ArticlePDF Available

Choice of Weapon or Weapon of Choice? Examining the Interactions between Victim Characteristics in Single-victim Male Sexual Homicide Offenders

Authors:

Abstract and Figures

As most studies report that the majority of sexual homicide offenders (SHOs) prefer to kill with their own hands, research has largely neglected to examine the choice of weapon by these offenders. The US Supplementary Homicide Reports show that although a large number of SHOs murder their victim using personal weapons (e.g. bare hands and manual or ligature strangulation), the majority use an alternative weapon (e.g. edged weapons, contact weapons, and firearms). The present study hypothesises that the choice of weapon is in part influenced by victim characteristics. To identify specific combinations and interactions between victim characteristics and the choice of a personal or edged weapon during the commission of a sexual homicide, a combination of exhaustive chi-square automatic interaction detector and conjunctive analysis is used on a sample of 2,472 single-victim male SHOs from a 36-year period of Supplementary Homicide Report data (1976–2011). Findings show that SHOs choose their weapon according to some victim characteristics. Implications of the findings are discussed in light of police suspect prioritisation. Copyright © 2014 John Wiley & Sons, Ltd.
Content may be subject to copyright.
This is an accepted manuscript version of an article published in the journal:
Journal
of Investigative Psychology and Offender Profiling
Copyright to the final published article belongs to 2016 John Wiley & Sons, Ltd.
Chan, H. C. O., & Beauregard, E. (2016). Choice of Weapon or Weapon of Choice?
Examining the Interactions between Victim Characteristics in Single‐victim
Male Sexual Homicide Offenders.
Journal of Investigative Psychology and
Offender Profiling, 13
, 70-88. doi: 10.1002/jip.1432.
C h a n e t a l .
P a g e | 2
Choice of Weapon or Weapon of Choice? Examining the
Interactions between Victim Characteristics in Single-
victim
Male Sexual Homicide Offenders†
HENG CHOON (OLIVER) CHAN1 and ERIC BEAUREGARD2
1 Department of Applied Social Sciences, City University of Hong Kong, Kowloon,
Hong Kong.
2 Center for Research on Sexual Violence, School of Criminology, Simon Fraser
University, Burnaby, British Columbia, Canada
†An earlier version of this paper was presented at the 15th International Academy of
Investigative Psychology Conference’s symposium on ‘Sexual Homicide and
SexualMurderers: Empirical Findings for Investigative Psychology’, 9 April 2014 in
London.
Abstract
As most studies report that the majority of sexual homicide offenders (SHOs) prefer
to kill with their own hands, research has largely neglected to examine the choice of
weapon by these offenders. The US Supplementary Homicide Reports show that
although a large number of SHOs murder their victim using personal weapons (e.g.
bare hands and manual or ligature strangulation), the majority use an alternative
weapon (e.g. edged weapons, contact weapons, and firearms). The present study
hypothesises that the choice of weapon is in part influenced by victim characteristics.
To identify specific combinations and interactions between victim characteristics and
the choice of a personal or edged weapon during the commission of a sexual
homicide, a combination of exhaustive chi-square automatic interaction detector and
conjunctive analysis is used on a sample of 2,472 single-victim male SHOs from a 36-
year period of Supplementary Homicide Report data (19762011). Findings show that
SHOs choose their weapon according to some victim characteristics. Implications of
the findings are discussed in light of police suspect prioritisation.
Keywords
Sexual homicide; sexual murderer; choice of weapon; murder weapon; victim
characteristic; offender profiling
Corresponding Author:
Heng Choon (Oliver) Chan, Department of Applied Social Studies, City University of
Hong Kong, Tat Chee Avenue, Hong Kong.
C h a n e t a l .
P a g e | 3
Email: oliverchan.ss@cityu.edu.hk
This is an accepted manuscript version of an article published in
Journal of
Investigative Psychology and Offender Profiling.
The copyright of this article is
reserved by 2014 John Wiley & Sons, Ltd.
C h a n e t a l .
P a g e | 4
Introduction
Sexual homicide events are rare and present a low base rate. Regardless of the high
level of public interest and media attention paid to sexually motivated homicides, as
well as increased empirical evidence gathered on the subject in recent years, the
global reported rates of sexual homicides are relatively low. In North America, sexual
murder accounts for about 1% of all reported homicides annually (Chan & Heide,
2009). Only 0.6% of the total reported homicide rate with a sexual element is
reported in the US between 1976 and 2007 (a 32-year period; Chan, Frei, & Myers,
2013a). Furthermore, there were only about 600 potential sexual homicide cases
investigated by the Royal Canadian Mounted Police in Canada between 1948 and
2010 (a 62-year period; Beauregard & Martineau, 2013). Burgess et al. (1986) defined
sexual homicide as the ‘killing [of] another in the context of power, control, sexuality,
and aggressive brutality’ (p. 252). One could reasonably infer that the ‘aggressive
brutality’ referred to in this definition is associated with the viciousness of the killing
method adopted by sexual homicide offenders (SHOs). Since the past decade, a
number of studies have examined the offender’s killing method in sexual homicides,
particularly the weapons used, through different lenses (e.g. offender and victim age,
sex, racial group, and their relationship prior to the offense) and with different
samples (e.g. the US, Canada, Germany, and the UK). Although a wealth of
knowledge has been accumulated over the years, little is known about whether
victim characteristics play a role in determining the choice of weapon made by sexual
murderers in killing their victim. This information could not only extend our
knowledge on sexualmurderers’ weapon choice used during their killings, but it could
also inform police practice in the formof police suspect prioritisation (i.e. offender
profiling). Perhaps, this could help the police to strategise their investigation plans.
