning of the body, (d) insertion of foreign objects into the victim's cavities, (e) evidence of sexual intercourse (oral, anal, or vaginal), and (f) evidence of substitute sexual activity, interest, or sadistic fantasy. This definition has been criticized to potentially present false positive (see e.g., Kerr et al., 2013;Stefanska et al., 2016) and consequently, all cases include in SHIelD presented at least two criteria of the FBI's definition to be considered sexual (see  for a complete review on this question). The data included in the database comes from various sources of information and were compiled by a team of crime analysts' experts in violent crimes. In each case, the information comes from investigative reports, offenders interview reports, autopsy reports provided by pathologists, psychological reports provided by a team of forensic psychologists, and reports provided by forensic experts. All this information is compiled by a unit of criminal analysts who enter the information in a database on the basis of a standardized grid. The presence of postmortem sexual activity is confirmed in this database both by autopsy reports and offenders' interviews. Nevertheless, it is possible that the presence of postmortem sexual activity was not identified by coroners or not confirmed by offenders' interviews and that cases were consequently not included in our sample.
For the current study we selected 109 solved sexual homicide cases where postmortem sexual acts were committed by offenders. The choice to select only solved case was made in order to utilize the offender's demographic and lifestyle information.

Measures
A total of 40 dichotomous variables (coded as 0, 1) were used to determine patterns of necrophilic behaviors in sexual homicide. Some of these variables were selected to develop the main classification model. These variables are those that were identified in previous studies and that allow to differentiate necrophilic behaviors of SHOs. Other variables are used to test the external validity of the model and are variables identified as important in the overall understanding of the criminal event.
Main Model. Previous studies have found that marital status and sexual dysfunctions are important information to understand the presence of necrophilic behaviors in sexual homicide (Chopin & Beauregard, 2020b;Hirschfeld, 1944;Rosman & Resnick, 1989;Stein et al., 2010). They also identified that the analysis of ante-mortem and post-mortem sexual and non-sexual behaviors were important to determine the level of necrophilic motivations (Aggrawal, 2010;Rosman & Resnick, 1989;Stein et al., 2010). Following these findings, we used ten dichotomous variables that represented these various aspects of necrophilic sexual homicides to the main model of our study: (1) offender was single, (2) offender reported sexual dysfunctions (i.e., unable to obtain an erection, unable to sustain an erection, conditional ejaculation), (3) offender perpetrated ante-mortem vaginal penetration, (4) offender perpetrated ante-mortem anal penetration, (5) offender perpetrated ante-mortem foreign object insertion, (6) offender mutilated victims' genitals, (7) method of killing: asphyxiation/strangulation, (8) offender used restraints, (9) victim was specifically targeted, 10) offender robbed items during crime.
Additional Variables. Complementary variables were used to provide a more comprehensive understanding of the main model. These variables add information related to offender, victim, and crime characteristics.
We used a total of 26 dichotomous variables to describe the crime characteristics. This set of variables was divided two subcategories of variables: Target selection and modus operandi characteristics.
Target selection. Previous studies have shown that victims' gender, age, and lifestyle were related to the post-mortem sexual acts (Chopin & Beauregard, 2020bRosman & Resnick, 1989;Safarik et al., 2002). We used a total of nine dichotomous variables to examine victim characteristics: (1) offender and victims were strangers (i.e., totally unknown at the time of the offense, (2) victim was female, (3) victim aged 65 years or more, (4) victim lived with somebody, (5) victim used alcohol/drugs prior to crime, (6) victim was assaulted while involved in domestic activities (e.g., watching TV, cooking, and cleaning home, etc.), (7) victim was assaulted while sleeping, (8) victim was assaulted while jogging, (9) victim was assaulted after he/she had been drinking at a bar.

