An Examination of the Home-Intruder Sex Offender

Abstract

One particular crime location in sexual assaults is the victim’s residence. Similar to sexual burglars, “home-intruder” sex offenders choose to assault the victim in her residence, most likely in their bedroom. The aim of the current study is to analyze modus operandi, temporal factors, and victim characteristics in a sample of 347 stranger sexual assaults committed by 69 serial sex offenders to determine which factors may be more relevant to sexual assaults committed in the victim’s residence compared with sexual assaults committed at another type of location. Our hypothesis is that offenders who choose to sexually assault victims in their home constitute a specific type of sex offender, one that resembles the sexual burglar. Results showed that modus operandi (e.g., burglary), temporal factors (e.g., time at crime scene with victim), and victim characteristics (e.g., age, victim-offender relationship) were significant in predicting whether the victim encounter, crime site, and victim release site were located at the victim’s residence or not. Moreover, these findings were generally significant across the three crime locations, which can be explained by the high consistency in location during home-intrusion sexual assaults. Situational crime prevention strategies aimed at making a residence less attractive for offenders should help reducing this particular type of sexual assault.

Keywords

home-intruder sex offender rape sexual burglary decision making prevention

Introduction

Stories of sexual assault describing offenders breaking into victim’s homes and assaulting the occupants inside cause people to feel unsafe in their homes and neighbourhoods (Francis, Soothill, & Dittrich, 2001; QMI Agency, 2012), and news reports describing cases of these can be found across various countries (e.g., Dufresne, 2017; “Vancouver Police Seek High-Risk Offender Antoine Naskathey,” 2017). Stranger sexual assaults can pose a particularly challenging task for law enforcement because there is no obvious tie between the offender and victim to help the police narrow down possible suspects. Given the serious nature of sexual crimes, it is essential to understand where, when, and who are committing these offences and to develop strategies that may prevent future offences from occurring.

Decision Making in Crime Site Selection

Criminological theories have tended to focus on criminality rather than the crime event itself to understand why an offender was motivated to commit their crimes (Deslauriers-Varin & Beauregard, 2010). However, the location of the offence can also provide crucial details for understanding a crime event. Research examining offenders’ decision-making processes for target and crime site selection have often focused on burglaries. It has been argued that burglars follow a spatially structured and sequential decision-making process when selecting their targets (Clarke & Cornish, 1986; Taylor & Gottfredson, 1986). They begin by selecting a suitable area and then they compare potential targets to determine which site is the most suitable (Bernasco, 2006). An interesting hybrid offence that appears to include both a burglary and a sexual-based assault has been identified by Warr (1988)—the home-intrusion rape—and illustrates the need for research on crime site selection to be expanded to crimes other than burglary. Warr (1988) argued that the common perception of rape as a crime that occurs in outside locations is false in a sizeable percentage of cases and instead can be an offence committed in a residential location, resembling a burglary. Hazelwood and Warren (1989) found that 50% of the cases of sexual assault they examined did occur in the victim’s home, supporting the assertion that sexual assaults occurring in an indoor location are not uncommon occurrences. Their study also showed that once an offender made the decision to offend in the victim’s residence for the first rape, that decision remained fairly consistent throughout the series of sexual assault. Factors that make a home attractive to burglars (e.g., ease of access and low visibility to the public) can also attract rapists. Warr (1988) asserted that a rape and burglary occurring at the same location is not an accident but instead a combined offence planned by an offender.

Sexual Burglary

Pedneault, Harris, and Knight (2012) highlighted a group of offences that combined rape and breaking and entering in a group of sex offender’s criminal histories. This is an important subtype of offending to study since it has been shown that sexual offenders who have committed burglaries have early onset criminal careers, more charges, and longer criminal careers (Harris, Pedneault, & Knight, 2013). A history of burglary is also more prevalent in samples of serial sex offenders (Hazelwood & Warren, 1989). However, a study by Park, Schlesinger, Pinizzotto, and Davis (2008) showed that there was no significant difference between serial and single rapist as to the stealing of property. As research on burglary suggests that it is a rational offence (Coupe & Blake, 2006; Nee & Taylor, 2000), Pedneault et al. (2012) believed that an examination of sexual burglary would shed light on the rationality of this type of sexual offence.

Explanations of sexual burglary usually fall into three categories (Pedneault, Beauregard, Harris, & Knight, 2015): (a) the burglary was the main goal of the offender and the sexual assault was just a “bonus” to the theft, when the offender unexpectedly found himself in a situation with a suitable victim, in a home he thought unoccupied (i.e., an opportunistic rape); (b) the sexual assault was the main goal of the offence and the theft was the bonus to the sexual offence (Scully & Marolla, 1985); and (c) the situational cues that make a location appealing for a burglary also make that same location appealing for a violent offence and the sexual assault is not accidental (Warr, 1988). Recent studies do not support the idea that burglary occurs as a bonus to a sexual assault since targets often feature characteristics that would be deemed counterindicative for a monetary-based burglary (Pedneault et al., 2015).

Pedneault et al. (2015), in their study of 224 incidents of sexual burglary, found that these incidents involved a rational organization and that there was limited support for the hypothesis that the sexual assaults were opportunistic crimes that were committed in addition to a regular burglary. For example, a number of situational cues indicated that the offenders planned for a sexual crime in addition to the break and enter, such as the finding that the sexual burglars would break into occupied homes and bring a weapon with them to the crime site. Victims were attacked when there was a lack of a capable guardian and when they were at their most vulnerable (i.e., asleep at their homes). Overall, these burglaries were dissimilar from regular burglaries and therefore it is unlikely that these offenders only planned for the break and enter and not a sexual crime as well. Still, support was provided for a rational organization of these sexual offences, notably found in the offenders’ modus operandi strategies, such as sexual burglaries being committed in apartments on lower floors, which would decrease the effort required to break into the home and later leave the scene. Such strategies make sense when understood as a cost minimization measure in the context of a sexual crime, while they may be questioned in the context of a monetary-based burglary (i.e., a burglar limiting himself to apartments, a smaller type of residence in which less affluent residents live and that contains fewer goods).

