Sexual Abuse of Elderly Victims Investigated by the Police: From Motives to Crime Characteristics

Elderly sexual abuse has been almost completely ignored from researchers and practitioners alike. However, the occidental population is aging and living longer, suggesting that the number of cases of elderly sexual abuse should increase. Moreover, elderly sexual assaults have been described as being more violent, resulting in more severe injuries, and are more frequently committed by strangers, making criminal investigations more difficult to solve. The current study aims to identify the various motivations associated with elderly sexual abuse and to test whether it is possible to link offender and modus operandi characteristics to these motivations. In other words, the main objective is to identify “why” the elderly are sexually abused, “how,” and “by whom”? Using two-step cluster analysis on a sample of 128 cases of extra-familial elderly sexual assaults (aged 65 years or more) from France, four clusters of offenders’ motivation were identified. Congruent with previous studies, results showed that elderly sexual abuse was motivated by sex, anger, and opportunities. However, a fourth cluster was identified, describing offenders motivated by experimentation. These offenders, in addition to being young with a lack of criminal experience, were also more likely to perform the most intrusive sexual acts and to use physical violence, sometimes to the point of killing their victim. To test the external validity of our cluster solution, a series of bivariate analyses were conducted. Results showed that the four motivations were also associated with specific offender and crime characteristics. These findings highlight the importance of looking at the motivations underlying elderly sexual abuse to suggest better interventions strategies as well as improve the criminal investigation of these cases.


Introduction
Sexual violence against elderly people is considered an awful and unusual crime as motivations underlying these acts defy rationality. Similar to child sexual abuse, sexual crimes against elders fall outside the stereotypes normally associated with sexual violence. Jones and Powell (2006) qualified elder sexual abuse as one of modern life's final taboos. This contrasts sharply with the image of the "classical rape," which describes a sexual assault committed by a stranger motivated by uncontrollable sexual impulses, during nighttime, at an outdoor deserted location, against a young woman (Anderson & Doherty, 2008). Lay people are unaware of elder sexual abuse (Simmelink, 1996) and elders themselves have generally no knowledge of the risk of sexual abuse victimization (Jeary, 2005). Empirical studies on children and adult sexual assaults (e.g., Proulx, Beauregard, Lussier, & Leclerc, 2014) have shown that the offending processes related to sexual abuse are more complex and diversified than the image of the "classical rape" described by Anderson and Doherty (2008). It is noteworthy that currently, there are only a few studies on elder sexual abuse (Ball, 2005;Bows, 2018;Jones & Powell, 2006), probably due in part to the difficulties obtaining data on this specific population.
However, more research is needed on elder sexual abuse for several reasons. First, the occidental population is aging and living longer (e.g., Jones & Powell, 2006), suggesting that the number of cases of elder sexual abuse should mathematically increase. Also, most of the existing research on this topic agree that elder sexual abuse is more violent than sexual assaults on other adults, resulting in more severe injuries (Burgess, Commons, Safarik, Looper, & Ross, 2007;Groth, 1978;Pollock, 1988;Safarik & Jarvis, 2005;Safarik, Jarvis, & Nussbaum, 2002). Elder sexual abuse may be committed in several contexts and by various types of offenders who can be strangers, care providers, partners, or care facility residents (Ramsey-Klawsnik, 1998, 2004Ramsey-Klawsnik, Mendiondo, Cecil, & Tooms, 2007). As shown by the literature review from Bows (2018), findings on the relationship between offenders and victims depend largely on the nature of the data. Research using police data usually identifies a majority of stranger offenders (Burgess et al., 2007;Burgess, Ramsey-Klawsnik, & Gregorian, 2008;Kennedy & Silverman, 1990;Safarik et al., 2002) whereas data from protective services or self-reported show more acquaintance and/or familial cases (Nóbrega Pinto, Rodrigues, Dinis-Oliveira, & Magalhães, 2014;Pollock, 1988;Ramsey-Klawsnik, 1991;Teaster, Roberto, Duke, & Kim, 2001). Cases committed by stranger offenders are particularly important to examine as they are harder to solve by the police (Beauregard & Martineau, 2017).
The current study aims to identify the various motivations associated with elder sexual abuse and to test whether it is possible to link offender and modus operandi characteristics to these motivations.