Analytical Strategy
The analytical strategy used in this study followed a two-step process. First, a latent class analysis (LCA) was computed to identify patterns associated with necrophilic behaviors in sexual homicide. LCA has been described as a statistical procedure used to identify 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). The goal of this procedure is to identify mutually exclusive classes using dichotomous variables (Collins & Lanza, 2010;Lanza et al., 2003Lanza et al., , 2007. LCA analysis is similar to cluster analysis but provide stronger models as it attributes class membership probabilities to each individual case. Seven models were computed and analyzed from a one-to-seven class solution (Table  1). Several fit measures were used to identify the model: Bayesian Information Criterion (BIC), log likelihood, likelihood ratio L 2 , degrees of freedom, Akaike Information Criterion (AIC) and entropy. We have tested for multicollinearity and no correlations were higher than .33 (Appendix section).
Second, we used additional variables describing offender, victim, and crime characteristics to test the external validity of the latent class solution and improve the depth of our model. Specifically, we used bivariate analysis (i.e., chi-square analysis 2 ) to identify significant differences between the different classes.

Latent Class Analysis
To determine patterns of necrophilic behaviors in sexual homicide, we used 10 dichotomous variables related to offender, victim, and crime characteristics. 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 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=109) (see e.g., Tein et al., 2013;VanVoorhis & Morgan, 2007), BIC is constantly increasing and is useless to determinate the best class solution. For the research, the AIC is more useful, and the smallest value suggests that the trade-off between fit and parsimony was achieved. Table 1 describes the model fit indices for the latent classes. The smallest AIC (1225.5535) suggested that the 4-class solution was the best fitting solution. Entropy for the 4-class solution was high (0.79) and suggested that predictors used are fit to classify the cases and that classes were sufficiently distinct (Schwartz, 1978). Table 2 and Figure 1 describe the 4-class solution representing the four different patterns followed by SHOs who perpetrated post-mortem sexual activities. The largest class corresponds to the class 1 including 32.87% of cases, while the smallest is class 3 including 16.03% of cases.
In class 1, labeled opportunist, offenders were more likely to be single (0.72) 3 and did not report any sexual dysfunctions (0.00). They were more likely to perpetrate ante-mortem vaginal penetration (0.64) and to rob some items from the victim (0.97). However, they were less likely to perpetrate ante-mortem anal penetration (0.18), foreign object insertion (0.00), and to mutilate victims' genitals (0.03). These SHOs were also less likely to use asphyxiation/strangulation as a method of killing (0.44), to use restraints, and to target a specific victim (0.33).
In class 2, labeled experimental, offenders were less likely to be single (0.07) or to report sexual dysfunctions (0.00). These offenders perpetrated ante-mortem vaginal penetration (0.61), anal penetration (0.54), and foreign object insertion (0.54). They were more likely to use asphyxiation/strangulation as a method of killing (0.79). These sexual murderers did not mutilate victims' genitals (0.07), did not use restraints (0.00), did not specifically target a victim (0.39), and did not take items from the victim (0.18). In class 3, labeled preferential, SHOs were more likely to be single (0.76) and report sexual dysfunctions (0.65) but did not perpetrate either ante-mortem vaginal penetration (0.00) nor ante-mortem anal penetration (0.00). They were unlikely to mutilate the genitals of their victims (0.18), use restraints (0.24), target a specific victim (0.29), or rob items (0.18), and never killed their victims using asphyxiation/strangulation (0.00), In class 4, sadistic, SHOs were not likely to be single (0.18) and to report sexual dysfunctions (0.18). These offenders perpetrated ante-mortem vaginal penetration (0.96), ante-mortem anal penetration (0.96), ante-mortem foreign object insertion (0.96), and mutilation of genitals (0.71). They also used restraints (0.96) as well as asphyxiation/strangulation as a method of killing (0.96). These offenders specifically targeted their victims (0.96) but did not rob items from them (0.00). Table 3 presents bivariate analysis of covariates. Offenders who possessed a sexual collection (χ 2 = 58.05, p = .000) were more likely to be included in class 4, while those that were characterized by a loner lifestyle (χ 2 = 24.45, p = .000) were more likely to be part of class 3. Offenders with previous criminal convictions (χ 2 = 9.34, p = .025) and those that used alcohol/drugs prior to crime (χ 2 = 17.02, p = .000) were more likely to be included in class 1 and 4. Victims assaulted by stranger offenders (χ 2 = 10.37, p = .016) and while they were involved in domestic activities (χ 2 = 31.49, p = .000) were more likely to be part of class 1. Victims assaulted while they were jogging (χ 2 = 15.98, p = .001) were more likely to be included in class 2, while victims aged 65 years or older (χ 2 = 12.45, p = .006) were more likely to be part of class 3. Finally, victims living alone (χ 2 = 8.76, p = .033), using alcohol/drugs (χ 2 = 20.25, p = .00), and who were drinking at a bar prior to be assaulted (χ 2 = 59.28, p = .00) were more likely to be included in class 4.