It appears that these sexual burglary offenders are purposely selecting the victim’s residence to commit their combined sexual assault and burglary. They are committing their offences at times when it is likely that the home will be occupied, which is a negative strategy for committing burglaries alone, but a beneficial strategy when you want a victim to be present. Beauregard, Proulx, Rossmo, Leclerc, and Allaire (2007) stated that their sample of home-intruder sexual offenders chose the “home-intruder” hunting style because of the advantages of private residences and it is likely that the sexual burglary offenders described by Pedneault et al. (2015) also benefitted by selecting the same type of private crime location. There are many similarities between sexual burglary offenders and home-intrusion and bedroom rape offenders because of the similar crime site selection, entailing a comparable set of crime costs and benefits.

The Bedroom Rapist

The previously discussed studies identified a specific type of offender that generally chose to encounter and attack a victim in her own home (i.e., Beauregard et al., 2007; Beauregard, Rebocho, & Rossmo, 2010; Pedneault et al., 2015; Warr, 1988). Two studies have focused exclusively on stranger bedroom rape crime events (Fossi, Clarke, & Lawrence, 2005; Lawrence, Fossi, & Clarke, 2010) and have confirmed the importance of including the context in the study of sexual assault. Victim behavior and victim characteristics (e.g., victim resistance during an assault) have often been studied to develop rape prevention strategies (Block & Skogan, 1986) but Fossi et al. (2005) believed that the sequence of a crime event would also be important to examine. They found that there are distinct groupings of sexual assault sequences that could be differentiated by the location of the offence. In one study, Fossi et al. (2005) chose to focus on one specific type of sexual assault location—bedroom rapes. Bedroom rapes were defined as sexual assault which occurred while the victim was in bed and the sequential analyses of these events came from witness statements of assaults which occurred in England. Results indicated that time and individual personality traits influenced the outcome of the assaults, as previous studies would have suggested, but the context of the offence also was an important factor. Overall, bedroom rapes which involved more than one sexual assault at the time of the crime involved an offender who acted out a consensual relationship script and did not use excessive force. The bedroom rapes, where only a single instance of sexual assault occurred, were characterized by an offender who used a more controlling verbal strategy.

Lawrence et al. (2010) also conducted a study on offenders’ verbal strategies during the course of stranger sexual assaults and separated the sexual assaults statements based on whether the assault could be defined as a bedroom rape or a nonbedroom rape. As with the previous studies focused on bedroom attacks there were differences between the sexual assaults that occurred in the bedroom versus those that did not, illustrating that this specific type of crime site location should be further examined. Specific differences noted in bedroom-based attacks were that a surprise attack by the offender on the victim was more likely than an offender who used a “con” approach (i.e., an offender who tried to trick the offender into the encounter). There was also more conversation and verbal intimacy found in the verbal strategies of the offender who committed a bedroom assault, but single bedroom rapes (compared with an offender who committed numerous rapes during one attack) showed conversations that were contained more aggression.

Aim of the Study

Recently, criminological research has turned toward the examination of sexual offender’s crime site selection and the influence that location can have on the crime event and offender. The aim of the current study is to analyze modus operandi, temporal factors, and victim characteristics in a sample of 347 stranger sexual assaults committed by 69 serial sex offenders to determine which factors may be more relevant to home-intrusion sexual assaults—that is, those offenders who commit the sexual assault in the victim’s residence—compared with sexual assaults committed at another type of location. Our hypothesis is that offenders who choose to sexually assault victims in their home constitute a specific type of sex offender, one that resembles the sexual burglar as discussed by Pedneault et al. (2015).

Method

Participants

The data were collected from a sample of sexual offenders who were incarcerated in a Correctional Service of Canada institution between 1995 and 2004. The sample consisted of sexual offenders who were convicted of a sentence of 2 or more years, and had committed two or more sexual assaults or other sex-related crimes (e.g., sexual homicide) involving a victim of any age and of any gender, who was a stranger to him (i.e., the offender and victim has no personal relationship prior to the offence). In total, 92 offenders matched the criteria and 69 of these individuals, whom together had committed a total of 347 sexual assaults (ranging from two to 37 sexual assaults each), were included in the study. Among the 23 excluded participants, only nine participants refused to participate, a remaining 11 participants were unavailable because of discipline problems, mental health state, or transfer to another institution and a further three were excluded after data collection due to missing data.

The age of the offenders at the time they began their crime series ranged from 18 to 55 years old, with a mean age of 29.3 (SD = 8.1). Among the participants, 92.5% had a prior criminal record before the onset of their series of sexual crimes and despite their involvement in a series of sexual crimes, the majority did participate in other crimes as well. Participants with a prior criminal record had an average of 1.7 (SD = 4.22) charges for sexual nonviolent crime, 4.5 (SD = 8.63) charges for sexual violent crimes, 14.4 (15.79) charges for nonsexual nonviolent crimes, and 3.5 (SD = 5.86) charges for nonsexual violent crimes. The victims ages ranged from 4 to 68 years old with a mean age of 18.4 (SD = 9.6). The mean age of female victims was 20 years old (SD = 9.3) compared with 12.1 years of age for male victims (SD = 7.9). The majority of victims were female (79.8%). Thirty offenders (43.5%) had adult only victims, 17 offenders (24.6%) had child only victims, and 22 offenders (31.9%) had both adult and child victims.