Elderly Sexual Abuse Explained by Anger
To understand elder sexual abuse, scholars have used various theories. The first one is a variation of the psychodynamic interpretation of rape (Brill, 1968), suggesting that victims are partially involved in their own victimization. This interpretation is mainly based on offenders' anger and suggests that the sexual assault is not the main motivation, but instead it is anger and the need for control of the victim (Burgess et al., 2007). Groth (1978) further suggested that victims are generally persons in authority. Empirical studies have shown that elder sexual abuse is the outcome of the offender's anger (Burgess et al., 2007;Groth, 1978Groth, , 1979Pollock, 1988;Safarik & Jarvis, 2005). For this motivation, sexual assault is not the intended goal, but the means to exert anger and control over the victim (Burgess et al., 2007). In a clinical study, Groth (1978Groth ( , 1979 has noted that some sex offenders targeting elderly people have reported familial problems during their adolescence mainly due to the unstable relationship of their parents, causing anger and a lack of respect for their mother. These offenders entertain a negative perception of their mother, leading them to search for an older victim who represents a surrogate for their mother and on which they will exert their revenge. Groth's (1978Groth's ( , 1979 findings showed that offenders motivated by anger are characterized by a brutal modus operandi and by the use of weapons during the attack. Previous studies (Burgess et al., 2007;Groth, 1978Groth, , 1979Lanyon, 1991;Pollock, 1988;Safarik et al., 2002) have suggested that offenders motivated by anger specifically search for a substitute to their source of anger to inflict excessive injuries. In the study of Burgess et al. (2008), 6% of assaulted elder people had severe injuries necessitating hospitalization in an intensive care unit. Some of these hypothesized characteristics associated with anger motivation have been tested in previous studies. For instance, Pollock (1988) compared five sex offenders of elderly people to seven sex offenders of younger victims. Although no significant differences were observed between the two groups in terms of childhood behavior problems (probably due to the small sample size), he noticed that the force used by sex offenders of elderly victims was greater than the one necessary to control the victim during the attack. Those findings are also confirmed by the study of Safarik and Jarvis' (2005), who showed that by examining the qualitative values of injury severity on a sample of 128 cases of sexual homicides of elderly people, these cases presented an excessive level of injury distinct from other violent crimes. Burgess et al. (2008) found that the majority of victims had genital injuries such as bruisings, abrasions, redness, swellings, and tears in the perianal area.
Other studies have highlighted more genital trauma among older victims (Cartwright, 1987;Muram, Miller, & Cutler, 1992). Muram et al. (1992) suggested that postmenopausal period increases the risk of genital injuries due to the reduced level of estrogen causing a vaginal mucosa atrophy. However, other studies have found no evidence of higher risk of genital trauma among older victim of sexual assault (e.g., Cartwright & Moore, 1989;Ramin, Satin, Stone, & Wendel, 1992). In their study, Burgess et al. (2007) found that approximately 22% of their sample corresponded to the vindictive motivation. They operationalized this motivation from evidence of misogynic anger, acts of sexual humiliation, and sadistic fantasies. Moreover, they made a distinction with pervasive anger motivation, for which women are not the exclusive source of anger (Burgess et al., 2007). Their findings showed that offenders motivated by anger present a high level of expressive aggression, juvenile and adult antisocial behavior, and a constant mood of anger. In the typology of elder-abuse offenders suggested by Ramsey-Klawsnik (2000), the sadistic offender is characterized by sexual assaults committed on elder people. Modus operandi used by this type of offender is based on humiliation, terror, and harm of other. They commit severe abuse inflicting notably physical assaults to genitals (Ramsey-Klawsnik, 2000).

