Studies exploring crime from the offender’s perspective is of crucial importance to the formulation of both theory and policy (Wright and Bennett 1990). Some researchers have neglected this perspective because: this kind of research is time consuming; such an approach requires that researchers deal directly with offenders, something that many criminologists have been reluctant to do; and the use of a research design that allows offenders to speak for themselves was considered “unscientific” by positivists (Bennett and Wright 1984).

Nonetheless, it has been noted that the face-to-face interaction which is part of collecting this information represents “the fullest condition of participating in the mind of another human being” (Lofland and Lofland 1984, p. 12). According to Ward and Hudson (1998), this tendency of most theorists to overlook the descriptive or microtheoretical level is unfortunate, as “such models serve to identify possible clinical phenomena and therefore describe the puzzles that subsequent theory sets out to explain” (Ward and Hudson 1998, p. 55).

Most of the work using the offender’s perspective to investigate decision making of criminals has been limited to property crimes. Such studies have been conducted on burglars (Bennett and Wright 1984; Cromwell et al. 1991; Rengert and Wasilchick 1985; Walsh 1986; Wright and Decker 1994; Wright and Logie 1988), robbers (Feeney 1986; Petrosino and Brensilber 2003; Walsh 1986; Wright and Decker 1997), shoplifters (Carroll and Weaver 1986; Cromwell et al. 1999), and auto theft (Fleming 1999). Only a few studies have been completed that aim at building descriptive models of the offense process directly from data provided by sex offenders.

Descriptive Models of Offense Processes in Sex Offenders

Using a qualitative approach, descriptive models of offense processes in child molesters (Ward et al. 1995) and in rapists (Polaschek et al. 2001) have been developed. Grounded theory consists of concepts that “are inductively derived from an initial set of qualitative descriptions or scripts, which, once coded into rudimentary categories, lead to the collection either of more descriptions or of quantitative data” (Ward et al. 1995, p. 454). Their models of offense processes provided a description of the cognitive, behavioral, motivational, and contextual factors associated with a variety of sex offenses. Although interesting and informative, these models present several limitations. First, as noted by the authors, their samples were small (26 child molesters and 24 rapists). Considering the heterogeneity of sex offender populations (Knight and Prentky 1990; Proulx et al. 1999), small samples may not contain a representative range of sexual assault behavior. Second, research on offense processes should not rely on study of the offender alone, but also on the entire criminal event, combining analysis of the offender, the victim, and the context (Meier et al. 2001). Third, these offense process models were too general, and neglected to describe in detail the different components of the modus operandi exhibited by sex offenders during the offense.

Hunting Process

Rossmo (1997) proposed that the hunting process of offenders include the search for a suitable victim and the method of attack; the first determines the selection of victim encounter sites, and the latter, attack sites. Rossmo identified four victim search methods in a sample of serial murderers. The crimes of the hunter (31.6%) are generally committed within the offender’s city of residence. The offenders set out from their home base, and search for suitable victims in the area within their awareness space. The poacher (54.8%) commits crimes by traveling outside its home city, or by operating from an activity node other than its home base. The troller (11.6%) is an opportunistic offender that encounters its victims through routine activities. Its crimes are often spontaneous, although sometimes this type of offender may have fantasized or planned its offense in advance so that he is ready when an opportunity presents itself. Trappers (1.9%) have an occupation or a position where potential victims come to them, or they can entice suitable victims into their home or a location they control by means of subterfuge (e.g., placing want ads, taking in boarders) (Rossmo 2000).

Three different types of offender attack methods were identified by Rossmo (1997). The raptor (78.7%) is characterized by attacking almost immediately upon encountering its victims. This is the most frequently used method for predators. The stalker (0%) follows, watches its victim, and waits for an opportune moment to attack. The attack, murder and victim release sites are thus strongly influenced by the victim activity space. Finally, the attacks of the ambusher (21.3%) are committed someplace where the offender has a great deal of control, such as its residence or workplace. This offender sometimes hides the bodies of the victims, most often on its property. Although Rossmo’s typology of hunting patterns illustrates well how the search for a suitable victim and the attack method influence each other and the geography of the crime (encounter sites and disposal/or victim release sites), no explanation is provided as to why offenders choose one method over another.

Many studies on serial rapists have focused on a quantitative description of modus operandi and personal characteristics of these criminals so that they can be used in offender profiling (Hazelwood and Burgess 1987; Hazelwood and Warren 1989, 1990). Although informative and useful, these studies neglect the cognitive processing of sexual aggressors during the commission of their crime. The use of a rational choice approach is an efficient way to further comprehend not only the “what” and the “why” behind criminal actions, but more importantly, their underlying rationality.

Rational Choice Perspective

In the rational choice approach to crime, the most important dimension is the presumption of rationality of human action, criminal or not (Cornish 1993). “Its starting point was an assumption that offenders seek to benefit themselves by their criminal behavior; that this involves the making of decisions and of choices, however rudimentary on occasion these processes might be; and that these processes exhibit a measure of rationality, albeit constrained by limits of time and ability and the availability of relevant information” (Cornish and Clarke 1986, p.1). Criminals will decide whether or not to commit a crime by weighing the efforts, rewards, and costs involved in alternative courses of action. This rational evaluation may seem limited or “bounded” in some crimes by time, effort, and available information; however, even violent criminals, including sex offenders, will exhibit a substantial degree of rationality (Rossmo 2000). According to Cornish (1993), in such a perspective, it becomes important to study situational variables such as crime scene and victim characteristics, and their choice structuring properties, in order to emphasize the distinctive nature of different person–situation criminal interactions.