Procedure

The procedure involved two stages: (a) in-depth semistructured interviews with all offenders to investigate each sexual crime, and (b) content analysis of police investigation reports. An instrument was developed from existing police questionnaires (e.g., Violent Crime Linkage Analysis System [ViCLAS], Violent Criminal Apprehension Program [ViCAP]) to guide the collection of information from the police investigation reports and the in-depth semistructured interviews with the offenders. The questionnaire included sections that collected information on precrime factors, target selection processes, modus operandi, postcrime factors, and geographic factors. Self-reported information was then compared with official data (i.e., police reports). In the case of a discrepancy on factual information (e.g., location of the crime), information from the official police data were used. Participants 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. Interviews were conducted in a private office, isolated from correctional staff and other inmates.

Measures

Dependent variables

The current study used three dependent variables to measure the environmental characteristics of three different crime locations. The dependent variables were crime site location, victim encounter location, and victim release location. These locations are important sites for examining environmental and geographical aspects of a serial rape event (Rossmo, 2000). The dependent variables were dichotomized to isolate the category of interest: whether or not each location site occurred at the victim’s residence (0 = not victim’s residence, 1 = victim’s residence). Consistency in offenders’ choice of environmental characteristics related to their crime site location has been found in previous studies (e.g., Lundrigan, Czarnomski, & Wilson, 2010). Despite the relative stability of the locations associated with sexual crimes, it was decided to look at each specific location. In total, 21.6% (n = 75) of the victim encounter sites were located at the victim’s residence, 21.3% (n = 74) of the crime sites were located at the victim’s residence, and 23.3% (n = 81) of the victim release sites were located at the victim’s residence.

Independent variables: Modus operandi strategies and planning variables

Five modus operandi strategies were included in the current study. The five variables were divided into two categories: planning variables and crime event variables. The two variables which made up the planning variables category were type of victim selection and crime premeditation. Type of victim selection was dichotomized into random and nonrandom selection (0 = nonrandom; 1 = random). Nonrandom selection refers to the offender not choosing a victim at random, but rather purposely selecting a victim independent from the situation. Random victim selection refers to the offender choosing the victim at random as a result of opportunity to do so. The second variable, crime premeditation (0 = no premeditation or unstructured premeditation; 1 = structured premeditation) refers to a thought process which manifests itself by preparation and planning. A sexual crime is premeditated when it is planned by the offender prior to its commission. The premeditation is structured when its level of planning is elaborate and involves specific components such as the victim’s identity, specific victim characteristics, locations at which the crime will be committed, strategies to commit the crime, and so on (Beauregard et al., 2010; Hewitt, Beauregard, & Davies, 2012). Three variables were included within the crime event modus operandi category. These variables included method to commit the crime (0 = noncoercive; 1 = coercive), the level of physical force used by the offender to secure the victim’s compliance and commit the crime (0 = no or minimum force; 1 = excessive force/more force than necessary), and whether the offender committed burglary in addition to the sexual assault (0 = no; 1 = yes).

Independent variables: Temporal variables

Three temporal variables were included in the current study. The first variable was offence timing (0 = weekday, 1 = weekend) and the second variable was time of the crime (0 = day, from 6:00 a.m. to 5:59 p.m.; 1 = night, from 6:00 p.m. to 5:59 a.m.). These temporal variables are relevant for determining the environmental characteristics of the offence at the victim encounter, crime site, and victim release sites. The third temporal variable was the amount of time the offender spent with the victim at the scene (0 = 30 min or less; 1 = more than 30 min). Based on a past study by Oziel, Goodwill, and Beauregard (2015), a short period of time was categorized as 30 min or less and a high period of time was more than 30 min spent with the offender. It is generally thought that committing a rape at the victim’s home is beneficial because of the lack of witnesses and thus one could hypothesize that offenders may be willing to spend a longer amount of time with the victim at the scene of the crime when in a private location that minimizes the risk of witnesses.

Control variables

There were three control variables included in the current study, the first variable was the age of the victim which was dichotomized into a victim who was a youth or adult victim (0 = age of victim is 0-17 years old; 1 = 18 years old and older).¹ The second variable was the sex of the victim (0 = male; 1 = female). Finally, the third control variable was the relationship between the victim and the offender (0 = stranger; 1 = offender and victim have seen/talked to each other). The current data consisted of all stranger sexual assaults with the offender and victim having no previous relationship prior to the offence committed; however, it is possible that the victim or the offender may have seen or talked to each other briefly before the crime took place, without the interaction being considered a previous relationship. Some past research has found that sexual assaults with nonstrangers were more likely to occur at the victim’s homes (Warr, 1988) so some prior contact between the offender and victim before the offence may also influence the location of the offence (see Table 1).

Table 1.

Descriptive Statistics.

Variable	n (%)
Victim characteristics
Age of victim
Youth (0-17)	197 (56.8)
Adult (18+)	150 (43.2)
Sex of victim
Male	70 (20.2)
Female	277 (79.8)
Victim-offender relationship
Nonstranger	226 (65.1)
Stranger	121 (34.9)
Modus operandi strategies offence planning
Crime premeditation
Nonstructured premeditation	41 (11.8)
Structured premeditation	306 (88.2)
Type of victim selection
Nonrandom	171 (49.3)
Random	176 (50.7)
Crime event variables
Method to commit crime
Noncoercive	144 (41.5)
Coercive	203 (58.5)
Level of force
No force or minimum force	291 (83.9)
Excessive force	56 (16.1)
Offender committed burglary
No	310 (89.3)
Yes	37 (10.7)
Temporal variables
Offence timing
Weekday	281 (81)
Weekend	66 (19)
Time of crime
Day	206 (59.4)
Night	141 (40.6)
Time at crime scene with victim
Less than 30 min	255 (73.5)
30+ min	88 (25.8)
Dependent variable
Victim encounter location
Not victim’s residence	272 (78.4)
Victim’s residence	75 (21.6)
Crime site location
Not victim’s residence	273 (78.7)
Victim’s residence	74 (21.3)
Victim release location
Not victim’s residence	266 (76.7)
Victim’s residence	81 (23.2)

Analytical Strategy

A two-step analytical process was used in the current study. First, bivariate analyses were conducted to determine which variables may be the most significant predictors to use in later multivariate analyses. Examination of the bivariate findings showed significant relationships between the dependent variables and a number of predictors. Second, multiple regression analyses were used applying generalized estimating equations (GEEs) performed in SPSS.