Elderly Sexual Abuse Explained by Sexual Motivation
A second theory is based mainly on a sexual motivation-namely gerontophilia-to explain elder sexual abuse, as mentioned in Psychopathia Sexualis, by Krafft Ebing (1886). Thus, elder sexual abuse is hypothesized to be driven by a paraphilia describing an attraction toward elder people. According to Burgess et al. (2007), such explanation suffers from a lack of empirical evidence. Nonetheless, some studies have shown that elder sexual abuse could also be motivated by sex (Ball, 2005;Burgess et al., 2007;Jeary, 2005). Krafft-Ebing noted that in many cases of gerontophilia, offenders were serial. This sexual attraction for older people has been described in previous case reports (e.g., Ball, 1998;Ball, Snowden, & Strickland, 1992), but Ball (2005) cautioned that there is no empirical evidence estimating the proportion of gerontophilic offenders among the ones assaulting elderly victims. Later, Burgess et al. (2007) suggested a less restrictive qualification for the sexual motivation. Based on the Massachusetts Treatment Center: Rape 3 (MTC: R3) typology, they operationalized sexual motivation using various criteria. In sum, the main preoccupation of these offenders is the need for sexual gratification. This need is often evidenced by the possession of pornographic material and the involvement in deviant sexual behavior (Burgess et al., 2007). Classifying 78 cases of rape implying elderly victims, they found that these offenders represented 44.15% of their sample (Burgess et al., 2007). These offenders are characterized by a low level of expressive aggression, an absence of anger, and a high level of sexual penetration (Burgess et al., 2007). These results were similar to Jeary (2005), showing that one third of his sample was motivated by sexual gratification. Jeary (2005) showed that most of these cases were young men who had broken into elders' residences and who presented sexual inadequacies. These results suggested that some offenders are not able to establish relationships and have sexual intercourse with sameaged people. This was reflected in Jeary (2005) study, as many of these offenders were not able to have an erection or to ejaculate during a consensual act. Thus, according to the sexual motivation hypothesis, these offenders have assaulted older victims to test their sexual performance and to obtain sexual gratification with a more vulnerable partner.

Elderly Sexual Abuse Explained by Opportunities
A third theoretical framework to explain elder sexual abuse focuses on the concept of opportunity, based on the lifestyle (Hindelang, Gottfredson, & Garofalo, 1978), routine activities (Cohen & Felson, 1979), and rational choice theories (Cornish & Clarke, 1986. Basically, such framework suggests that crime is the result of an opportunity encountered by a motivated offender when the victim is vulnerable and with no one around to protect her (also referred as a capable guardian). Such opportunities are assessed by the offender through an economic analysis of crime (i.e., costs vs. benefits). Empirical evidence has partially confirmed the role of opportunity in elder sexual abuse. For instance, Jeary (2005) suggested that rape of elderly people could occur during the course of a burglary, where the main motivation was to burglarize residences to finance drugs and/or alcohol consumption. These offenders target elderly residences because of the reduced security and the knowledge that older people have a tendency to keep more cash in their home. Based on the MTC: R3 typology, Burgess et al. (2007) have shown that 14.28% of their sample corresponded to the opportunist offender. They suggested that these offenders are impulsive with a low level of planification and premeditation. Sexual assaults perpetrated by opportunist offenders are characterized by a lack of sexual penetration and the commission of acts of fondling and kissing (Burgess et al., 2007). Opportunist offenders are mostly characterized by the presence of juvenile and adult antisocial behaviors (Burgess et al., 2007).

Aim of Study
Previous studies on the topic have not only been scarce, but suffer from small sample size, the absence of multivariate analyses, and a limited number of variables investigated. Finally, the three motivations associated with elder sexual abuse discussed in the literature have been identified through clinical observations and have not been tested empirically, despite the fact that several authors have used them in their own study. To help the investigation of elder sexual abuse, the current study uses information from a national police database. Although very few cases of elder sexual abuse are ever prosecuted, this database includes only the cases that are the most challenging to solve (e.g., stranger offender) with information that is available to the police during their investigation. Therefore, the aim of the study is twofold. First, the study will identify the various motivations associated with elder sexual abuse. Second, the study will examine whether these motivations are associated with offender and crime characteristics. We can derive three hypotheses from these two research goals: Hypothesis 1: Elderly sexual offenders present different motivations, such as opportunistic, vindictive (anger), and sexual. Hypothesis 2: There exists a relationship between the offender's motivation and his modus operandi. Hypothesis 3: There exists a relationship between an offender's characteristics and his motivation for elder sexual abuse.