WEAPONS USED IN SEXUAL HOMICIDES
A plethora of literature has demonstrated that the use of firearms is not the leading
cause of death of sexual homicide victims. Sexual murderers, in comparison with
other types of killers, are likely to use more ‘intimate’ and ‘up-close’ methods to kill
their victim (Chan & Heide, 2008; Van Patten & Delhauer, 2007). Personal weapons,
such as asphyxiation and beating, are the most commonly used killing methods in
sexual homicides, whereas the next frequently documented murder weapons are
edged (e.g. knives) and contact (e.g. blunt objects) weapons (e.g. Chan & Heide, 2008;
Chan, Heide, & Myers, 2013b; Beauregard, Stone, Proulx, & Michaud, 2008; Fisher &
Beech, 2007; Greenall & Richardson, 2014; Harbot & Mokros, 2001; Proulx, Blais, &
Beauregard, 2007; Myers & Chan, 2012; Myers, Eggleston, & Smoak, 2003; Safarik,
2002; Safarik, Jarvis, & Nussbaum, 2002). However, it should be noted that the
majority of past research examining the weapon used in sexual homicides committed
by male sexual murderers consists of mixed types of victims (i.e. adult women and
men, elderly women, and children). Only a handful of studies focused specifically on
children, adult women, adult men, and the elderly (Chan & Heide, 2009).
C h a n e t a l .
P a g e | 5
Sadistic fantasy has long been associated with sexual homicide, and the
relationship between the offender’s sadistic personality and the choice of murder
weapon has also been substantially studied (Beauregard & Proulx, 2002; Beauregard,
Proulx, & St-Yves, 2007; Proulx, et al., 2007). Compared with the more ‘distant’ type
of killing method, such as the use of firearms, close-contact killing methods are able
to provide fantasyprone SHOs a means by which to achieve their sadistic
psychological gratification, or sexual euphoria, through the expression of power,
anger, or a combination of both, against the victim (Chan, Heide, & Beauregard,
2011). Overkill is not uncommon in sexual homicides (Beauregard & Martineau, 2013;
Chan & Heide, 2009; Redojević, Radnić, Petković, Miljen, Čurović, Čukić, et al., 2013).
According to Beauregard and Proulx (2002), sadistic SHOs are more likely to sexually
humiliate their victims, use physical restraints to control their victims, and perform
body mutilation on their victims. In comparison with non-sadistic sexual killers, the
use of a personal weapon is the most common cause of victim death by sadistic
sexual murderers.
In addition, to satisfy the sexual killers’ psychosexual needs, practicality is also
a consideration in their murder weapon selection. The victim’s vulnerability (i.e.
potential physical resistance) and environmental concern (i.e. potential situational or
external interference with the assault) play a vital role in the offenders’ choice of
weapon. Personal weapons are frequently used by SHOs, especially in murdering
child and elderly victims who are more physically vulnerable (Beauregard, et al., 2008;
Safarik, et al., 2002). Personal (e.g. strangulation) and edged (e.g. stabbing) weapons
are commonly seen in the sexual killing of adult women (Chan & Heide, 2009).
Objects or tools, other than personal weapons, are utilised by SHOs in killing adult
men (Beauregard & Proulx, 2007). To a certain extent, the sexual murderers’ choice of
weapon is determined by their offending motive, which could be to satiate their
sadistic gratification through performing sadistic acts on their victim, or to kill to
avoid detection (Greenall, 2012; Kerr, Beech, & Murphy, 2013; Lobato, 2000).
Offenders with the latter motive are more likely to complete their offense with as
little extraneous activity as possible and exert only the amount of violence required
to kill their victim (Balemba, Beauregard, & Martineau, 2014). Hence, physical beating
and strangulation of their victim are uncommon amongst this type of SHOs
(Beauregard & Proulx, 2002).
Sexual killers’ age and body build also dictate their murder weapon selection.
Depending on the offenders’ physical strength and their victim’s physical
vulnerability, contact and personal weapons, as well as firearms, are commonly used
in sexual homicides committed by juveniles (Myers, 2002; Myers & Chan, 2012; Myers,
et et al., 2003). It is noteworthy that multiple killing methods are being used in many
juvenile sexual homicides as well. Interestingly, recent studies by Chan and Frei (2013)
and Chan et al. (2013a) on weapons used by female sexual murderers found that
firearms are their weapon of choice. The authors suggested that such a choice in
female sexual homicides was partly due to the preference for a weapon that requires
less physical force, particularly against victims with larger physical size and strength.
C h a n e t a l .