The current data consists of a series of sexual crimes committed by serial sexual offenders. Each offender committed a minimum of two sexual offences and therefore it is possible that correlations exist between crime events committed by the same offender. The possibility of correlation means it cannot be assumed that independence exists between each observation and, thus, the analytic strategy chosen needs to take this into account. Treating observations as though they are independent, which is one assumption of regression, could potentially have negative consequences such as underestimating standard errors or models having inefficient estimators, meaning that there is greater mean square error (Johnston & Stokes, 1997). GEEs were chosen as the appropriate method because it is an extension of generalized linear models that allows for data that includes correlated responses (Garson, 2013; Liang & Zeger, 1986).

The repeated offences committed by the same offender are incorporated in the GEE model based on a working correlation matrix. The correct correlation structure is often unknown to the researcher and must be estimated. A number of options relating to correlation structure exist, the most common being independent (all cases are assumed to be independent of one another), exchangeable (assumes the correlation between observations is the same across all cases), and unstructured (no assumption is made about the correlations between observations and no pattern is assumed); however, GEEs are generally robust to misspecification of the working correlation matrix and is thus an attractive strategy for correlated data (Letourneau, Levenson, Bandyopadhyay, Armstrong, & Sinha, 2010). In the current study, working correlation matrices were analyzed for victim encounter location, crime site location and victim release locations and the independent structure was found to be the appropriate structure based on comparison of the quasilikelihood criterion (QIC); the structure with the lowest QIC value being chosen. The independent structure assumes that all cases within the data set are independent of one another (Liang & Zeger, 1986).

As QIC values can be used to determine which correlation structure offers the best model fit, the QIC values can be used to choose the best subset of predictors to include in GEE models (Garson, 2013). Initially all significant variables identified from the bivariate findings were included in the model but examination of the QIC values suggested that the removal of the offender weapon use and forensic awareness variables created a better model fit. These two variables were not found to be significant during GEE analyses either so the researcher chose to remove them from the final models.²

Results

Table 2 shows the bivariate analysis results for the victim encounter site. There is a significant association between encounter locations occurring at a victim’s home and eight of the variables included. Victim residence encounter sites were more likely when the victims were adults and when the offence occurred at night compared with the day. Further, offenders who choose victim residence encounter locations were more likely to have premeditated and selected their victims ahead of time and were more likely to use coercion and a weapon. Finally, home-intruder sex offenders were more likely to commit a burglary in addition to the sexual assault and show signs of being forensically aware.

Table 2.

Bivariate Associations Between Predictors and Victim Encounter Location (N = 347).

	Victim residence % (n)	Nonresidence % (n)	χ²
Age of victim (18+)	69.3 (52)	36.0 (98)	.277***
Victim-offender relationship (stranger)	68.0 (51)	64.3 (175)	−.032
Sex of victim (female)	88.0 (9)	77.6 (211)	.107
Crime premeditation (structured)	94.7 (71)	86.4 (235)	.155*
Nonrandom victim selection	64.0 (48)	45.2 (123)	−.176**
Coercive method to commit crime	81.3 (61)	52.2 (142)	.243***
Level of force (excessive)	20.0 (15)	15.1 (41)	.055
Offender committed burglary	32.0 (24)	4.8 (13)	.363***
Offender brought rape kit	10.7 (8)	8.8 (24)	.026
Offender forensically aware	64.0 (27)	34.2 (93)	.250***
Weapon use	42.7 (32)	24.6 (67)	.164**
Offence timing (weekday)	77.3 (58)	82.0 (223)	.049
Time of crime (night)	61.3 (46)	34.9 (95)	.221***
Time at crime scene with victim (30+ min)	18.7 (14)	27.6 (74)	−.085

p < .05. **p < .01. ***p < .001.

To identify the differences between offenders who offend at the victim’s residence versus other crime site locations, bivariate relationships between the crime site and crime offence variables were also examined. Table 3 shows that there is a significant association between home-intrusion attack crime sites and nine of the variables included in the current study. Victim residence crime sites were more likely when the victims were adults and when the offence occurred at night, compared with the day. Furthermore, offenders who choose the victim residence as crime site locations were more likely to have premeditated and selected their victims ahead of time and were more likely to use coercion, excessive force, and a weapon. Finally, home-intruder sex offenders were more likely to commit a burglary during the sexual assault and show signs of being forensically aware.³

Table 3.

Bivariate Associations Between Predictors and Crime Site Location (N = 347).

	Victim residence % (n)	Nonresidence % (n)	χ²
Age of victim (18+)	75.7 (56)	34.4 (94)	.341***
Victim-offender relationship (stranger)	73.0 (54)	63.0 (20)	−.086
Sex of victim (female)	91.9 (68)	76.6 (209)	.157
Crime premeditation (structured)	95.9 (71)	86.1 (235)	.125*
Nonrandom victim selection	66.2 (49)	44.7 (122)	−.176**
Coercive method to commit crime	85.1 (63)	51.3 (140)	.281***
Level of force (excessive)	24.3 (18)	13.9 (38)	.116*
Offender committed burglary	36.5 (27)	3.7 (10)	.436***
Offender brought rape kit	14.9 (11)	7.7 (21)	.102
Offender forensically aware	67.6 (50)	33.3 (91)	.286***
Weapon use	47.3 (35)	23.4 (64)	.216***
Offence timing (weekday)	86.5 (64)	79.5 (217)	−.073
Time of crime (night)	67.6 (50)	33.3 (91)	.286***
Time at crime scene with victim (30+ min)	23.6 (17)	26.2 (71)	−.024

p < .05. **p < .01. ***p < .001.