Data
Sample. The sample used for this study comes from a French national police database managed by the Ministry of Interior in France, and it has traditionally been used to complement the criminal investigation as well as to identify violent serial offenders. All criminal events included have taken place on the French Territory between 1979 and 2014 and are complete (no attempt) extrafamilial sexual assaults against elderly victims. Crime analysts are specifically responsible for entering the information about the cases based on the criminal investigation files, reducing considerably the number of missing values. Information included in these files is mainly filled out by police officers, but also other experts were involved in the investigative process (e.g., coroner, psychologist). Crime analysts responsible for the information included in the database assure a certain quality control by checking for inconsistencies. Although superior to most police data, it is important to keep in mind that the data used in this study has been collected for operational and not research purposes. Sexual abuse has been operationalized in this study as the occurrence of at least one of the following acts: penile, digital, or foreign object penetration (vaginal and/or anal), foreplay (i.e., fellatio, masturbation, cunnilingus), or fondling. Although there is no consensus as to what age should be considered as "elderly" (Ball, 2005;Lea, Hunt, & Shaw, 2011), Lea et al. (2011) suggested that it constituted a central methodological question. Davis and Brody (1979) considered in their study victims aged 50 years or above, Muram et al. (1992) considered 55 years or above, whereas Ball (2005) noted in his review that some studies went as far as considering an age cut-off of 80 years. Despite the use of different thresholds in the literature, the current study chose 65 years or above1 to identify elderly victims based on the suggestion from Lea et al. (2011) who recommended using retirement age as the main criteria as this often involves significant changes in terms of lifestyle, routine activities and consequently, risk exposure. Therefore, the sample used is this study includes 128 offenders who have committed extra familial sexual abuse against 128 victims.
Victims. Victims included in the sample are mostly females (98.40%). They are aged, on average, of 76.8 years (SD = 8.12) with a maximum of 94 years. White victims represent the majority (61.72%) followed by North African (13.28%) and Black ones (12.80%). A majority of them were not single (79.80%) and did not live alone (85.16%) at the time of the offense. Finally, victims have been mainly sexually assaulted while involved in domestic activities (e.g., watch TV; 37.50%), when they were sleeping (31.30%) or when walking (25%).

Analytical Strategies
To answer our research question, we have analyzed the data in two stages. First, to classify the 128 offenders according to their motivation, we performed a two-step cluster analysis using the Schwarz Bayesian Criterion (BIC) as an indicator for the best cluster solution. Nine dichotomous (coded as 0 = no, 1 = yes) variables were selected to measure the offenders' motivations as presented in the previous studies (i.e., opportunistic, vindictive, and sexual motivation; Burgess et al., 2007;Jeary, 2005). For the opportunistic motivation, we have considered the following variables: (a) nontargeted victims, (b) offender is familiar with the crime scene, (c) lack of sexual penetration, and (d) resistance of victims. For the vindictive motivation, we have considered (e) the strong level of physical force used by the offender, (f) the anger of the offender, and (g) the presence of humiliation acts. Finally, for the sexual motivation we have included (h) the presence of paraphilic behaviors2 and (i) the existence of a sexual collection. We have tested for multicollinearity and no correlations were higher than .27 (appendix section).
To test the external validity of our motivation classification, the second stage of the study consisted of examining the bivariate relationships (through ANOVAs and chi-square analyses) between the cluster solution and additional variables related to offender and modus operandi characteristics. As to the modus operandi, 16 dichotomous variables (coded as 0 = no, 1 = yes) were selected: (a) offender is a stranger, (b) con approach, (c) surprise approach, (d) blitz approach, (e) the crime location was deserted, (f) witness could see or interrupt the assault, (g) the aggression has taken place in a public area, (h) the aggression occurred in an outdoor location, (i) vaginal intercourse, (j) anal intercourse, (k) foreplay, (l) fondling, (m) use of a weapon during the crime, (n) victim was beaten, (o) victim has been intentionally released, and (p) victim has been killed.
As to the offender characteristics, 13 variables were selected: (a) offender age, (b) offender is single, (c) offender is living with his parents, (d) offender is living with his partner, (e) offender is living alone, (f) offender is of medium build, (g) offender experienced sexual dysfunctions, (h) offender consumed psychoactive substances, (i) offender was frequently engaged in social activities, (j) offender was a loner (k) offender had no fixed address (homeless), (l) offender was frequently engaged in criminal activities, and (m) offender was handicapped (physically and/or psychologically). Table 1 shows a four-cluster solution from the two-step cluster analysis. These four groups correspond to the different motivations involved in the sexual assaults of elderly women. The four-cluster motivations are as follows: sexual (n = 33; 25.78%), opportunistic (n = 38; 29.69%), experimental (n = 44; 34.38%), and vindictive (n = 13; 10.15%).