Aims of the Study

The aim of this study is to provide a descriptive model of a specific component of the offense process, namely the hunting process of serial sex offenders from the offenders’ account. Based on the rational choice perspective, the model presents a description of the decision making surrounding the hunting process used by serial sex offenders to commit their crimes. Finally, the model may be employed to better understand the criminal behavior of serial sex offenders and thus inform clinical practice, crime prevention, and offender profiling.

Method

Participants

The participants in this study included any individual who committed two or more sexual assaults or sex-related crimes (e.g., sexual homicides) on a stranger victim (i.e., offenders with no personal relationship with the victim prior to the day the offense was committed) of any age (there is no universally accepted definition of serial sex offenders). Events, such as incarceration, that interrupt a series do not preclude offenses taking place before or after the said event from being included in a series (Alston 1994). The serial sex offenders selected were all incarcerated in a Correctional Service of Canada penitentiary (institution where inmates serve a sentence of 2 years or more), located in the province of Quebec, Canada. They were recruited from institutions of different security levels in order to obtain a broad range of offenders, and therefore a wide range of responses. A record was kept by the first author of all sex offenders convicted of a sentence of greater than 2 years in the province of Quebec, Canada, between 1995 and 2004. This list was subsequently used to identify and select the offenders in this study. Among more than a thousand sex offenders, 92 matched to our criteria, and 72 agreed to participate in the research, though only 69 provided qualitative material for the analysis. Among the 20 participants who were not included, only 9 offenders refused to participate in the study; and the other 11 were not available to meet because of their mental state, discipline problems, or because of a transfer to another institution. The majority of our sample was white (91.3%); they were on average 30.7 years old when they started their series of crime (S.D. = 9.4). Almost half (46.4%) of the offenders were married or in a relationship at the time of their crimes. The majority (89.9%) were French speaking, 39.6% were unemployed, and only 10.1% had no prior criminal record before starting their series of crimes. For those with a record, they had an average of 1.0 criminal charge (S.D. = 3.1) for sexual nonviolent crime, 2.9 charges (S.D. = 6.3) for sexual violent crimes, 11.9 charges (S.D. = 19.6) for nonsexual and nonviolent crimes, and 2.5 charges (S.D. = 4.4) for nonsexual and violent crimes. It is also noteworthy that 39.1% of our sample have never owned a car.

Procedure

The rational choice perspective provided the main set of theoretical principles used to guide our data collection. A specially constructed instrument was built to collect information from police investigative reports (present in the institution’s case file) and to conduct in-depth semi-structured interviews. This questionnaire includes five sections that permit the collection of information on pre-crime factors, hunting patterns, modus operandi, post-crime factors and geographic behavior. The main advantage of semi-structured interview over other methods is that it allows subjects to speak freely and at length using their own concepts and terminology. It can also be conducted in a relatively informal, nonthreatening manner, providing the researcher the opportunity to develop a relationship of trust and confidence (Bennett and Wright 1984). The reliability of responses in our study was monitored by checking for and questioning inconsistencies. However, most of the subject matter covered in the interviews was not included in official reports. Nonetheless, one study comparing the responses of offenders to official records revealed close agreement between the two (West and Farrington 1977). In order to minimize response distortion, offenders were promised complete anonymity and confidentiality, and a guarantee that their information provided could not be used in any way against them by the Correctional Service of Canada. Inmates, however, were told that if during the course of the interview, the name of a potential victim or someone who is in danger was brought up, the interviewer would have an obligation to inform the concerned authorities. Interviews were conducted in a private office, isolated from correctional staff and other inmates. They lasted from 2 to 12 h, depending on the number of crimes committed and the participants’ verbosity. Due to the sensitive nature of the conversations, permission was not requested to tape record the interviews, although extensive verbatim notes were taken whenever possible. No participant was paid for participating in the study.

Data Analysis

As dictated by the rational choice perspective, the serial rapists’ hunting process has been broken down into a series of sequential events or steps that allow exploration of the different aspects of their decision making related to their assaults. The rational choice perspective has been used as a heuristic device to frame offender decision making regarding their hunting process (Petrosino and Brensilber 2003). Following Clarke and Cornish’s (1985) decision making model, qualitative responses from serial sex offenders were organized into two major areas: (1) victim search methods, and (2) offender attack methods (see Fig. 1).