Victim release location was dichotomized into whether the victim was released in their own home or at another location. Bivariate analyses show that eight of the variables are significant, the same results as was found for the victim encounter site. Again, victims were more likely to be released in their residence when the victims were adults and when the offence occurred at night. Offenders were more likely to have premeditated and selected their victims ahead of time and use coercion and a weapon. Victim’s residence release sites were more likely to involve offenders who would also commit a burglary during the sexual assault and show signs of being forensically aware.

Next, the independent variables were tested at the multivariate level using GEE analyses for the three locations.⁴ As to victim encounter crime site, sexual offences that involved adult victims (OR = 4.626; p < .05) were again more likely to have an encounter location at the victim’s home than at another type of location. The relationship between victim and offender is also a significant predictor of where the offence will take place. Sexual crimes with victims and offenders who had some previous contact with each other prior to the offence (OR = 5.085; p < .01) are more likely when the victim is encountered at their home, compared with when the victim and offender had no contact with each other previously. Time offender spent at the scene with the victim proved to be a significant predictor of whether the encounter location was the victim’s residence. The offender was more likely to spend a shorter amount of time at the scene (i.e., 30 min or less) when the encounter began at the victim’s home (OR = .180; p < .01). In regard to modus operandi strategies, two predictors were significant. The method to commit the crime was significant with a noncoercive method style (OR = 3.672; p < .05) more likely to be used by an offender when the crime encounter location was not at the victim’s home. Second, premeditation was significant (OR = 4.456; p = .05) with a structured premeditation more likely to be used by an offender when the crime encounter location was the victims home. Third, crime events where the offender committed burglary during the sexual assault crime event (OR = 8.256; p < .01) were more likely to be sexual crimes that began at the victim’s residence compared with a different location.

When examining the crime site location, two of the control variables were significant predictors of whether or not the crime site location occurred inside the victim’s residence. Sexual offences that involved adult victims (OR = 5.798; p < .05) were more likely to take place at the victim’s home than with victims under 18 years of age. The relationship between victim and offender was also a significant predictor of where the offence will take place. Sexual crimes with victims and offenders who had some previous contact with each other prior to the offence (OR = 3.966; p < .05) are more likely to occur at a location that is the victim’s home than a different location site. Only one temporal factor—time offender spent at the scene with the victim—proved to be a significant predictor of whether the crime site occurred at the victim’s residence. The offender was more likely to spend a shorter amount of time at the scene (i.e., 30 min or less) when the attack occurred at the victim’s home (OR = .289; p < .05). In regard to modus operandi strategies, three predictors were significant. The method to commit the crime was significant with a noncoercive method style (OR = 4.375; p < .05) more likely to be used by an offender when the crime scene location was not at the victim’s home. Premeditation (OR = 7.145; p < .05) was also a significant factor, with home-intrusion sexual assault crime site locations being more likely to occur with offenders who used structured premeditation. Finally, crime events where the offender committed burglary prior to or after the sexual assault (OR = 10.246; p < .01) were more likely to occur at a victim’s residence (see Table 4).

Table 4.

Bivariate Associations Between Predictors and Victim Release Location (N = 347).

	Victim residence % (n)	Nonresidence % (n)	χ²
Age of victim (18+)	66.7 (27)	36.1 (96)	.261***
Victim-offender relationship (stranger)	69.1 (56)	63.9 (170)	−.046
Sex of victim (female)	87.7 (71)	77.4 (206)	.108
Crime premeditation (structured)	95.1 (77)	86.1 (229)	.118*
Nonrandom victim selection	65.4 (53)	44.4 (118)	−.178**
Coercive method to commit crime	79.0 (64)	52.3 (139)	.230***
Level of force (excessive)	22.2 (18)	14.3 (38)	.091
Offender committed burglary	30.9 (25)	4.5 (12)	.361***
Offender brought rape kit	13.6 (11)	7.9 (21)	.083
Offender forensically aware	63.0 (51)	33.8 (90)	.251***
Weapon use	40.7 (33)	24.8 (66)	.149**
Offence timing (weekday)	81.5 (66)	80.8 (215)	−.007
Time of crime (night)	61.7 (50)	34.2 (91)	.237***
Time at crime scene with victim (30+ min)	21.8 (17)	26.8 (71)	−.048

p < .05. **p < .01. ***p < .001.

Finally, multivariate analyses show that sexual offences that involved adult victims (OR = 3.398; p < .05) were more likely to have a victim release location at the victim’s home than at another type of encounter location (see Table 5). The relationship between victim and offenders was also a significant predictor of where the release location would be. Sexual crimes with victims and offenders who have had some previous contact with each other prior to the offence (OR = 2.844; p < .05) are more likely to have release sites at the victim’s home than when the victim and offender have had no contact with each other previously. As with crime site and encounter location only one temporal factor—time offender spent at the scene with the victim—proved to be a significant predictor of whether the encounter location was the victim’s residence. The offender was more likely to spend a shorter amount of time at the scene (i.e., 30 min or less) when the victim was released at the victim’s home (OR = .321; p = .05). Regarding modus operandi strategies, the method to commit the crime was no longer a significant predictor of victim release site. However, whether the offender committed burglary in addition to the sexual assault (OR = 7.097; p < .01) was still significant, with offenders who committed burglary being more likely to release their victims at the victim’s residence compared with a different location. An offender using structured premeditation was also more likely when the victim release site was at the victim’s bedrooms and was statistically significant (OR = 3.398; p < .05).

Table 5.

Generalized Estimating Equation Model Predicting Victim Encounter, Crime, and Victim Release Locations.