Classification of Offenders Depending on Their Motivations
The first cluster is related to a sexual motivation. All offenders motivated by sex perform sexual penetration (100%) and present paraphilic behaviors, and almost one quarter of them possess a sexual collection (n = 24; 20%). The second cluster describes an opportunistic motivation. These offenders typically do not commit sexual penetration, half of them are unfamiliar with the location of the crime, and one third (34.2%) do not target their victims prior to their assaults. In cases of opportunistic motivation, victims typically resist during the assault (78.90%), and these offenders do not display indicators of sexual deviance nor do they possess a sexual collection. The third cluster describes a motivation that we labeled as "experimental." In these cases, offenders perform sexual penetration (100%) but no indicators of sexual paraphilias are found. Similar to the opportunistic motivation, victims typically resist during their assault (81.80%); however, most offenders are familiar with the location of the crime and select their victims. The last cluster describes a vindictive motivation. Offenders presenting a vindictive motivation are all angry during the sexual assault (100%), and they are more likely to perform acts of humiliation (46.20%) in comparison with the other clusters. However, most of these offenders do not perform acts of sexual penetration (85%).    Table 3 shows findings of the relationship between offenders' characteristics and the offenders' motivations in sexual assault of elderly women. As can be seen, experimental offenders are significantly younger (29 years old) compared with other groups, especially the opportunist offenders (37 years old). Also, sexually motivated offenders are more likely to be single (93.94%) compared with opportunist (65.79%), experimental (56.82%), and vindictive (53.85%) offenders (Cramer's V = .33, p ≤ .01). Vindictive offenders are more likely to live with a partner at the time of offense (Cramer's V = .22, p ≤ .1) than the other offenders, whereas sexually motivated offenders are more likely to live alone (Cramer's V = .31, p ≤ .01). Vindictive offenders are less likely to present a medium built compared with other offenders (Cramer's V = .25, p ≤ .05). Interestingly though, sexually motivated offenders are the most likely to present sexual dysfunctions (Cramer's V = .31, p ≤ .01). Opportunist offenders are significantly less likely to use psychoactive substances at the time of offense (Cramer's V = .29, p ≤ .01), whereas sexually motivated offenders are more likely to be loners compared with the other offenders (Cramer's V = .29, p ≤ .01).

Discussion
Although several typologies of sexual aggressors of women have already been identified (see Proulx & Beauregard, 2009 for a review), very few studies have examined specifically those sex offenders who target elderly victims. There are several reasons why elderly victims should be the focus of more studies. First, elderly victims are more likely to be attacked by strangers (e.g., Kennedy & Silverman, 1990), which complicates the investigation (see Beauregard & Martineau, 2017). Second, they are most likely to be victimized in their own homes (e.g., Muram et al., 1992), which combined with the fact that they are more likely to live alone, reduces the chance of someone interrupting or witnessing the assault. Third, due to a general difference in physical size and strength, elderly victims are usually less capable of resisting a physical attack compared with younger females (e.g., Nelsen & Huff-Corzine, 1998), and, therefore, they are less likely to avoid injuries during the crime (Safarik et al., 2002). It is noteworthy that this increased risk of injury of elderly female is also due in part to the biological aging process (e.g., following menopause, there is an increased risk of osteoporosis). A fourth reason as to why it is important to focus on elderly victims is brought on by the routine activities perspective (Cohen & Felson, 1979). The routine activities theory suggests that the risk of a crime occurring increases when there is convergence in time and space of an attractive/vulnerable target, a motivated offender, and the absence of appropriate guardianship. Thus, partly due to an increased risk of widowhood and longer life expectancy (Safarik et al., 2002), approximately 80% of elderly people who live alone are female (Moen, 1996), which could be perceived by a motivated offender as more vulnerable considering a greater lack the guardianship compared with younger women living with parents, boyfriends, or husbands. Finally, the research on this specific group of offenders is of the utmost importance as was highlighted by Safarik, Jarvis, and Nussbaum (2000), demographic data show that the combined effect of greater life expectancy and the "baby boomers" getting older will naturally lead to an increase in the elderly population, which is in majority composed of women. One of the main questions raised with regard to elder sexual abuse is the question of motivation. Some have suggested that these acts were mainly motivated by a sexual attraction to older women-referred as gerontophilia. Others instead have insisted on a motive of anger, more specifically the sexual assault serving to achieve revenge against a woman. Finally, others have simply suggested that elderly victims were targeted just because of the opportunity. These different motivations suggest that elder sexual assaults represent a heterogeneous phenomenon. Thus, in one of the few studies on the topic, Burgess et al. (2007) classified 77 cases of elderly sexual assaults using the MTC: R3 classification (Knight & Prentky, 1990). What they found was that all three motivations were represented in their sample (i.e., sexual, vindictive/anger, and opportunistic). However, one of the issues with this study is the fact that Burgess et al. (2007) did not actually empirically test the presence of these motivations, choosing instead to classify their cases with an existing classification. Although these findings are interesting on their own, our study used a different approach. Instead of relying on an existing typology, we submitted our data to an empirical test. The results are very informative. We were not only able to replicate the motivations identified by Burgess et al. (2007), but we also identified an additional motivation never discussed before in previous studies: experimental.