Fig. 1
figure 1

Rational choice heuristic

Full size image

Results

Victim Search Methods

Phase 1: Routine Activities of Offenders and Victims

Offenders were asked about their routine activities just before the commission of their crimes. Routine activities are defined as the activities and behavior engaged in by people on a regular (daily or weekly) basis (Rossmo 2000). Brantingham and Brantingham (1984) note that “most criminals are predominantly non-criminals; that is, they spend most of their time in non-criminal pursuits” (p. 350). Phase 1 of the model shows that a majority of serial rapists were engaged in non-criminal routine activities prior to their crimes. Table 1 reports that 25% of the offenders were engaged in an occupation where they could come into contact with potential victims. Such occupations include work in facilities that have potential victims in vulnerable positions (e.g., hospitals, nursery homes, kindergartens, babysitting), and team sports coaching or other activities that attract potential victims (e.g., bike repairing, operating a convenience store). Some offenders (20%) reported that they were engaged in an inside recreational activity, such as drinking and/or taking drugs at home, at a bar or a party, or watching a movie (“I was in a bar to celebrate getting my income tax check and was drinking a few beers”). Some serial sex offenders encountered their victims while traveling from one place to another (17%; “I was on my way home from a bar and I noticed a girl walking alone”). Others (3%) were engaged in drug-related activities (e.g., waiting or looking for drug dealer), and 6% were not able to recall what they were doing prior to committing their crimes. For a significant proportion of serial sex offenders (42%), however, most of their time was spent prowling for victims, working towards achieving their sexual goals. While they were seemingly engaged in normal activities such as shopping, walking, and driving, they were really searching for a potential target (“I was walking through the different stores in the mall looking for a good looking girl that I could attack”). These offenders often prowl for long hours before selecting a victim corresponding to their suitability criteria. This is in agreement with the findings of Davies and Dale (1995) who reported that several serial rapists in their sample were extensive prowlers, some of whom covered relatively long distances using a vehicle or public transportation. Ouimet and Proulx’s (1994) study on serial child molesters suggested that because some of these individuals lack occupational stability, they spend a disproportional large amount of their time and energy on making contacts with potential victims and as such, their routine activities may be determined by the opportunity structure associated with their crimes (note that the percentages may add to more than 100 because of multiple responses).

Table 1 Variables related to victim search methods
Full size table

The activity in which victims were most often engaged in prior to a crime occurring was traveling from one place to another (39%; see Table 1). Victims were leaving work, driving home after shopping, hitchhiking, getting off bus, walking home, or leaving a bar. Other victims (25%) found themselves in the offender’s home or workplace prior to the crime. Offenders were able to attract the victims to their homes and workplaces by a subterfuge or manipulation, or were simply put in contact with them through offender’s occupation. In 22% of the cases, victims were home, most often sleeping. Another 22% of the victims were participating in outdoor recreational activities (e.g., playing outside on the street, in a park or at a swimming pool), whereas in 20% of cases, victims were engaged in indoor recreational activities (e.g., in a bar, at a rock concert, at a party, or shopping). Finally, in 12% of the cases, victims were at work, either as clerks in a store or business, or on the street as prostitutes.

Results from Rossmo’s (2000) study on serial murder showed that most of the victims were at home (30.9%), engaged in prostitution activities (22.5%), or walking/jogging (21.9%). No victims were at an outdoor recreation site and only 1.1% were at work. Some victims were at a bar or nightclub (10.7%), at another type of social event (5.1%), or commuting (6.2%). Victims of our sample were engaged in similar routine activities than the victims in Rossmo’s study, but in slightly different proportions. It can be hypothesized that serial sex offenders differ from serial murderers on their choices of targets, which will influence their hunting fields.

Phase 2: Choice of Hunting Field

Hunting fields designate the type of area where offenders hunt for victims. It must be distinguished from hunting grounds defined by Rossmo (2000) as the territory (a geographic area) within which an offender searches for victims (Rossmo 2000).

More than half of the subjects (57%) interviewed reported hunting in specific places to find a victim. As can be seen in Table 1, six different hunting fields were identified from the answers given by offenders. Hunting fields used by serial sex offenders were most often acquired through local visibility (57%). In these cases, the offender contacts the victim in a public place often frequented by the offender himself, such as a bar, a park, a shopping mall, or on the street. Offenders explained that this choice of hunting field was motivated by the fact that it is an easy way to find attractive targets, to gain access to them, and to estimate their vulnerability. Some of these places were also known by offenders to be both isolated and attractive to potential victims (“When I talk to women I get rejected because I don’t look good. When I go to the mall, I see plenty of good looking women, so instead of being told ‘no’, I jump them on their way back to their cars ”). Further, some offenders (19%) decided to search for victims through their own occupation. These offenders get involved in a type of work, volunteering or leisure activity that brings potential victims to them (e.g., hockey coach, janitor, owner of convenience store, kindergarten teacher). According to the offenders, their occupation serves as an effortless hunting field because the victims, most often children, come to them on their own. Moreover, because of the offenders’ positions, victims were in a vulnerable situation and could be easily manipulated, reducing the risk of the crime being reported. (“I told them that I had the power to cure their sicknesses and help them in their studies. They’d come over, I’d have them smoke up and tell them that they’d have to release all tension and take off their clothes. I could then touch them and they’d think that it was part of the healing”. “At the daycare, the kids are there, available and very vulnerable, especially during their afternoon naps. Parks and schoolyards demanded too much effort and was too risky because anything could happen.”)

Some serial sex offenders hunt for victims through families (16%). The offender first gets acquainted with a family (most often with a female living alone with her children or vulnerable because of a drug and/or alcohol problem), and then offers different services, particularly babysitting, in order to gain access to the children. Offenders explained that it was easy to find single mothers who need some kind of help, especially in poorer neighborhoods. By offering their services and not requesting anything in return, the offender is able to gain the mother’s trust and get access to her children. The offender invests some time in the pseudo-relation with the mother, and then exploits any opportunities to offend.