Discussion
This study investigated patterns of necrophilic behaviors in sexual homicide cases. Specifically, this research aimed to empirically determine whether SHOs who perpetrated necrophilic acts were driven by a unique motivation. To answer this research question, we used a sample of 109 sexual homicide cases committed in France and Canada between 1948 and 2018, where post-mortem sexual acts were identified.

Toward a New Classification of Necrophilic Behaviors Patterns in Sexual Homicide
As discussed in previous studies, several aspects are relevant to explaining the motivations associated with the commission of necrophilic behaviors (see e.g., Rosman & Resnick, 1989;Stein et al., 2010). Hence, offender characteristics (e.g., social integration, sex life, substance use), sexual (ante-mortem sexual acts, sadistic processes), and non-sexual behaviors (e.g., theft of objects, victim selection) are all associated with this paraphilia. In contrast to certain classifications based on some clinical observations and theoretical assumptions (e.g., Aggrawal, 2010), empirical classifications are identified from an adequate number of observations, which increase the generalization of findings. Using latent class analysis on a sample of 109 sexual homicide cases characterized by the presence of postmortem sexual activities, the best model suggested that SHOs involved in necrophilic behaviors can be divided into four groups: Opportunistic, experimental, preferential, and sadistic.
Opportunistic. The opportunistic offenders constitutes the most important group of our model. Offenders are mostly single, consumed alcohol prior to the crime, and are more likely to have prior criminal convictions. These characteristics are congruent with the ones observed in previous studies on individuals for which necrophilic acts are not the primary motivation for the crime (Aggrawal, 2010;Rosman & Resnick, 1989;Stefanska et al., 2017;Stein et al., 2010). Moreover, the presence of previous criminal convictions confirmed the findings of Beauregard et al. (2018), suggesting that SHOs are mostly versatile and may engaged in various criminal activities. These offenders assaulted their victims while they were involved in domestic activities or sleeping. This finding suggests that most victims were in their residence at the time of the crime. Interestingly, we observed that crimes perpetrated by opportunistic offenders were characterized by two main aspects: the presence of robbery and the low diversity of ante-mortem sexual acts perpetrated by the offender. Based on previous studies, two interpretations are possible. On one hand, taking items belonging to the victim can happen after the sexual assault and the homicide (see e.g., Chopin & Beauregard, 2020b;Safarik et al., 2002). On the other hand, taking items from the victim may be the primary motivation, while the sexual acts and the homicide are occurring as an afterthought (see e.g., Beauregard et al., 2020;Chan et al., 2019;Chopin & Beauregard, 2020bPedneault et al., 2015). We argue that the latter is more likely. The combination of theft, in the victim's residence, and with a low level  of diversity in the sexual acts committed suggest that the primary motivation of these offenders was to perpetrate a burglary. The confrontation with the victim gave them the opportunity to perpetrate ante-mortem and post-mortem sexual acts. Moreover, our findings are in line with those of Stein et al. (2010), who suggested that necrophilic behaviors can be a consequence of alcohol/drug intoxication by offenders. Similarly, our findings showed that most opportunistic necrophiles consumed drug/alcohol prior to the crime. Perkins (2008) also noted that necrophilia is often not the primary motivation of a sexual homicide but that postmortem sexual acts can be perpetrated for sexual gratification purposes without resistance and without making matters worse since a killing has already occurred" or because "feelings of anger toward the victim for resisting and dying during the course of a violent sexual assault, resulting in anger being directed toward the corpse" (p. 81).