Experimental Offender
The experimentally motivated offender constitutes the most important group of the classification identified. These offenders are typically younger, and they are medium build. They choose to attack their victims-not necessarily a stranger-mostly at deserted locations that they are not familiar with. They are very likely to encounter victim resistance and they are willing to use physical force, from different types of approach (i.e., con, blitz, and surprise). Interestingly, these offenders are the ones more likely to see their level of force leading to the death of the victim. Although they do not exhibit deviant behavior or possess a sexual collection, they are more likely to penetrate the victim, both vaginally and anally.
The experimentally motivated offender presents similar characteristics of the opportunistic offender, but also some important differences. We believe that due to a lack of experience, this offender is likely to attack the victim somewhere he is not familiar with and when facing victim resistance, he is likely to resort to physical violence to a point where it can even lead in some instances to a fatal outcome (see Beauregard & Mieczkowski, 2012;Mieczkowski & Beauregard, 2010). However, despite this obvious lack of planning and control at the crime scene, this offender seems to know exactly what he wants out of this attack. Thus, the experimentally motivated offender is looking to accomplish certain sexual acts-the most intrusive ones, namely, vaginal and anal penetration, which suggests that some may be sexually motivated. It is possible that such motivation is an artifact of a younger elderly sex offender who is just starting his sexual criminal career. The "experimental" offender could represent a hybrid type of sex offender, meaning that their experience with elder sexual abuse is not necessarily defined. On one hand, it is possible that after experimenting with elder sexual abuse, they will develop a clear sexual motivation that will drive their sexual offending. On the other hand, it is also possible that some of these offenders will consider elderly victims just as a good opportunity for sexual offending. This type is particularly important as it illustrates the complexities associated with elder sexual abuse. It also allows to consider the dynamic aspect of elder sexual abuse and the fact that some offenders may commit a sexual assault against an elderly victim without having to be classified under the three main motivations that were previously identified. The following case example illustrates the experimentally motivated offender: At age 19, Ted sexually assaulted a woman aged 84 years old. Ted was living in the same neighborhood of the victim. The victim, who was a widow and a loner, lived alone. Prior to the assault, Ted had consumed alcohol. Ted surprised the victim alone in her residence and restrained her. During the assault, Ted forced the victim to perform fellatio, he masturbated, and he penetrated the victim vaginally, both with his fingers and his penis. During the assault, Ted tried to negotiate with the victim and to reassure her. As the victim tried to resist during the assault, Ted beat her. Following the sexual assault, Ted left the victim residence. Although injured, the victim survived the attack. The police in charge of the investigation found traces of semen at the crime scene.

Sexual Offender
In addition to the offenders motivated to experiment, our findings showed that some offenders present a sexual motivation to attack elderly victims, thus confirming our first hypothesis. The sexual motivation is evidenced by the offender being involved in different sexual behaviors with the victim. Not only will these offenders commit foreplay and fondling, but similar to the experimentally motivated offender, they are also likely to penetrate both vaginally and anally the victim, which is congruent with our second hypothesis. Moreover, these offenders display deviant behaviors (e.g., exhibitionism, voyeurism, zoophilia) and some of them possess a sexual collection. This is similar to what Burgess et al. (2007) found using the MTC:R3 typology, that is that the main preoccupation of these offenders is the need for sexual gratification, which is often exhibited by a high level of sexual penetration, the possession of pornographic material and the involvement in deviant sexual behavior. After the attack, they are more likely to intentionally release the victim. Offenders presenting a sexual motivation are typically single, live alone, use alcohol and/or drugs, and tend to avoid contact with other people, confirming our third hypothesis. In addition, our findings showed that offenders motivated by sex often present a sexual dysfunction. This is congruent with Jeary (2005) findings who showed that most of these offenders were young men who had broken into elderly residences and who presented sexual inadequacies. According to his findings, many of these offenders are not able to establish relationships-and sexual relationships-with same-aged people, often not being able to have an erection or to ejaculate during a consensual act. Assaulting elder victims becomes for these offenders a strategy to test their sexual performance. The following case illustrates the case of an offender motivated by sex: Robert was 24 years old when he raped a 69-year-old victim who was living alone and had difficulties to move herself. At the time of the offense, Robert was single, and he lived with his parents. Robert had been involved in exhibitionism and voyeurism, especially with older women. He had reported a sexual attraction for older women-i.e., gerontophilia. In addition to his polymorphic sexual deviance, Robert sometimes experienced sexual dysfunctions, more often with having difficulties having or sustaining an erection during intercourse. Prior to the sexual assault, Robert had consumed alcohol and marijuana. Although he didn't know the victim, Robert had targeted her as he knew she was living alone and that she was vulnerable. He broke into her residence and started sexually assaulting her while she was sleeping. He penetrated the victim, both vaginally and anally, also using his fingers. During the attack, the victim tried to resist. Robert did not use physical violence but was able to control her. He also tried to negotiate with her and to reassure her. Following the sexual assault, Robert left the victim's residence and soon after the police discovered traces of semen at the crime scene.