A few serial sex offenders (13%) targeted their victims by hunting in private or semi-private places. According to Rossmo (2000), “there is a hierarchy of spaces that influence criminal action: public space (e.g., street), semi-public space (e.g., open front yard), semi-private space (e.g., fenced backyard), and private space (e.g., house). As an offender moves from the street to an apartment building parking lot, to inside the building itself, and to the interior of an individual apartment, he is progressively entering more private space and concomitantly increasing his risk” (p. 124). The victim’s home, most often an apartment, was the type of private place preferred by offenders, whereas elevators or laundry facilities found in apartment buildings were typical of semi-private places. Offenders explained this choice of hunting field by its highly arousing quality (e.g., thrill, excitement). Finally, two offenders (3%) reported hunting for victims through prostitution areas and two others (3%) through newspaper advertisements. The choice of the hunting field is directly influenced by the selection of a specific victim type (i.e., prostitute) for the former. For the latter, advertisements were a means to meet potential victims as the offenders did not know any potential victims in their own neighborhood.

For most of the serial sex offenders in our study, hunting fields were characterized by what Brantingham and Brantingham (1995) call crime attractors, places that attract offenders through their reputation for crime opportunities that trigger their criminal intent. In his study on 357 cases of serial murder, Newton (1992) found such killers hunted their victims in public places (31%), in homes or private places (16%), in the victim’s home (10%), on prostitution strolls (5%), in occupational settings (5%) (e.g., hospitals, nursing homes, clinics), at work (3%), on the road with hitchhikers (1%), and through classified advertisements (0.5%). In 23% of the cases, serial murderers used different hunting fields to find a victim. These results are similar to ours.

Phase 3: Victim Selection

Targeting of a specific victim was carried out by 51 offenders (74%). Nine offenders (13%) could not recall or did not want to answer the question. Non-random victim selections, or victim selection based on specific criterion, requires more searching on the part of the offender as compared to random victim selection.

Seven factors have been identified as being related to victim selection. The most important one was the location and availability of the victim (45%). For offenders, easy victim access, low socioeconomic status neighborhoods, geographic isolation, and houses with large windows where the victim’s behavior can be observed from outside, are all helpful features (“she was there, all alone and no one was around”). Other offenders mentioned some of their victims were outside (e.g., leaving a bar, waiting for the bus), but close to their homes.

The general physical appearance of the victim was the second most important factor (38%). The offenders explained that they chose their victim because he/she was beautiful, sexy, had blond hair and blue eyes, dressed in a specific fashion (skirt, shorts, bathing suit), delicate, tall, and with nice skin. However, only 7% of offenders mentioned that the victim’s sexual physical appearance played a role in their choice of target. For these offenders, the victim had to have big breasts, a nice figure and no pubic hair. As one offender said, “she had to have big breasts, otherwise I didn’t get excited, I wouldn’t get hard.”

The victim’s vulnerability is another important factor associated with choice of specific victims (30%). Features associated with victim vulnerability include : young, alone or alone with children, on the street (hitchhiking), naïve, handicapped, and fragile. When the victim was a child, most often a young boy, vulnerability is associated with specific characteristics, for example, a child with family problems, without supervision, always on the street and in need of help (“when I saw a kid that always had a key around his neck and who was badly dressed, I knew that his parents didn’t take care of him and that I could easily approach him to offer him things... I knew that he’d accept because he had nothing”).

For some offenders (22%), the victim’s age (pre-pubescent, pubescent, or adult) was an important factor in their choice, whereas for others (22%), it was the victim’s personality. Victim personality types of interest to offenders include “easy girls”, or a caring, affectionate, and extraverted victims willing to smile. Finally, 10% of offenders claimed that it was the victim’s behavior which guided their choice. Behavioral characteristics reported include a victim who dances and moves well, victim exhibitionists, and victims who talk sexually “dirty” or in an open manner.

Stevens (1998) investigated the rationale behind certain characteristics of the modus operandi and the hunting patterns of serial rapists from the offender’s perspective. Using a grounded-theory approach, the author surveyed 61 serial rapists on their targeting techniques and identified three rationales of reason for victim selection. Most rapists (69%) mentioned vulnerability as the strongest reason to attack a female. Youth, helpfulness, submissiveness, and females who decrease their defense capabilities (placing themselves in risky situations such as getting drunk at a party) were reported characteristics of vulnerable targets. For 15% of the sample, the situation provided the main reason for targeting a specific victim. Some offenders seek specific situations or circumstances, as opposed to specific individuals, and generally encountered their victims in particular social environments, such as at a place of employment. Finally, 11% of the serial rapists reported that encounter of a random victim was the reason to assault her. While these offenders were engaged in other criminal activities (e.g., burglary, robbery), they came upon a victim by chance and spontaneously assaulted her (Stevens 1998). A study conducted by the FBI’s National Center for the Analysis of Violent Crime (NCAVC), which included interviews with 41 incarcerated serial rapists responsible for more than 837 rapes (Hazelwood and Warren 1989) showed, however, that the criterion most reported by serial rapists for target selection was victim availability (98%), gender (95%), age (66%), location (66%), and race (63%). Physical characteristics were reported by only 39% of offenders, similar to the results found in our study.