Opportunistic offenders assaulted acquaintance victims in the majority of cases. This information combined with the absence of sexual sadism suggests that the murder is not part of a deviant script. As previously discussed in the research on sexual homicide, the death of the victim is not necessarily part of a deviant process and can be accidental (i.e., due to the use of an excessive amount of physical force to overcome the victim resistance) or instrumental (i.e., offender willingness not to be denounced to the police by the victim) (see Carter & Hollin, 2014;Chopin & Beauregard, 2019a, 2019bStefanska et al., 2017). Considering that opportunistic offenders are more likely to use specific strategies to avoid police detection, it can be argued that killing the victim presented itself as an opportunity to avoid being identified by the police.
Experimental. Experimental offenders are usually in a relationship, used alcohol/drugs prior to the crime, and assaulted only female victims. Looking at the crime-commission process, we observed two major tendencies. First, analyses indicated that an important diversity of ante-mortem sexual acts were committed by these offenders. Experimental offenders were more likely to vaginally and anally penetrate the victim as well as commit foreign object insertion and fondling. Second, our findings indicated that these offenders did not engage in ante-mortem and post-mortem sadistic acts. The experimental offenders class is congruent with sexually-motivated SHOs characterized by the diversity of sexual acts they perpetrate (see e.g., Chopin & Beauregard, 2020b;Stefanska et al., 2015). Based on the diversity of ante-mortem sexual acts perpetrated by experimental offenders, we can assume that they were primarily motivated to have sexual intercourse with a living victim. This is in line with findings from Rosman and Resnick (1989) who stated that despite the commission of postmortem sexual acts, some SHOs involved in necrophilic behaviors prefer sexual contact with living partners. This idea is reinforced by the fact that most of these offenders were in a relationship at the time of the crime.
It is possible that the commission of post-mortem sexual acts by these offenders is motivated by the experimentation of new sexual practices. In other words, postmortem sexual activities was not the primary motivation for these SHOs but given the situation, they experimented with this deviant sexual behavior. Such interpretation is congruent with observations made by Rosman and Resnick (1989) who found that pseudonecrophiles had a transient attraction to corpses.
Preferential. Preferential offenders are mostly single, report sexual dysfunctions, and have a loner lifestyle. These characteristics are congruent with several studies which identified a marginal lifestyle in some necrophilic offenders (Chopin & Beauregard, 2020b;Hirschfeld, 1944;Rosman & Resnick, 1989;Stein et al., 2010). These characteristics suggest that these individuals present relational difficulties and fail to achieve normal relationships (i.e., social, sexual).
This group of SHOs present two possible patterns of "true necrophilia" (Rosman & Resnick, 1989, p. 156). First, in some cases these offenders tried to have ante-mortem sexual interactions (e.g., foreplay) and faced with the failure of successfully completing the crime (e.g., unable to obtain or sustain an erection; victim's resistance), they chose to kill their victim to have an unresisting and unrejecting partner, as described by Rosman and Resnick (1989). Second, in other cases where no ante-mortem interactions were observed, the offender main motivation was to obtain a corpse for sexual purpose (Aggrawal, 2010;Rosman & Resnick, 1989). In these cases, the victim's death is motivated by the necrophilic urge described by Aggrawal (2010, p. 73) as 'the urge to kill because a dead body is required for necrophile activities'. In such situations, postmortem sexual activities are more likely to be combined with body mutilation and criminal dismemberment (see e.g., Holmes, 2017;Purcell & Arrigo, 2006). Several studies (Püschel, 1987;Püschel & Koops, 1987) suggested that offensive mutilation (i.e., mutilations are part of the deviant process) is often observed in cases of necrophilia. These extreme deviant sexual behaviors can be motivated by the willingness of necrophiles to have sexual activities with body parts of corpses (see e.g., Boureghda et al., 2011;Chopin & Beauregard, in press;.