P a g e | 6
The Present Study
Although some studies have devoted their efforts to specifically investigating the
differences in murder weapon(s) used by SHOs on the basis of the offender’s and
victim’s age, sex, race, and relationship prior to the offense (e.g. Chan & Frei, 2013;
Chan, et al., 2013a; Chan & Heide, 2008; Chan, Heide, & Myers, 2013b), little is known
about whether the victim characteristics play a role in determining the choice of
weapon made by sexual murderers in killing their victim. Therefore, the present study
aims to examine the interactions between victim characteristics and the choice of
weapon type made by single-victim male SHOs. On the basis of the knowledge
produced by previous research, it is hypothesised that the sexual murderers’ choice
of weapon typeas being a personal weapon, an edged weapon, or other weapon
types—is in part influenced by the victims’ characteristics.
Method
Data and Procedure
The data used in this study were derived from the Federal Bureau of Investigation’s
(FBI) Supplementary Homicide Reports (SHRs) for the years 1976 through 2011 (Fox
& Swatt, 2008; FBI, 2008–2011). This offense data consist of both the offender’s and
victims demographics, as well as the characteristics of each homicide event reported
to the FBI by participating law enforcement agencies across the nation. For the
purpose of this study, sample subjects were individuals arrested for homicide in
conjunction with ‘rape’ and ‘other sexual offenses’ (coding terms used in the SHR). It
should be noted, however, that the homicides recorded in these data were classified
as sexual in nature by the investigating law enforcement agencies, and not by the
authors of this study. During this 36-year period of examination, 686,398 individuals
were arrested for homicides, and only 5,914 (0.86%) cases with pertinent
offenserelated information were classified as sexual homicide. In this study, 2,472
cases of single-victim male SHOs with no missing values on tested variables were
selected. Cases involving female SHOs, multiple-victim incidents, and missing values
on tested variables were removed from the analysis.
Measures
Sexual murderers in this study were categorised as juveniles (aged 17 years and
younger) or adults (aged 18 years and older). Victims, however, were divided into
four age groups: (1) child (aged 12 years and younger); (2) adolescent (aged 13 to 17
years); (3) adult (aged 18 to 59 years); and (4) the elderly (aged 60 years and older).
Coded by the FBI, the offender’s and victim’s racial groups examined in this study
were White, Black, and others (Asian and Pacific Islander, and American Indian or
Alaskan Native). It is important to note that the SHR database does not code
offenders and victims as multiracial, and it no longer records Hispanic origin.
For the purpose of the analyses, a number of dichotomous variables were
created. The offenders’ choice of weapon type was categorised as either a personal
C h a n e t a l .
P a g e | 7
weapon or an edged weapon. Personal weapons include killing with hands and feet,
strangulation, beating, asphyxiation, drowning, and defenestration (i.e. the act of
throwing someone out of a window), whereas edged weapons are referred to as
different types of knives. Victims were dichotomised as adults (i.e. aged 18 years and
older) or non-adults (i.e. aged 17 years and younger; adolescents and children), and
either strangers (i.e. no known relationship prior to the offense) or non-strangers.
Sexual homicides were each classified as either an intraracial (i.e. the offender killed
within his or her own race) or interracial (i.e. the offender killed outside of his or her
race) killing. Guided by the US Census Bureau’s grouping system, the urbanness level
of the offense location was categorised as either in a more populated (i.e. large and
small cities with a population of at least 2,500) or less populated geographical area
(i.e. suburban and rural areas).
Sample subjects and offense characteristics
The demographic characteristics of the SHOs and their victim and the descriptive
findings of the offenders’ offending process are shown in Table 1.
Table 1. Demographic characteristics and the dependent and independent
descriptive variables of the single-victim male sexual homicide sample extracted from
the Uniform Crime Reports (US): Supplementary Reports, 19762011 (N= 2,472)
Variables
Number of
cases
Percent of total
(100.0%)
Demographic
Sexual homicide offender’s age group
Juvenile sexual murderer
205
8.3
Adult sexual murderer
2,267
91.7
Sexual homicide offender’s race
White
1,494
60.4
Black
978
39.6
Victim’s age group
Child
328
13.3
Adolescent
276
11.1
Adult
1,549
62.7
Elderly
319
12.9
Victim’s race
White
1,799
72.8
Black
627
25.4
Others
46
1.8
Dependent
Use of an edged weapon
Yes
724
29.3
No
1,748
70.7
C h a n e t a l .
P a g e | 8
Sexual homicide offenders
Of 2,472 sexual homicide cases examined in this study, nearly 92% of the offenders
were at least 18 years of age, with a total offender mean age of 27.1 years (standard
deviation [SD] = 8.62). Approximately 60% of these sexual murderers were identified
as White offenders
Victims of sexual homicide offenders
The average age of the victim was 32.5 years (SD = 21.06), with nearly 76% of them
being at least 18 years of age. A large majority of the victims were women (81.1%),
and approximately three-quarters of them were White (72.8%) and adults (73.4%).
Types of weapon used
Pertaining to the SHOs’ choice of weapon, 42.9% of the offenders used a personal
weapon to kill their victims, whereas 29.3% of them chose to use an edged weapon
to commit their offense.