	Victim encounter locationOR (SE)	Crime locationOR (SE)	Victim release locationOR (SE)
Age of victim (18+)	4.626 (.714)*	5.798 (.712)*	3.398 (.610)*
Victim-offender relationship (seen + talked to)	5.085 (.590)**	3.966 (.596)*	2.844 (.519)*
Crime premeditation (structured)	4.456 (.762)†	7.145 (.975)*	4.431 (.7445)*
Random victim selection	1.240 (.800)	1.126 (.827)	.995 (.743)
Coercive method to commit crime	3.672 (.601)*	4.375 (.654)*	2.962 (.571)
Level of force (excessive)	.809 (.840)	1.236 (.830)	1.148 (.760)
Offender committed burglary	8.256 (.765)**	10.246 (.743)**	7.097 (.783)**
Offence timing (weekend)	2.507 (.659)	1.021 (.623)	1.697 (.598)
Time of crime (night)	1.920 (.844)	1.991 (.862)	1.734 (.786)
Time at crime scene with victim (30+ min)	0.180 (.633)**	0.298 (.548)*	0.321 (.583)†

Note. The specified working correlation structure is independent. The reference category for each independent variable is provided in parentheses. OR = odds ratio.

†

p = .5. *p < .05. **p < .01.

Discussion

The current study examined stranger “home-intrusion” sexual assaults, referring to sexual assaults which occurred at a victim’s residence, along with temporal factors and offender modus operandi strategies to gain a better understanding of these types of sexual assaults. The victim characteristics, along with temporal and modus operandi factors that were significant for home-intrusion sexual assaults were generally significant across the three crime locations which can be explained by the high consistency in location during home-intrusion sexual assaults, except for method of crime commission which was not a significant variable for the victim release site.

The first significant victim characteristic was age, the results indicating that adults were more likely to be targets of home-intrusion sexual assaults compared with younger victims. This finding is consistent with a routine activities approach. Generally, both adults and children spend a large portion of their time at home, which makes their home an opportune place to encounter a victim. Bernasco and Nieuwbeerta (2005) noted that offenders are likely to choose an area to commit their crimes if they deem that location to be a suitable place to find appropriate targets. A target is also deemed more suitable by an offender if it lacks capable guardianship (Cohen & Felson, 1979). Adults are more likely to be home alone, without a capable guardian, compared with children who are supervised by a parent or caregiver at most times. In comparison, a female who lives alone may be deemed a more suitable target by a sexual offender because there is little chance of anyone interrupting the attack. This difference in guardianship between adults and children explains why adults were more likely to be encountered and attacked at their residence in the present study. This lack of guardianship and potential witnesses would also explain why the victim’s home is also used as an encounter and release site.

The second significant victim characteristic found across encounter site, crime site, and victim release site was that the offender and victim were more likely to have had some contact with each other before the crime occurred, when the offence was a home-intrusion sexual assault. Although it has been established previously that victims of nonstranger rapes are often victimized in their own homes (McDermott, 1979; Warr, 1988), the current findings indicate that even stranger rapes do not necessarily occur in public locations and that the crime location can be related to the type of relationship (however superficial) the victim and offender have. From a rational choice perspective, the selection of a residence as the location of a sexual assault makes intuitive sense: The settings are private and reduce the possibility of a third-party intervention. It is therefore unsurprising that both stranger and nonstranger offenders opt to commit rape in residences, although gaining access might be complicated when the victim does not know the offender. It is possible that the limited contact between offender and victim noted in the current findings was enough to facilitate entry, similar to the results of Pedneault et al. (2015), which found that about 20% of all sexual burglaries involve some trickery and short prior contact by the offender with the victim to gain access to the victim’s living space.

Although the victims of home-intrusion sexual assaults were more likely to have had some contact with the offender prior to the offence, it was not found that victims were selected beforehand. Even though victim selection was not found to be a significant factor, results do indicate that home-intruder sex offenders use some form of structured premeditation in planning these offences. The victim was more likely to be encountered and released at their homes if the crime is premeditated by the offender. The location of the crime was also more likely to be at the victim’s home when the crime was structurally premeditated. These findings would suggest that the offender is purposely hunting for victims in this type of location because they know they can find potential victims there. Hewitt et al. (2012) found that victims of sexual assault were more likely to be encountered in private or indoor locations when the crime was structurally premeditated.

In terms of the modus operandi variables, victims are more likely to be encountered and assaulted in their residence when the offender utilizes a coercive method to commit the crime. Sexual coercion involves the use of threats, force, or another oppressive strategy to gain victim compliance (Balemba & Beauregard, 2012). This finding was expected as Beauregard et al. (2007) found similar results in their examination of the home-intrusion rape track. The home-intruder committed their sexual assaults at an indoor or private location and used coercion and force to commit their offence. The current examination of home-intrusion sexual assaults suggests that these offenders are more likely to use coercion in their encounters with victims and during the commission of their attack; however, use of force (operationalised as the use of more than necessary force) was not found to be a significant modus operandi variable in the current study. From a rational choice perspective, it is possible that offenders needed coercion to gain control of the situation but did not need to escalate force and exert more effort, since victim was at the desired crime site with enough privacy to limit the possibility of interruption.

A victim was more likely to be encountered, assaulted and released in their home when the offender spent a shorter amount of time at the crime scene. A “short” amount of time was operationalised as 30 min or less (based on past research by Oziel et al., 2015). Home-intruder sex offenders were more likely to spend 30 min or less with their victims compared with others. Committing a sexual assault at the victim’s home can be beneficial because of the reduced risk of potential witnesses since the crime site is removed from the public view. Because of this, it was hypothesized that an offender would spend a longer amount of time with the victim at the crime scene; however, results indicate the opposite. Although the rational choice approach generally supports that an offender should limit the amount of time spent at the crime site to reduce the risk of apprehension, it is unclear why this would be the case more so than in nonresidential rapes. Perhaps the unfamiliarity and lack of predictability of the location sways the offender toward staying a shorter amount of time even though the location is hidden from public.