Sadistic. Sadistic offenders share all the characteristics associated with sadistic SHOs from the literature. First, they have a relatively common profile and are well socialized (i.e., in a relationship, not a loner lifestyle) which is congruent with previous studies (see e.g., Brittain, 1970). We observe that these individuals possessed a sexual collection (i.e. movies, pictures involving deviant sexual behaviors) allowing them to feed their deviant sexual fantasies (see e.g., Brittain, 1970;Chopin & Beauregard, 2019b, 2019c. Second, our results showed that the modus operandi used by offenders included in this class is consistent with the one described by previous studies of sadistic SHOs (Chopin & Beauregard, 2019c;Myers et al., 2019;Proulx et al., 2007;Reale et al., 2017). This includes both the approach used (i.e., con approach strategy, targeting vulnerable victims, use of restrains, see Beauregard & Martineau, 2016;Ressler et al., 1986Ressler et al., , 1988, ante-mortem sexual acts (i.e., the combination of vaginal and anal penetration, foreign object insertion, see Myers et al., 2019), acts of torture (i.e., vaginal/ anal fisting, mutilation of genitals, see Myers et al., 2019), methods of killing (i.e., strangulation and asphyxiation, see Chopin & Beauregard, 2020a;Healey et al., 2013;Myers et al., 2019;Reale et al., 2017), and the ability to avoid police detection (i.e., body concealment, see Chopin & Beauregard, 2020a;Reale et al., 2020). As to the mutilation of genitals,  noted that their presence are fundamental to the expression of violent sexual sadistic fantasies and/or necrosadistic fantasies. The high probability of occurrence of these acts by SHOs included in this class confirms this assumption. This class of offenders is congruent with previous studies which identified that the necrophilic behavior is often associated with, or is part of sexual sadism (see e.g., Aggrawal, 2010;Holmes, 2017;Krafft-Ebing, 1886;Pettigrew, 2019c;Purcell & Arrigo, 2006). Moreover, Rosman and Resnick (1989) found that the pseudonecrophile category includes sadistic cases and determine that most pseudonecrophilic killers and necrophilic murderers had history of sadistic acts. Necrophilia in a sadistic sexual process can have different purposes. First, according to the sexual sadism scale developed by Myers et al. (2019), necrophilic acts can be considered as an ultimate way of degradation and humiliation. Second, as mentioned by Stein et al. (2010), in some cases necrophilic acts were perpetrated because ante-mortem behaviors as well as the murder were not sufficiently satisfying to the offender. For the sadistic offenders, post-mortem sexual acts are considered as secondary acts of a process focused on victims' humiliation, torture, and death.

Necrophilic Behaviors in Sexual Homicide: Primary Motivation to Kill or Secondary Deviant Behavior?
One of the underlying questions for this study of necrophilic behaviors in sexual homicide was to determine whether the murder is directly associated with the achievement of the deviant fantasy to commit sexual acts with corpses. First, in the four-class model we identified, four classes of offenders (i.e., opportunistic, experimental, preferential, sadistic) who perpetrated ante-mortem sexual and non-sexual acts with their victims. These classes of SHOs involved in necrophilic behaviors correspond with offenders associated with the pseudonecrophile murders identified by Rosman and Resnick (1989). This category of offenders had sexual activities with corpses during the crimecommission process but these acts were not the primary motivation for their crime nor the main object of their sexual fantasies (Rosman & Resnick, 1989). Moreover, these classes are also congruent with Stein et al. (2010) findings that none of the SHOs involved in necrophilic behaviors they investigated specifically perpetrated homicides to obtain a corpse for sexual purposes (Stein et al., 2010).