Anger Offender
Other offenders who attack elder victims are clearly motivated by anger. Such offenders typically use physical force, present an affect of anger prior to the crime (see Burgess et al., 2007), and do not hesitate to humiliate the victim during the attack. Such humiliation can be verbal (e.g., insulting the victim) or physical (e.g., forcing the victim to take a degrading position). According to Groth (1979), these offenders entertain a negative perception of their mother, leading them to search for an older victim who represents a surrogate for their mother and on which they will exert their revenge. These offenders do not exhibit signs of planning as evidenced by their use of a blitz approach at a location where they can be seen or even interrupted. However, the assault is always violent as the victim is likely to be beaten, even if she is not resisting the attack, which is consistent with findings from Pollock (1988). This is also similar to Groth's (1978Groth's ( , 1979 description, showing that offenders motivated by anger are characterized by a brutal modus operandi. Although Groth (1979) suggested that the use of a weapon was also typical of the offender motivated by anger, it is possible that this difference with our own findings can be explained by the fact that our sample comes from France, where the use of weapon is less prevalent than in the United States. Nonetheless, both our findings and Groth (1979) show that the focus of this attack is on the use of violence, not the sexual nature of the acts (Burgess et al., 2007). Thus, very few offenders will penetrate the victim, but most will attempt foreplay. These offenders usually live with a partner and some use alcohol and/or drugs. The following case example illustrates an offender motivated by anger: Barry was 30 years old when he committed a brutal sexual assault on a single woman aged 79 years old who lived alone. He was not married, and he lived alone. Although he did not use drugs or alcohol prior to the assault, he was feeling especially angry on that night. During the day, a woman at the bank had refused him a transaction and, according to his own perception, treated him badly. Barry had experienced several similar situations in the past, always involving women. Barry did not know the victim personally but was feeling very angry at women-all women. The night of the crime, he decided to break into her residence from a partially open window while the victim was watching TV. He immediately jumped on the victim, severely beat her, and started to asphyxiate her. He masturbated, forced the victim to perform fellatio on him and decided to urinate on her. Following the sexual assault, Barry left the victim's residence while the victim laid unconscious on the floor. The police recovered forensic evidence at the scene of the crime.

Opportunistic Offender
Finally, some offenders of elder victims seem to be motivated by the simple presence of an opportunity. These offenders are not familiar with the crime scene; they have not targeted the victim, who is a total stranger. The opportunistic character of the attack is also evidenced by the fact that the victim will resist the assault, which can take place where someone could see and/or interrupt, in a public area, and outdoors. This is in accordance with the findings from Burgess et al. (2007) who have suggested that opportunistic offenders are impulsive with a low level of planification and premeditation. According to Jeary (2005), these opportunistic rapes of elderly people could occur during the course of a burglary, where the main motivation was to burglarize residences to finance their drugs and/or alcohol consumption.
The total lack of planning is also confirmed by the absence of penetration; however, these offenders will get involved in foreplay and fondling with the victim (Burgess et al., 2007). These offenders are typically older (at least compared with the other three groups) and most of them are single. The following case example illustrates the offender motivated by an opportunity: Richard, 39 years old, sexually assaulted a woman aged 65 years old who was married and had an active social life. At the time of the crime, Richard was also married, and he lived with his wife. Richard was hanging out at a park located in a residential neighborhood where the victim lived. Richard only noticed the victim when she parked her car. It was the first time he saw her. As the park seemed completely deserted, he decided to surprise the victim from behind and threatened the victim to follow him in a wooded area of the park. Richard kissed the victim, sexually touched her, masturbated, and forced her to perform fellatio. A man who was walking his dog saw the assault, which forced Richard to leave the victim there and flee the crime scene. The police were not able to recover any forensic evidence at the crime scene.