Interestingly, 28% of offenders admitted that in some cases they had targeted another specific victim but were unable to assault her. Reasons for such changes of plan were diverse. For example, some rapists claimed that the situational factors were just not right because the victim was not alone, the victim was not easily available, someone nearby might interfere, or simply that the location was too risky. For others, the victim’s behavior made them change their mind. In one case, the victim’s speech prior to the crime lowered the offender’s sexual arousal, whereas in another case, the offender lost his patience and left before the victim became available (“I was waiting for her, she had finished working but she didn’t want to come out, she wouldn’t stop talking with the other one... I couldn’t take it anymore, I was fed up and left”). Moreover, in some criminal events, victims were able to escape from offenders prior to the sexual assault. In explaining why he did not assault the selected victim, one offender claimed that he was able to resist his desire to offend.

Although the part of our model related to victim search methods is presented in a linear fashion, it should be seen as a dynamic process, much like the dynamic model of crime site selection developed by Brantingham and Brantingham (1978). In their model, the motivated offender is influenced by past criminal events and by the site and situation of a potential crime, and is triggered by events typically part of daily routine activities, and shaped by the environment, past experience, and the crime templates. These templates are the set of cues, cue sequences, and holistic cue clusters used to identify suitable targets, and that are incorporated in the target selection either at a conscious level or automatically. When the template is constructed, it becomes relatively fixed, and influences future search behavior, thus becoming reinforced. The templates present similarities that can be identified because the spatial and temporal distribution of offenders and victims are patterned and because human environmental perception has certain universal properties. Finally, targets are selected from the offender’s awareness space. Awareness space is defined as “all the locations about which a person has knowledge above a minimum level even without visiting some of them... Awareness space includes activity space (the area within which most of a person’s activities are carried out, within which the individual comes most frequently into contact with others and with the features of the environment), and its area enlarges as new locations are discovered and/or new information is gathered” (Clark 1990, pp. 24–25). The targets will then be assessed against the criteria of suitability and risk, and evaluated in relation with the offender’s template. Finally, rational choices are made by the offender and specific targets are selected (Rossmo 2000). Our results are congruent with this model as offenders, through their routine activities, will encounter potential victims and select them in places suitable for an attack. These hunting fields, if successful, will be reinforced and will be used again during the search process for suitable victims.

Offenders’ Attack Methods

Phase 4: Method of Approach

Once a victim has been targeted, the serial sex offender uses different methods of approach. The results in Table 2 show that the preferred method used by serial sex offenders is the trick/false identity approach (48%). Offenders explained that such a method allows them to gain the victims’ trust, avoiding scaring the victims and hurting them physically. Moreover, some offenders explained that the trick/false identity method was only used to get close physically to the victims, to gain access to them, and sometimes, to distract the victim just before the attack. One offender explained that:

when I saw a girl that I wanted to attack, I would ask her for the time and when she looked at her watch I had my chance to push her right away and then rape her.

Table 2 Variables related to offender attack methods
Full size table

Some serial sex offenders are far more violent and will use physical violence from the start in their victim approach (25%). Offenders explained that violence helps to surprise the victim and prevent them from resisting or escaping. Some offenders use this method to make the victim lose consciousness and then do whatever they want to them; for others, it is only for the thrill. For only a few offenders (9%), threats are used alone or in conjunction with physical violence to approach the victims. Threats are thus mainly used to scare the victims, to “freeze” them, to prevent them from alarming witnesses or resisting, and to make sure they will be completely submissive during the sexual assault.

Three methods are used by sexual offenders specifically against children: seduction/persuasion (13%), money/gift (16%), and games (9%). These methods help offenders make contact with the victims slowly and to gradually estimate their chance of succeeding in getting the victim involved in sexual activities. Offenders have mentioned that by acting in such a manner they were able to get the victim to consent to the sexual contacts and to regularly return to them, often with other potential victims.

Acting directly on the victim (15%) is a method that serial sex offenders use to take advantage of the surprise effect, often because they are not able to otherwise approach the victims (e.g., by talking to a woman). These offenders mentioned that they do not want to risk being rejected, or they simply wanted to take advantage of the fact that the victim was sleeping. It is interesting to note, however, that 36% of serial sex offenders use more than one method to approach their victims (“it always depended on the victim, when they were sleeping, I’d touch them directly, otherwise I would do it through a game because its more subtle, otherwise, with the babysitter, I pretty much just jumped her”).

Ouimet and Proulx’s (1994) study perfectly illustrates the switch of methods used to approach victims. The authors argue that if child molesters are not able to find a suitable victim near their home, they may have to travel further. As they journey greater distances, convincing children to return home with them becomes harder (few are willing to take a car trip with a stranger). The offender has to adapt his hunting strategies and use a more direct approach, such as a direct attack or an ambush, which in turn can lead to an increase in the level of violence of the crime.

Hazelwood and Warren (1990) found the surprise approach, which involves the rapist waiting for the victim or subduing her while she was sleeping, was the most frequently used. The con approach, the second most commonly used by serial rapists, involves subterfuge and the ability of the offender to interact with women. The blitz approach (a direct, physical assault that subdues and injures the victim) was the method used least often. These results are different from our findings and one explanation might be that Hazelwood and Warren (1990) only focused on adult female victims. It can be hypothesized that different types of victim will trigger different strategies to approach them.