Finally, home-intrusion sexual assaults encounter sites, crime sites, and release sites were also more likely to involve offenders who also stole from the victim prior to or after the sexual assault. Pedneault et al. (2015) did not find support for the explanation that sexual burglaries were opportunistic, “bonuses” to residential thefts, in which a female house occupant was encountered in the course of burglarizing a residence for monetary purposes. The reasoning for this was because sexual burglaries appeared to be significantly different from nonsexual burglaries (e.g., sexual burglaries were more likely to occur in apartments containing fewer goods and in occupied residences at times when it was more likely for a victim to be home). Arguably, the victim’s residence generates criminal opportunities for the offender to commit both a sexual assault and a burglary offence and thus these locations would also be a beneficial choice for offenders who want to commit both types of offences. It can be hypothesized that offenders are choosing the victims home as a suitable location because of a rational decision-making process determining that these locations provide the most benefits and opportunity for the commission of both types of offences.

Conclusion

The current study aimed to add to the existing knowledge of home-intrusion stranger sexual assaults and the offending behavior used at these locations. It was found that modus operandi, temporal factors, and victim characteristics were significant in predicting whether the victim encounter, crime site, and victim release site were located at the victim’s residence or not.⁵ Even though a victim’s home as a crime location comes with certain risks (e.g., being a space that is unfamiliar to the offender or the increased difficulty in knowing whether more than one individual is home), these home-intruder sex offenders must view the location as being beneficial in a way that outweighs the risks. Some major benefits of the location would be the lack of visibility for potential witnesses, no need to bring the victim to the crime site, and a high likelihood that a victim would be available at the location. It appears that offenders are aware of the potential risks and rewards of this location and change their offending behavior and choices accordingly. As hypothesized, it seems that offenders who target victims in their own home fit the profile of a specific type of sex offender, one that has been described in previous studies as a “sexual burglar” (see Pedneault et al., 2015).

The current study is unique in that it focuses on serial rape crime events and not specifically on the offender. This allows for situation crime prevention measures to be put forth. There are a number of crime prevention measures that could be implemented to potentially reduce home-intrusion sexual assaults. On the outside of a potential victim’s home, it is recommended that doors and windows be kept locked, camera and alarm systems be placed at entryways, and strict doorway access into apartment buildings be kept, for example, by having all individuals scan access cards or key fobs to gain access to the building. Alarms and cameras will act as “guardians” for individuals that live alone which may decrease the likelihood of an offender choosing to offend at the location. Keeping blinds closed at night so that offenders are not able to hunt for potential victims from outside the home and hinder their ability to see objects of value would also be highly recommended. Inside the home, panic buttons or easy access to emergency calling on cell phones could help in situations where an offender may have already gained entry. Other measures could be introduced in outside areas surrounding residences to make it more difficult for potential offenders to break into a home. For instance, keeping residences well-lit and having cameras outside apartment or condo buildings. Finally, increased police patrols or neighbourhood watch programs in residential areas could impede potential offenders. Overall, it is important that we be aware of environmental cues that may make potential victims more susceptible to being victimized. Increasing the risks for offenders to target and attack victims and their homes are possible when these environmental structures are modified.

This research is not without its limitations. The selection of serial sex offenders for the purpose of this study prevents us to generalize our findings to the entire population of sex offenders. Specifically, it is possible that the crime location decisions are very different (e.g., more impulsive) in cases of nonserial sex offenders compared with serial sex offenders. Despite this obvious limitation, such decision provided the study with a couple of strengths. Serial sex offenders were specifically targeted for the sample, as they are more likely to face a variety of situations and, accordingly, are more likely to make a variety of choices during the decision-making process of each crime committed. Moreover, the stranger-victim element provides a further refinement to preserve a relative homogeneity in the situation faced by offenders; prior relationship between the offender and the victim arguably heavily influences the locations where the crimes are likely to take place. Also, the small sample size restricted the kinds of statistical analyses able to be performed and due to the distribution of the sample, some variables with theoretical relevance could not be included in the study. It is unclear how the exclusion of these variables may have affected the results. In addition, crimes analyzed in this study were all committed by incarcerated offenders only; therefore, it is possible that the results may not be generalizable to offenders who have not been apprehended. Finally, misreporting of modus operandi strategies and other crime characteristics by offenders, whether intentionally or due to misconception of memory, is possible with self-reported events of crimes, although the original interviewer was able to look to official reports if obvious discrepancies in self-reporting occurred. Despite these limitations, the current findings are theoretically and practically relevant for researchers and potentially law enforcement officials as well.

Future research should attempt to examine whether other situational, modus operandi, or victim characteristics may be more likely in home-intrusion sexual assaults compared with other locations. For instance, the type of victim residence (apartment, house, etc.) or the number of residents in the home. A qualitative analysis of offender’s reasoning for why they choose particular crime site locations would also provide a greater understanding of their decision-making process.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Balemba

Beauregard

(2012). Reactions to resistance: The role of contextual factors in sex offending. Violence and Victims, 27, 148-165.

Beauregard

Bouchard

(2010). Cleaning up your act: Forensic awareness as a detection avoidance strategy. Journal of Criminal Justice, 38, 1160-1166.

Beauregard

Proulx

Rossmo

D. K.

Leclerc

Allaire

(2007). Script analyses of the hunting process of serial sex offenders. Criminal Justice and Behavior, 34, 1069-1084.

Beauregard

Rebocho

Rossmo

(2010). Target selection patterns in rape. Journal of Investigative Psychology and Offender Profiling, 7, 137-152.

Bernasco

(2006). Co-offending and the choice of target areas in burglary. Journal of Investigative Psychology and Offender Profiling, 3, 139-155.