Victimoffender relationship and geographical urbanness level of the offense
occurrence
In terms of the victimoffender relationship, 34.6% of the victims were strangers to
their offender with no prior known relationship. Interestingly, 80.6% of the sexual
killings in this study were intraracial oriented, in which the offender and his or her
Use of a personal weapon
Yes
1,060
42.9
No
1,412
57.1
Independent
Female victim
Yes
2,005
81.1
No
467
18.9
Adult victim
Yes
1,814
73.4
No
658
26.6
Stranger victim
Yes
855
34.6
No
1,617
65.4
An interracial killing
Yes
1,993
80.6
No
479
19.4
Offense committed in a more populated
area
Yes
1,015
41.1
No
1,457
58.9
C h a n e t a l .
P a g e | 9
victim were from the same racial background. Findings also show that 58.9% of the
sexual murders were committed in suburban and rural areas, which are less
populated geographical environments.
Analytic strategy
As the main goal of the study was to examine the interactions between victim
characteristics and the choice of weapon by SHOs, a two-step strategy was used for
the analyses. First, chi-square automatic interaction detector (CHAID) analysis was
used to identify the interactions between the victim characteristics and the choice of
weapon, as well as to identify those most strongly associated with the choice of
weapon. Next, conjunctive analysis procedures were used to assess the impact of the
interaction of those factors, and the relative strength of each combination actually
found in the data. This was performed by determining the odds of any particular
combination that led to the choice of a specific weapon type.
Chi-square automatic interaction detector
Chi-square automatic interaction detector is a type of decision-tree technique that
can be readily used to assess categorical predictor variables and the strength and
significance of their association with an outcome state. This technique alternately
relies on Pearson association or likelihood ratio measures by which it automatically
computes a series of crosstabulations for all pairs of independent variables (Kass,
1980). The most significant of these cross-tabulation results is then incorporated into
a classification tree. The tree divides the data into mutually exclusive subsetsor
nodesthat account for all of the data and that best describe the dependent
variable (Kass, 1980).
The ordering of each successive split of the data is an important feature of
CHAID. The top node contains all of the data, and the most significant variable
associated with the dependent variable determines the first split. Each node, or
group of cases, is then sequentially split on the basis of the significance of the
variables and interactions within that node. This splitting process is continued until a
stopping rule is invoked or until there are no more variables that significantly split
the remaining cases. In the present analysis, a variation on the CHAID procedure
known as exhaustive CHAID was used. Exhaustive CHAID optimises the selection of
the appropriate variable splitting by a more thorough generation of predictor-to-
outcome comparisons (Biggs, DeVille, & Suen, 1991).
Conjunctive analysis
Miethe, Hart, and Regoeczi (2008) proposed another technique that is useful
for exploring multivariate relationships amongst categorical variables. Conjunctive
analysis is an extension of tabular contingency analysis and qualitative comparative
analysis and consists of compiling a matrix of all possible interactive combinations
for a given series of binary categorical variables. The list of all possible combinations
is then organised along an analytic dimension, and interactive relationships amongst
C h a n e t a l .
P a g e | 10
variables are assessed in relation to some qualitative premise of interest. Conjunctive
analyses have been performed to investigate different outcomes related to violent
crime and policing issues (e.g. Beauregard & Mieczkowski, 2012; Hart & Miethe, 2008;
Mieczkowski & Beauregard, 2010, 2012).
The appeal of conjunctive analysis resides in its presentation of all possible
interaction terms and the assignment of odds or probabilities associated with each
particular interaction complex. Another important element, and one particularly
relevant to the choice of weapon in sexual homicide, is that the technique permits an
exploration of interactions amongst categorical variables, that is, that the same
outcome can be created by more than one multivariate configuration of the
antecedent conditions that the analysis is exploring. By using this approach and
considering all possible combinations of variables (i.e. a saturated model), one can
target three different aspects of the causeeffect relationship: (1) the determination
of the fewest number of factors that appear to be related to the outcome state; (2)
the potential identification of some set of necessary conditions (i.e. elements that
appear in every case of the outcome state); and (3) the potential identification of
some set of sufficient conditions (i.e. elements that, if present, result in the outcome
state).
The limitation of conjunctive analysis is that the matrix of interaction terms
can quickly grow very large (e.g. the number of theoretical combinations for a binary
variable is 2n). Thus, a matrix with more than five or six binary variables quickly
becomes unwieldy. The conjunctive matrix allows for the calculation of the odds of
the occurrence of any particular n-way interaction term. Overall, the dynamics of the
conjunctive analysis approach suggests a technique that identifies important
variables that appear to be associated with the dependent or predicted state, whilst
linked to a method that considers all of the interactions of those variables.
In this study, five dichotomised variables relevant to victim characteristics were
selected. For any given dichotomised variable, the possible number of combinations
is 2n, where n is the number of variables included in the matrix. Odds ratios were
calculated, which describe the probability that a specific type of weaponwhether an
edged weapon or a personal weaponwas used to commit a sexual homicide for
each combination of victim characteristics. For instance, an odds ratio greater than
1.0 indicates that this particular combination increases the probability, whereas an
odds ratio of less than 1.0 means it decreases the probability that an offender will
choose to use an edged weapon or a personal weapon during a sexual homicide.