Second, we found in our LCA model one class-labeled preferential offenderscharacterized by a crime-commission process which we hypothesize to be mainly focused on the obtaining and the commission of sexual acts with a corpse. This category is congruent with Aggrawal's (2009aAggrawal's ( , 2009bAggrawal's ( , 2010 homicidal necrophile and Rosman and Resnick (1989) genuine necrophilic murderer. These offenders, considered as the 'true necrophile' (Rosman & Resnick, 1989, p. 156), present a sexual attraction to corpses and focus their crime-commission process on this aspect.

Conclusion
This study investigated patterns of sexual homicide where necrophilic acts were perpetrated. We used a sample of 109 solved cases that occurred in France and Canada between 1948 and 2018. The objective of this study was to empirically determine whether SHOs involved in necrophilic behaviors represent a homogeneous type of offender. As an underlying question, we aimed to determine whether post-mortem sexual activities were directly related to the victim's death or if there was evidence to suggest that necrophilic behaviors were not the offender's primary motivation. Using latent class analysis, we generated a four-class model of SHOs involved in necrophilic behaviors : Opportunistic, experimental, preferential, and sadistic. Our findings showed that preferential offenders are the only ones who specifically kill their victims in order to have sex with their corpses. For sadistic offenders, the victim's death is part of the sexual deviance and post-mortem sexual acts are a way to pursue the humiliation and degradation of the victim. Experimental and opportunistic offenders killed their victims to attempt to avoid police detection and perpetrated post-mortem sexual acts because they had the opportunity to do so or because it was part of secondary deviant sexual fantasies to experiment.
This study presents several implications. Our findings provide empirical evidence that SHOs involved in necrophilic behaviors constitute a heterogeneous population of offenders. This could have implications for practitioners working with these individuals. From an investigative perspective, criminal investigators should be aware that the presence of post-mortem sexual acts does not refer to a single category of SHOs and that it is important to observe other crime scene behaviors to determine working hypotheses and suspect prioritization (e.g., identification of robbery, ante-mortem interactions, sadistic process, etc.). Moreover, as SHOs involved in necrophilic behaviors most often target victims they know, the police should prioritize looking at the victims' entourage for potential suspects. From an intervention perspective, practitioners should adapt the management and treatment of these individuals based on the motivation for the necrophilic acts. For instance, preferential offenders would benefit from an intervention directly targeting their sexual attraction to the corpse, whereas for the other groups of SHOs involved in necrophilic behaviors, the intervention should focus on the main paraphilias associated with the crime commission such as sadism, erotophonophilia (i.e., sexual gratification found in the act of committing murder) or raptophilia (i.e., sexual gratification found in the act of committing a rape).
Although the current study presents novel and interesting findings, it is not without limitations. This study is based on police data which are known to present biases in terms of validity and reliability (see Aebi, 2006;Chopin & Aebi, 2018. First, this data includes only cases reported to authorities and we cannot exclude that some cases of homicide were never reported. However, we can assume that it represents a very limited number of cases as the dark figure (i.e., number of unreported or undiscovered crimes) for homicide is especially low (see e.g., Aebi & Linde, 2012). Second, we cannot exclude that in some cases post-mortem sexual activities were not identified by coroners and investigators and consequently were not included in our sample. Finally, we included only solved cases in our sample and therefore, it is unknown whether the identified patterns may also apply to unsolved cases.
Future studies need to test the validity of this classification with datasets from other countries. More empirical research is also needed to analyze developmental and psychological risk factors of the different type of SHOs involved in necrophilic behaviors. Finally, future research could investigate in greater details the influence of some of the variables included in the current study, such as the different types of post-mortem sexual acts perpetrated as well as the detailed previous criminal convictions.