Lebeau (1987) focused on how the journey to rape varies as a function of the offender’s approach method. He studied 320 cases of rape committed by single and serial offenders. He found that offenders, both single and serial, traveled the shortest mean distance to assault their victims when they illegally entered the victim’s residence. This suggests that offenders travel shorter distances when using a method linked to crimes against property. Thus, in relationship with our results, this could mean that offenders who break into the victim’s house live nearby.

Phase 5: Attack Location Choice

Serial sex offenders were asked why they assaulted their victims where they did. Their answers were varied. For some offenders (23%), the only reason for the location was because the victim was there. These offenders admitted that the situational and environmental cues were not taken into consideration when they decided to attack the victim (“the victim was there and I didn’t even see the witnesses that were close by, I didn’t care, I just wanted to grab her boobs”). A few others (3%) mentioned that they had no choice but to attack the victim at that particular location because he or she was going to escape or be no longer available (“I had to attack her there because she was getting in her car and she’d be gone, I didn’t have a choice on the place”).

Many serial sex offenders were selective in their choice of attack location. Table 2 shows that 35% of offenders have specifically chosen an isolated area (e.g., the offender’s home) to attack the victim in order to prevent being seen or disturbed by witnesses, and to have some time alone with the victim (“I always hung out in the woods near the bike path, plenty of people went through there because it was a shortcut. I noticed that women used the path too”). Davies and Dale (1995) observed that such areas are often more hazardous for potential victims but safer for the offender, their disadvantage being smaller numbers of available targets. Fisher (1980) suggested that the choice of an isolated area to attack a victim might be indicative of an intention to injure. Ouimet and Proulx (1994) observed that in the case of child molesters, the offender’s home appears to be the best possible location to commit an offense because it offers several advantages over competing locations. In a home, the child might feel more secure and more willing to participate in sexual contact, allowing the offender to engage in satisfying sexual activities. Child molesters are also skillful in manipulating children in order to keep the victim from telling someone about the sexual contacts (Leclerc et al. 2005). These factors might explain why most child molesters hunt for targets in the immediate vicinity of their residence as the shorter the distance from the victim encounter site to the offender’s home (the attack site), the greater the likelihood of the offender convincing the victim to voluntarily accompany him.

Another group of offenders (12%) revealed that their choice of attack location was mainly motivated by places known to be frequented by victims. In these cases, the offender is looking for a specific type of victim and he knows where he can find them (“for a whore, you go to the whore district”). Attack location choice has also been influenced by the availability/accessibility of the victims. They explained that they engaged in their offending process not only because the victims were there, but because they could have easy access to them. Finally, 12% of offenders reported not knowing or having forgotten what was the reason for their attack location choice. Hazelwood and Warren study (1989) found that the scene of the sexual assault was relatively consistent from one rape to another; in 50% of the cases, the rape occurred in the victim’s home. In 6% of the cases, the sexual assault occurred on the street, while in another 6%, the rape was committed in a parking lot or a highway. Few offenders attacked their victims in their own residence, a public facility, or at the victim’s place of work. Again, these inconsistencies in results may be linked to the different types of victims involved in serial sex crimes.

Phase 6: Method to Bring Victims to Crime Sites

The methods used by offenders to bring a victim to the crime site vary slightly from the methods to approach a victim, mainly because in 41% of the cases the victims are not moved. Some offenders (9%) use seduction/persuasion, money/gift, or threats to bring the victim to the crime site. Only 3% of the offenders used games or acted directly on the victims without any strategy, whereas 35% tricked the victims. Use of physical violence only occurred in 17% of the cases. Serial sex offenders (33%) sometimes used more than one method to bring the victims to the crime site. The rationale for these methods was similar to that for the method of approach.

Phase 7: Crime Location Choice

As almost half of the participants (41%) did not move the victims between the attack and the crime, the underlying rationale was very similar to that for the attack location choice. Most offenders who moved the victim between the attack and crime location mentioned that they chose this crime location because it was the victim’s residence or it was in an isolated area.

Phase 8: Method to Commit the Crime

The methods used to commit the crime are more violent than the methods used to approach the victims or to bring them to the crime site. Seduction/persuasion (1%), money/gift (7%), and games (6%) were methods used only by a minority of offenders, reportedly most often so as not to scare the victims, avoid using violence, and prevent police reporting (“I just wanted to have sexual contacts with the kid, I didn’t want to scare or hurt him, when he started screaming I stopped right away”). Two offenders (3%) used drugs and/or alcohol to commit their crimes because it helped to decrease the victim’s inhibitions and prevented them from resisting during the crimes. Tricking or using a false identity was used by 15% of serial sex offenders to impress their victims, convince them not to resist and to consent to the sexual contacts, and prevent them from reporting the crimes to the police.