Result
Figure 1 shows the findings of the exhaustive CHAID on the choice of an edged
weapon (hereafter, knife is used interchangeably) to commit a sexual homicide.1 The
analysis shows that the most important predictor is the age of the victim, with
1The tree depth was set at three levels with a minimum of 100 cases in the parent
nodes (which could be split further by other variables), and 50 cases in the child
C h a n e t a l .
P a g e | 11
nodes (which could not be split further). Cross-validation was utilised with three
sample folds to ensure greater validity.
Figure 1. Exhaustive chi-square automatic interaction detector decision tree of victim
characteristics and geographical urbanness level on offenders’ choice of an edged
weapon
.
offenders being more likely to use a knife to kill an adult (31.9%, N= 579) than a non-
adult victim (22.0%, N= 145). Moreover, the findings indicate that SHOs who target
an adult male victim from a greater geographically populated area are more likely to
kill using a knife (42.5%, N= 82). However, offenders targeting non-adult victims are
more likely to use a different weapon to murder their victim (78.0%, N= 513).
The Offener's Weapon as an
Edged Weapon =
29.3% N= 2.472
29% 71%
Non-adult Victims =
22.0% N= 658
22% 78%
Adult Victims = 31.9%
N=1814
32% 68%
Female Victims = 30.6%
N= 1445
31% 69%
Male Victims = 37.1%
N= 369
37% 63%
Lower Populated Areas =
31.2% N = 176
31% 69%
Higher Populated Areas =
42.5% N= 193
43% 57%
Victim-Age Group
Victim Sex
Geographical Urbanness Level
C h a n e t a l .
P a g e | 12
Figure 2 presents the results of the exhaustive CHAID on the offender’s choice
of a personal weapon to commit a sexual homicide. Interestingly, the most important
predictor is no longer the age of victim, but the victim’s gender. As a result, SHOs
targeting a female victim are more likely to kill with their own hands (47.6%, N= 955)
compared with the offenders who target a male victim (22.5%, N= 105). Moreover,
findings indicate that SHOs who target a female non-adult victim are more likely to
kill with their own hands (56.4%, N= 323) compared with the offenders who target an
adult male victim (14.9%, N= 55). Findings also indicate that the choice of a personal
weapon depends on the degree to which the geographic area is considered urban.
Sexual murderers are more likely to kill with their own hands when they target an
adult female victim from a more populated environment (47.2%, N= 285).
Figure 2. Exhaustive chi-square automatic interaction detector decision tree of victim
characteristics and geographical urbanness level on offenders’ choice of a personal
weapon.
Victim Age Group
Geographical
Urbanness Level
The Offender's Weapon as
a Personal Weapon =
42.9% N=2472
43% 57%
Female Victims = 47.6%
N=2005
48% 52%
Non-adult Victims = 56.4%
N=573
56% 44%
Adult Victims = 44.1%
N=1432
44% 56%
Lower Populated Areas =
41.9% N=828
42% 58%
Higher Populated Areas =
47.2% N=604
47% 53%
Male Victims = 22.5%
N=467
23% 77%
Non-adult Victims = 51.0%
N=98
51% 49%
Adult Victims = 14.9%
N=369
15% 85%
Victim Sex
Victim Age Group
C h a n e t a l .
P a g e | 13
The second step of the analysis used conjunctive analysis procedures to assess
the relative strength of each combination of the victim characteristics actually found
in the data, by determining the odds of any particular combination leading to the
choice of a specific weapon type. Table 2 presents the matrix produced by the
conjunctive analysis of the victim characteristics on the offender’s choice of an edged
weapon during a sexual homicide. Insofar as considering the odds of a decision to
use a knife during a sexual homicide, it appears that the most likely combination is
Table 2. Conjunctive analysis of victim characteristics on an edged weapon used
during sexual homicide events rank ordered by odds ratio (N= 2,472)
Combinati
on
number
Victim
as
female
Victim
as an
adult
Victim as
a
stranger
An
interracia
l killing
Edged
weapon (%)
(N = 724)
Non-edged
weapon (%) (N=
1,748)
Odd
ratio
14
No
No
Yes
No
2 (50.2%)
2 (50.2%)
2.41
12
No
Yes
No
No
19 (47.5%)
21 (52.5%)
2.18
13
No
No
Yes
Yes
11 (42.3%)
15 (57.7%)
1.77
11
No
Yes
No
Yes
84 (36.7%)
145 (63.3%)
1.4
9
No
Yes
Yes
Yes
28 (36.4%)
49 (63.6%)
1.38
1
Yes
Yes
Yes
Yes
133 (33.5%)
264 (66.5%)
1.22
8
Yes
No
No
No
10 (33.3%)