Conclusion
Although the heterogeneity of sex offenders is a well-established fact (Knight & Prentky, 1990), there exists a group of sex offenders who have escaped much scrutiny. Sex offenders who target elderly victims share one characteristic with child molesters: They both target potentially vulnerable individuals. This is very important considering the fact that the elderly population is expected to grow in the coming years, with the combined effect of baby boomers aging and longer life expectancy, especially for women. This suggests that cases of elder sexual assaults may increase as well.
Our findings allowed to confirm our three hypotheses: elder sexual abusers present different motivations and these motivations are associated with specific modus operandi and offender characteristics. This is important at different levels. First, this heterogeneity of motivation involved in sexual assault of elderly victims suggests that it may be possible to identify different treatment targets for these offenders. Although the motivation is an aspect usually hard to uncover, our results show that with the right modus operandi characteristics, it is possible to identify different motivations for offending. Based on these behaviors, it is thus possible to infer offenders' motivations. These motivations could help clinicians working with sex offenders to better understand the offending process and may focus some interventions. For instance, an offender motivated by anger could probably benefit from an anger management intervention, also focusing on his hostility against women.
Another implication of our findings relates to risk management. Considering the high level of violence involved in sexual assaults of elderly victims, it becomes not only necessary to assess the risk of recidivism but to assess the risk that recidivism becomes lethal (Beauregard & Martineau, 2017). Once again, the identification of various motivations may help clinicians involved in risk assessment to identify those offenders who are more at risk of killing their victims.
Finally, another implication of our findings relates to the criminal investigation. As we established, it is possible to link specific offender characteristics to the various offending motivations. Such results, if replicated in future studies, could become useful for the police when trying to prioritize suspects, especially as many of these cases involve stranger offenders.
Despite all these interesting findings, our study is not without its limitations. Limitations of the current study can be found at three different levels: police data, analysis, and the operationalization. Official data from police files present several limitations (Aebi, 2006;Chopin & Aebi, 2019, 2018. First, it is well-known that police data are only available for cases that have been brought to the attention of the police. The dark number of sexual assault in France is relatively high (Chopin, 2017), and it is reasonable to assume that several cases of elder sexual abuse are never reported to the police. Therefore, our results are not generalizable to all cases of elder sexual abuse. Second, as mentioned previously, the nature of the data is likely to influence some of the findings. Considering we used police data, stranger offenders are overrepresented in this sample and this should be taken into consideration when interpreting the current findings. Third, given the purpose of the database used for the study, our sample had no intrafamilial cases, which, once again, is not representative of elder sexual abuse. Fourth, previous studies on elder sexual abuse have shown that some of these sexual assaults occur in care facilities. Unfortunately, we did not have this specific information in the data, making it impossible to examine this specific context of elder sexual abuse.
As to the analyses, due to the relatively small sample size, we had to limit the number of variables to include in our analyses. It is possible that the inclusion of more variables would have allowed to identify more types of offenders, or at the very least, to better detail the ones we have identified (e.g., vindictive vs. pervasive anger). Finally, as to the operationalization, considering that there is no consensus as to what constitute an elderly victim-i.e., what is the right age cut-off-comparisons with previous studies become difficult. Although we have remained conservative in using 65 as the cut-off, it is possible that this is too low, or that with a longer life expectancy, this cut-off will need to be readjusted and placed higher in the near future.
Future studies need to consider a different age cut-off to see if different findings emerge when focusing on older victims (i.e., 75 or more). Moreover, future studies need to replicate the existing results we have on elder sexual assaults. For instance, Safarik et al. (2000) mentioned that the police were using the finding that the age of the offender and the victim were negatively correlated. Although interesting and potentially useful, such finding was not replicated in the current sample from France. Therefore, it becomes important for future studies to consider different samples from different countries and from different data sources to acknowledge the diversity of contexts.