Violent methods were used by more offenders at this stage of the hunting process. Some offenders (25%) reported the use of threats to prevent victims, especially younger victims, from resisting (verbally and physically) and escaping. Offenders explained that the use of threats during the sexual assaults served to scare, intimidate, humiliate and control their victims. When victims were older, more offenders (52%) resorted to physical violence to commit their crimes. Again, physical violence was mainly used to surprise the victims, prevent them from resisting, and to control them during the assaults. For some offenders, physical violence was also a part of their deviant sexual fantasies and the killing of the victim after the rape was their primary goal. Other offenders stated that they wanted to render the victim unconscious, so they could do whatever they wanted to the victims during the sexual assault. Almost half of the offenders interviewed (42%) showed a change in crime commission methods from one victim to another. One serial sex offender reported:

I would always use a knife or my gun, that way I could hold the victim in place and stop her from resisting. I would go into her room and I would stare at my triumph: it’s my victim, I succeeded, this is it. There was a feeling of anticipation, she is going to be mine. I put my hands over her mouth so that she couldn’t scream and I could threaten her with the knife or gun, depending on the distance between us. When I was close it was the knife because the victim would stay still so she wouldn’t get cut. I would also verbally threaten her, shut up or I’m gonna kill you... do as I say and everything will be alright. When I followed the victims I threatened with my gun because it was more effective during the day. When I was on the subway, I would rub myself directly on them or I’d wait to be alone in the wagon and I would jump them... it wasn’t as long that way

Results from Hazelwood et al. (1989) are somewhat different. They found most offenders used a threatening physical presence and/or verbal threats to control the victim and in half of the cases, a weapon (most often a knife) was displayed. The majority of the serial rapists they studied used minimal to no physical force during the rapes. It can be hypothesized that the more frequent use of a weapon may influence the level of violence displayed during the crime.

Stevens’ (1998) found that only a minority of serial rapists exhibited an excessive use of force before, during, and after the attack. These individuals often lacked control during their crimes. Most serial rapists, however, claimed that minimal violence was used, and only then to gain the victim’s compliance. In 13% of the accounts, no violence was used during the rape. Some offenders (16%) described intimidation as a way to shock victims into submission. Finally, nealy half of the subjects used either moderate violence (a blow, slap, push, threat to use a weapon) to gain the victim’s attention, or limited violence (a level of violence more than necessary to gain the victim’s submission). Most stopped their attacks when the victim refused to submit, rather than resort to an escalation of violence (Stevens 1998). As we can observe, results from Stevens are closer to ours in terms of the level of violence used.

Finally, it is possible that the level of violence displayed during the crime commission is influenced by other factors not mentioned by our subjects. For example, Warren et al. (1999) found that serial rapists who sexually assaulted their victims for long periods of time, and used more profanity, were more likely to escalate in their level of violence than sex offenders who did not exhibit such behavior.

Phase 9: Victim Release Location Choice

This crime process ends with the victim release. The majority of offenders (79%) stated they left their victims at the crime scene after the sexual assault was completed. Either the offender or the victim left after the sexual assault was completed (“I was finished so I got out of there, I think that she stayed to get dressed, I didn’t care”). Some offenders had no choice as to the victim release location because the victims were able to escape, they resisted too much (9%), or because the offenders were interrupted or surprised by witnesses or bystanders (7%). Other offenders, however, were careful in their choice of victim release location. In the sexual homicide cases, offenders disposed of the victim’s body in an area where it would not be discovered (7%; e.g., a wooded area, in the garbage).

I had to get rid of the body because it would smell in my apartment. I had a large box that I wasn’t using so I put her in and put all kinds of garbage on top so that no one would feel like going through the box. Afterwards, I threw it in the garbage. When the garbage men came, I was nervous, I was scared that they would take a look inside because of the weight. They ended up taking it and they never found it. For the second one, I didn’t have a box, so I decided to cut her into pieces and put the parts in different garbage bags. So that it would look like real garbage, I put my garbage into the bags. Then, I put them in the alley for garbage pick-up like the first time. Except this time, some kids were playing hockey in the alley and one of them opened a bag with his stick and saw the victim’s head. The cops came and they found the bags. On top of it, there was a bill in one of the bags with my name and address on it. They came straight to me.

Some offenders took special care to release the victim far from the offender’s home (4%) in order not to be associated with the victim or so the victim would not recognize the offender’s residence or neighborhood (“I didn’t want her to be able to recognize my place, so I brought her very far, in another neighborhood”). A few remorseful offenders (9%) took the time to release their victims at busy locations where they were sure the victims would be helped and take home (“I felt bad about what I did, and I didn’t want to leave her alone on the road, there was no one around... I took her near the place where she said that her ride was going to pick her up later”).

Discussion

This study explores the hunting process of serial sex offenders based on the offenders’ accounts. “Descriptive models represent the most fundamental level of theory building; they are judged successful to the extent that they capture the essential elements of how offenders go about committing their criminal acts” (Polaschek et al. 2001, p. 537). The major purpose of our investigation was to thus identify a descriptive model specific to the hunting process of serial sex offenders from a rational choice perspective in order to better comprehend the choices made by offenders during this process. This hunting model (see Fig. 2) has proven useful for at least three different reasons. First, the hunting model emphasizes the notion of geography of crime (spatial mobility, choice of crime locations, hunting fields), a dimension that has been neglected in the study of sex offenders (see Beauregard et al. 2005b).

Fig. 2
figure 2

A descriptive model of the hunting process of serial sex offenders

Full size image

Second, it permits a better understanding of the relationships between the geographic and the behavioral components of the sexual assault. For example, the hunting model shows how an approach victim location may influence the type of strategy used by the offender. Third, the hunting model identified permits also to investigate important components in the offending process of serial sex offenders. The hunting model details this specific process of the offending process and explores some dimensions neglected by other researchers in the field of the offending process of sex offenders (see Polaschek et al. 2001; Ward et al. 1995). These include: (1) routine activities of offenders and victims prior to the crime, (2) distinguishing between method to approach, method to bring to crime site, and method to commit the crime, and (3) exploring the different locations associated with the crime (encounter, attack, crime, and victim release).