20 (66.7%)
1.21
4
Yes
Yes
No
No
55 (31.8%)
118 (68.2%)
1.13
2
Yes
Yes
Yes
No
49 (29.2%)
119 (78.8%)
0.99
3
Yes
Yes
No
Yes
202 (29.2%)
492 (70.9%)
0.99
5
Yes
No
Yes
Yes
30 (23.1%)
100 (76.9%)
0.72
10
No
Yes
Yes
No
5 (22.7%)
17 (77.3%)
0.71
7
Yes
No
No
Yes
76 (20.5%)
295 (79.5%)
0.62
15
No
No
No
Yes
11 (19.3%)
46 (80.7%)
0.58
6
Yes
No
Yes
No
4 (13.8%)
25 (86.2%)
0.39
16
No
No
No
No
1 (9.1%)
10 (90.9%)
0.24
Table 3. Conjunctive analysis of victim characteristics on a personal weapon used during sexual
homicide events rank ordered by odds ratio (N= 2,472)
Combination
number
Victim as
female
Victim as
an adult
Victim as
a stranger
An
interracial
killing
Personal
weapon (%) (N
= 1,060)
Non-personal
weapon (%)
(N= 1,412)
Odds
ratio
16
No
No
No
No
8 (72.7%)
3 (27.3%)
3.55
6
Yes
No
Yes
No
19 (65.5%0
10 (34.5%)
2.53
15
No
No
No
Yes
36 (63.2%)
21 (36.8%)
2.28
7
Yes
No
No
Yes
213 (57.4&)
158 (42.6%)
1.8
5
Yes
No
Yes
Yes
70 (53.8%)
60 (46.2%)
1.55
4
Yes
Yes
No
No
81 (46.8%)
92 (53.2%)
1.17
3
Yes
Yes
No
Yes
317 (45.7%)
377 (54.3%)
1.12
2
Yes
Yes
Yes
No
70 (41.7%)
98 (58.3%)
0.95
C h a n e t a l .
P a g e | 14
#14 (i.e. victim being male, non-adult, stranger to the offender, and of a different
race than the offender). Of the eight combinations presenting an odds ratio greater
than 1.0 (#1, 4, 8, 9, 11, 12, 13, and 14), five involve a male victim (#9, 11, 12, 13, and
14), whereas five others involve an adult victim (#1, 4, 9, 11, and 12). Sexual killers are
more likely to use a knife (#12 and 14) when the victim is male and of a different race
and whether the victim is not a stranger adult (OR = 2.18) or a stranger non-adult
(OR= 2.41). Sexual killers are the least likely to use a knife during the sexual homicide
(OR= 0.24) when the victim is male, non-adult, not a stranger, but of a different race
from the perpetrator (#16). Interestingly, in cases where the typical victim is targeted
by the SHO (i.e. female, adult, stranger, and of the same race as the offender), the
odds of using a knife are positive (OR= 1.22).
As to the use of a personal weapon (Table 3), it appears that the most likely
combination is #16 (i.e. male, non-adult, not a stranger, and of a different race than
the perpetrator). Of the seven combinations presenting an odds ratio greater than 1
(#3, 4, 5, 6, 7, 15, and 16), five involve a female victim (#3, 4, 5, 6, and 7), and five
involve a non-adult victim (#5, 6, 7, 15, and 16). Sexual murderers are more likely to
kill with their own hands (#16) when the victim is male, non-adult, not a stranger, but
of a different race than the offender (OR = 3.55). Interestingly, this combination is
also the least likely to lead to a decision to use a knife.
To push the analyses a step further, a new dimension to the geographical
urbanness level of the offense occurrence was added. As research is showing that the
type of environment may have an influence on the decisions made by offenders
during their crimes, a variable looking specifically at the geographical urbanness level
of where the crime took place (i.e. a more versus less populated area) was included.
The addition of this fifth variable increased the number of possible combinations to
32 in the conjunctive matrix. Table 4 presents the matrix produced by the conjunctive
analysis of the victim characteristics, whilst taking into account the urbanness level
on the offender’s choice of weapon during a sexual homicide. Insofar as considering
the odds of a decision to use a knife during the sexual homicide, it appears that the
most likely combination is #14 (i.e. victim being male, non-adult, stranger, of a
different race than the perpetrator, and from a more populated area). Sexual
murderers are more than seven times more likely to use a knife with such a victim
(OR= 7.24). Of the 17 combinations presenting an odds ratio greater than 1.0 (#1, 3,
4, 9, 10, 11, 12, 13, 14, 17, 18, 24, 25, 27, 28, 29, and 30), 12 involve an adult victim
1
Yes
Yes
Yes
Yes
164 (41.3%)
233 (58.7%)
0.94
8
Yes
No
No
No
12 (40.0%)
18 (60.0%)
0.89
13
No
No
Yes
Yes
6 (23.1%)
20 (76.9%)
0.4
12
No
Yes
No
No
9 (22.5%)
31 (77.5%)
0.39
9
No
Yes
Yes
Yes
12 (15.6%)
65 (84.4%)
0.25
10
No
Yes
Yes
No
3 (13.6%)
19 (86.4%)
0.21
11
No
Yes
No
Yes
31 (13.5%)
198 (86.5%)
0.21
14
No
No
Yes
No
0 (0.0%)
4 (100.0%)
0.15
C h a n e t a l .