Bernasco

Nieuwbeerta

(2005). How do residential burglars select target areas? A new approach to the analysis of criminal location choice. British Journal of Criminology, 44, 296-315.

Block

Skogan

W. C.

(1986). Resistance and nonfatal outcomes in stranger-to-stranger predatory crime. Violence and Victims, 4, 241-253.

Clarke

R. V.

Cornish

D. B.

(1986). Modeling offenders’ decisions: A framework for research and policy. Crime and Justice, 6, 147-185.

Cohen

L. E.

Felson

(1979). Social change and crime rate trends: A routine activity approach. American Sociological Review, 44, 588-608.

10.

Coupe

Blake

(2006). Daylight and darkness targeting strategies and the risk of being seen at residential burglaries. Criminology, 44, 431-464.

11.

Deslauriers-Varin

Beauregard

(2010). Victims’ routine activities and sex offenders’ target selection scripts: A latent class analysis. Sexual Abuse: A Journal of Research and Treatment, 22, 315-342.

12.

Dufresne

(2017, June 08). Sex offender found guilty of assault, attempted B&E of elderly woman’s home. CBC News. Retrieved from http://www.cbc.ca/news/canada/british-columbia/sex-offender-found-guilty-of-assault-attempted-b-e-of-elderly-woman-s-home-1.4150855

13.

Fossi

J. J.

Clarke

D. D.

Lawrence

(2005). Bedroom rape: Sequences of sexual behavior in stranger assaults. Journal of Interpersonal Violence, 20, 1444-1466.

14.

Francis

Soothill

Dittrich

(2001). A new approach for ranking “serious” offences: The use of paired-comparisons methodology. British Journal of Criminology, 41, 726-737.

15.

Garson

G. D.

(2013). Generalized linear models. Available from http://statisticalassociates.com/

16.

Harris

D. A.

Pedneault

Knight

R. A.

(2013). An exploration of burglary in the criminal histories of sex offenders referred for civil commitment. Psychology, Crime & Law, 19, 765-781.

17.

Harris

D. A.

Smallbone

Dennison

Knight

R. A.

(2009). Specialization and versatility in sexual offenders referred for civil commitment. Journal of Criminal Justice, 37, 37-44.

18.

Hazelwood

R. R.

Warren

J. I.

(1989). The serial rapist: His characteristics and victims. FBI Law Enforcement Bulletin, 58, 10-18.

19.

Hewitt

Beauregard

Davies

(2012). “Catch and release” Predicting encounter and victim release location choice in serial rape events. Policing: An International Journal of Police Strategies & Management, 35, 835-856.

20.

Johnston

Stokes

(1997). Applications of GEE methodology using the SAS system. In Proceedings of the 10th Annual NorthEast SAS Users Group (pp. 881-886). Cary, North Carolina: SAS Institute Inc.

21.

Lawrence

Fossi

Clarke

(2010). A sequential examination of offenders’ verbal strategies during stranger rapes: The influence of location. Psychology, Crime & Law, 16, 381-400.

22.

Letourneau

E. J.

Levenson

J. S.

Bandyopadhyay

Armstrong

K. S.

Sinha

(2010). The effects of sex offender registration and notification on judicial decisions. Criminal Justice Review, 35, 295-317.

23.

Liang

K. Y.

Zeger

S. L.

(1986). Longitudinal data analysis using generalized linear models. Biometrika, 73, 13-22.

24.

Lundrigan

Czarnomski

Wilson

(2010). Spatial and environmental consistency in serial sexual assault. Journal of Investigative Psychology and Offender Profiling, 7, 15-30.

25.

McDermott

M. J.

(1979). Rape victimization in 26 American cities. Washington, DC: U.S. Government Printing Office.

26.

Nee

Taylor

(2000). Examining burglars’ targets selection: Interview, experiment or ethnomethodology? Psychology, Crime & Law, 6, 45-59.

27.

Oziel

Goodwill

Beauregard

(2015). Variability in behavioral consistency across temporal phases in stranger sexual offences. Journal of Police and Criminal Psychology, 30, 176-190.

28.

Park

Schlesinger

L. B.

Pinizzotto

A. J.

Davis

E. F.

(2008). Serial and single-victim rapists: Differences in crime-scene violence, interpersonal involvement, and criminal sophistication. Behavioral Sciences & the Law, 26, 227-237.

29.

Pedneault

Beauregard

Harris

D. A.

Knight

R. A.

(2015). Rationally irrational: The case of sexual burglary. Sexual Abuse: A Journal of Research and Treatment, 27, 376-397.

30.

Pedneault

Harris

D. A.

Knight

R. A.

(2012). Toward a typology of sexual burglary: Latent class analysis. Journal of Criminal Justice, 40, 278-284.

31.

QMI Agency. (2012, September 18). Sex attacker breaks into four homes overnight in Etobicoke. Toronto Sun. Retrieved from http://www.torontosun.com/2012/09/18/sex-attacker-breaks-into-four-homes-overnight-in-etobicoke

32.

Rossmo

D. K.

(2000). Geographic profiling. Boca Raton, FL: CRC Press.

33.

Scully

Marolla

(1985). Riding the bull at Gilley’s: Convicted rapists describe the rewards of rape. Social Problems, 32, 251-263.

34.

Taylor

R. B.

Gottfredson

(1986). Environmental design, crime, and prevention: An examination of community dynamics. Crime and Justice, 8, 387-416.

35.

Vancouver police seek high-risk offender Antoine Naskathey. (2017, February 10). CBC News. Retrieved from http://www.cbc.ca/news/canada/british-columbia/vancouver-police-seek-high-risk-offender-antoine-naskathey-1.3977685

36.

Warr

(1988). Rape, burglary, and opportunity. Journal of Quantitative Criminology, 4, 275-288.