Although few studies have been conducted on the rationality of sex offending, this study clearly demonstrates that even sex offenders act in a rational, although sometimes bounded, way during the commission of their crimes. Despite the fact that not all serial sex offenders hunt for victims in the same way, our descriptive model shows that the decision-making surrounding the hunting process fluctuates according to offenders’ strategies, type of victim and their reaction, situational context of the crime, and environment. For example, the rationale behind the use of a trick/false identity victim approach strategy may change during the commission of the crime. Moreover, the strategy used to approach a specific victim may vary by contextual setting (different time or location), or by victim type. Our model is in direct accordance with Rossmo’s hunting typology and the dynamic target site selection model of Brantingham and Brantingham (1978, 1993).

Although preliminary, we believe our model may have several implications for clinical practice with sex offenders, crime prevention strategies, and offender profiling. First, the hunting process model could be used in treatment programs in order to better detail the offense chain of the criminal, and it would complement the relapse prevention model (Pithers 1990) and the offense models of Ward et al. (1995) and Polaschek et al. (2001). These models emphasize the cognitive, emotional, and motivational factors associated with the offense, but neglect to detail the offending behavior, including the hunting process.

Second, the different components of the model could be used by crime prevention strategists to adapt victim responses to different behaviors of attackers. Knowing why a sex offender acts in such a manner might help victims prepare different strategies to escape or resist the offender. We can improve on uninformed confrontational or nonresistance advice, by learning from the decision making of serial sex offenders and creating of new advice adapted to the situation and the offending behavior. Our model emphasizes the fact that the majority of victims of serial sex offenders are encountered in public places, where the risk for the offender is high. Also, many offenders use different strategies during the assault to prevent or stop the victim from resisting (verbally or physically) in order not to have to harm physically the victim. This suggests these offenders are concerned the victim might alert someone. It also indicates that some are willing to resort to physical violence. Moreover, it seems that offenders prefer to attack their victims in isolated areas. Finally, our results show that violence is intensified in the later stages of the crime (during the commission of the assault), as compared to the earlier stages. It may be that victims should try to resist the offender during the initial encounter or when he is trying to bring them to the crime scene, instead of when he is about to commit the sexual assault.

Third, the descriptive model may prove useful in both geographic and psychological profiling. Pioneers of criminal profiling have stated that in order to correctly predict the offender’s characteristics, the profiler has to be able to get “inside the criminal mind.” Until now, studies on psychological and geographic profiling have focused on the “how,” the “where,” the “when,” and the “who,” but have neglected the “why”. Our study was not interested in the different motivations of offenders. This is something that has been studied by several researchers (e.g., Groth et al. 1977; Knight and Prentky 1990). The goal of our investigation was to learn why serial sex offenders acted the way they did during the criminal hunting process; in other words, why they committed the assault the way they did it. We investigated the decision-making process of these offenders in order to better understand why an offender acted in a certain way during the criminal event. Criminal investigators must understand the rationale of this “seemingly irrational” behavior in order to better infer the type of criminal, and to establish proactive strategies that might prevent another sexual assault.

Our model also makes a contribution to the theoretical assumptions underlying offender profiling. Results from a previous study (Beauregard et al. 2005a) indicated the modus operandi of sex offenders was not immutable, but instead is a dynamic process more influenced by situational factors than personal characteristics (e.g., sexual interests). The present study demonstrates, again, that serial sex offenders do not always exhibit a consistent modus operandi during their crimes, but rather are affected by the context and situational factors. Our results underline the necessity of including situational factors in any model of offender profiling in order to reliably infer offender characteristics.

There are, however, a number of limitations to this study. First, the small sample size means that we may not have captured the complete range of serial sex offenders’ hunting methods. Second, not much is known about the external validity of our results. For example, the decision-making of incarcerated serial sex offenders may be different from perpetrators who have not been apprehended (Petrosino and Brensilber 2003). Third, the data collection methodology relied mainly on retrospective self-reports. This type of research may suffer from intentional or unintentional retrospective distortion, even if appropriate interviewing techniques are employed to enhance the level of detail (Polaschek et al. 2001). Finally, the methodology we used did not permit us to examine every relationship between each of the components and phases of the model.

In future studies, we plan to identify different scripts that take into account every phase of the model, as well as all relevant situational factors. The script is “generally viewed as being a special type of schema, known as an event schema, since it organizes our knowledge about how to understand and enact commonplace behavioral processes or routines” (Cornish 1994, p. 32). Quantitative methods such as corresponding analysis and clustering techniques will enable us to examine the relationships between the behavioral and geographic components of the hunting process. Further, it is important for other researchers interested in this matter to collect descriptions of criminal hunting processes in order to test the existing model and establish its validity and reliability (Polaschek et al. 2001). Other offense process variables also need to be investigated and incorporated into the model. These should include pre-crime factors (alcohol and/or drug consumption, use of pornography, deviant sexual fantasies, affective state), modus operandi variables (use of weapons, restraints, and/or rape-kits, forensic awareness, acts committed), and post-crime behavior (reasons for leaving the crime scene, attitudes during police interrogation) in order to establish a complete and detailed decision-making model of offense patterns.