P a g e | 15
Table 4. Conjunctive analysis of victim characteristics and geographical urbanness level on an edged weapon
used during sexual homicide events rank ordered by odds ratio (N= 2,472)
Combination
number
Higher
populated
area
Victim as
female
Victim as
an adult
Victim as
a stranger
An interracial
killing
Edged weapon
(%) N=724
Non-edged
weapon (%)
N=1748
Odds ratio
14
Yes
No
No
Yes
No
1 (100.0%)
0 (0.0%)
7.24
12
Yes
No
Yes
No
No
12 (48.0%)
13 (52.0%)
2.23
28
No
No
Yes
No
No
7 (46.7%)
8 (53.3%)
2.11
29
No
No
No
Yes
Yes
8 (44.4%)
10 (55.6%)
1.93
11
Yes
No
Yes
No
Yes
52 (43.3%)
68 (56.7%)
1.85
9
Yes
No
Yes
Yes
Yes
13 (38.2%)
21 (61.8%)
1.49
13
Yes
No
No
Yes
Yes
3 (37.5%)
5 (62.5%)
1.45
24
No
Yes
No
No
No
7 (36.8%)
12 (63.2%)
1.41
4
Yes
Yes
Yes
No
No
29 (36.3%)
51 (63.7%)
1.37
10
Yes
No
Yes
Yes
No
5 (35.7%)
9 (64.9%)
1.34
18
No
Yes
Yes
Yes
No
27 (35.1%)
50 (64.9%)
1.3
25
No
No
Yes
Yes
Yes
15 (34.9%)
28 (65.1%)
1.29
1
Yes
Yes
Yes
Yes
Yes
53 (34.0%)
103 (66.0%)
1.24
30
No
No
No
Yes
No
1 (33.3%)
2 (66.7%)
1.21
17
No
Yes
Yes
Yes
Yes
80 (33.2%)
161 (66.8%)
1.2
3
Yes
Yes
Yes
No
Yes
88 (31.8%)
189 (68.2%)
1.12
27
No
No
Yes
No
Yes
32 (29.4%)
77 (70.6%)
1
20
No
Yes
Yes
No
No
26 (28.0%)
67 (72.0%)
0.94
19
No
Yes
Yes
No
Yes
114 (27.1%)
303 (72.7%)
0.91
8
Yes
Yes
No
No
No
3 (27.3%)
8 (72.7%)
0.91
21
No
Yes
No
Yes
Yes
24 (25.8%)
69 (74.2%)
0.84
2
Yes
Yes
Yes
Yes
No
22 (24.2%)
69 (75.8%)
0.8
15
Yes
No
No
No
Yes
6 (22.2%)
22 (77.8%)
0.69
23
No
Yes
No
No
Yes
54 (21.3%)
200 (78.7%)
0.65
7
Yes
Yes
No
No
Yes
22 (18.8%)
95 (81.2%)
0.56
31
No
No
No
No
Yes
5 (16.7%)
25 (83.8%)
0.48
5
Yes
Yes
No
Yes
Yes
6 (16.2%)
31 (83.8%)
0.47
22
No
Yes
No
Yes
No
3 (15.0%)
17 (85.0%)
0.43
32
No
No
No
No
No
1 (14.3%)
6 (85.7%)
0.4
6
Yes
Yes
No
Yes
No
1 (11.1%)
8 (88.9%)
0.3
16
Yes
No
No
No
No
0 (0.0%)
4 (100.0%)
0.27
26
No
No
Yes
Yes
No
0 (0.0%)
8 (100.0%)
0.14
(#1, 3, 4, 9, 10, 11, 12, 17, 18, 25, 27, and 28), 11 involve a male victim (#9, 10, 11, 12,
C h a n e t a l .
P a g e | 16
13, 14, 25, 27, 28, 29, and 30), and 10 involve a stranger victim (#1, 9, 10, 13, 14, 17,
18, 25, 29, and 30). Moreover, of the 17 combinations presenting a greater likelihood
Table 4. Conjunctive analysis of victim characteristics and geographical urbanness level on an edged weapon
used during sexual homicide events rank ordered by odds ratio (N= 2,472)
Combination
number
Higher
populated
area
Victim as
female
Victim as
an adult
Victim as
a stranger
An interracial
killing
Edged weapon
(%) N=724
Non-edged
weapon (%)
N=1748
Odds ratio
16
Yes
No
No
No
No
4 (100.0%)
0 (0.0%)
11.99
6
Yes
Yes
No
Yes
No
8 (88.9%)
1 (11.1%)
10.66
5
Yes
Yes
No
Yes
Yes
25 (67.6%)
12 (32.4%)
2.78
31
No
No
No
No
Yes
19 (63.3%)
11 (36.7%)
2.30
15
Yes
No
No
No
Yes
17 (63.0%)
10 (37.0%)
2.27
7
Yes
Yes
No
No
Yes
71 (60.7%)
46 (39.3%)
2.06
32
No
No
No
No
No
4 (57.1%)
3 (42.9%)
1.78
23
No
Yes
No
No
Yes
142 (55.9%)
112 (44.1%)
1.69
22
No
Yes
No
Yes
No
11 (55.5%)
9 (45.0%)
1.63
21
No
Yes
No
Yes
Yes
45 (48.4%)
48 (51.6%)
1.25
3
Yes
Yes
Yes
No
Yes
132 (47.7%)
145 (52.3%)
1.21
4
Yes
Yes
Yes
No
No
38 (47.5%)
42 (52.5%)
1.21
2
Yes
Yes
Yes
Yes
No
43. (47.3%)
48 (52.7%)
1.19
20
No
Yes
Yes
No
No
43 (46.2%)
50 (64.9%)
1.15
1
Yes
Yes
Yes
Yes
Yes
72 (46.2%)
84 (53.8%)
1.14
8
Yes
Yes
No
No
No
5 (45.4%)
6 (54.5%)
1.11
19
No
Yes
Yes
No
Yes
185 (44.4%)
232 (55.6%)
1.06
17
No
Yes
Yes
Yes
Yes
92 (38.2%)
149 (61.8%)
0.82
13
Yes
No
No
Yes
Yes
3 (37.5%)
5 (62.5%)
0.80
26
No
No
Yes
Yes
No
3 (37.5%)
5 (62.5%)
0.80
24
No
Yes
No
No
No
7 (36.8%)
12 (63.2%)
0.77
18
No
Yes
Yes
Yes
No
27 (35.1%)
50 (64.9%)
0.72
28
No
No
Yes
No
No
4 (26.7%)
11 (73.3%)
0.48
14
Yes
No
No
Yes
No
0 (0.0%)
1 (100.0%)
0.44
9
Yes
No
Yes
Yes
Yes
8 (23.5%)
26 (73.5%)
0.41
12
Yes
No
Yes
No
No
5 ()20.0%
20 (80.0%)
0.33
29
No
No
No
Yes
Yes
3 (16.7%)
15 (83.3%)
0.27
27
No
No
Yes
No
Yes
15 (13.8%)
94 (86.2%)
0.21
11
Yes
No
Yes
No
Yes
16 (13.3%)
104 (86.7%)
0.20
30
No
No
No
Yes
No
0 (0.0%)
3 (100.0%)
0.19
25
No
No
Yes
Yes
Yes
4 (9.3%)
39 (90.7%)
0.14
10
Yes
No
Yes
Yes
No
0 (0.0%)
14 (100.0%)
0.05
C h a n e t a l .
P a g e | 17
of using a knife, nine involve a victim of the same race as the perpetrator (#1, 3, 9, 11,
13, 17, 25, 27, and 29), and nine others involve a sexual homicide committed in an
environment that has a greater population (#1, 3, 4, 9, 10, 11, 12, 13, and 14). SHOs
are the least likely to use a knife during the sexual homicide (OR= 0.14) when the
victim is male, adult, stranger, and of a different race than the offender, and when the
crime is committed in an area that is highly populated (#26).
As to the decision to kill with their own hands (Table 5), it appears that the
most likely combination is #16 (i.e. victim is male, non-adult, not a stranger, of a
different race than the perpetrator, and from a more populated environment),
followed closely by combination #6 (i.e. victim is female, non-adult, a stranger, of a
different race than the perpetrator, and from a more populated area). When using
these combinations, SHOs are approximately 12 (OR = 11.99) and 11 (OR = 10.66)
times more likely to kill the victim with their own hands during the sexual homicide.
Of the 17 combinations presenting odds ratios greater than 1.0 (#1, 2, 3, 4, 5, 6, 7, 8,
15, 16, 19, 20, 21, 22, 23, 31, and 32), 13 involve a female victim (#1, 2, 3, 4, 5, 6, 7, 8,
19, 20, 21, 22, and 23), and 11 involve a non-adult (#5, 6, 7, 8, 15, 16, 21, 22, 23, 31,
and 32) who is not a stranger to the offender (#3, 4, 7, 8, 15, 16, 19, 20, 23, 31, and
32). Moreover, of the 17 combinations presenting a greater likelihood of sexual
murderers killing their victims with their own hands, 10 are committed in an area with
a greater population (#1, 2, 3, 4, 5, 6, 7, 8, 15, and 16), and nine others involve a
victim of the same race as the offender (#1, 3, 5, 7, 15, 19, 21, 23, and 31). Sexual
killers are the least likely to kill with their own hands (OR = 0.05) when the victim is
male, adult, stranger, and of a different race than the perpetrator, and when the
crime is committed in a more populated environment (#10). Interestingly, this
combination was also less likely to be observed when SHOs were using a knife,
suggesting that when offenders target such a victim in a more populated area, they
are less likely to both use a knife and kill with their own hands. Finally, it is
noteworthy that when positive odds ratios are observed for the use of a knife to kill
the victim, the odds are negative for the use of a personal weapon, except for three
specific combinations (#1, 3, and 4).
Discussion
The present study used the SHR data that spanned over a 36-year period (19762011)
to examine the interactions between victim characteristics and choice of weapon
made by 2,472 single-victim male sexual murderers. Weapons were categorised as
either personal or edged weapons. Using a two-step strategy (i.e. exhaustive CHAID
and conjunctive analysis), the victim characteristics that are most strongly associated
with a particular choice of weapon, as well as the odds of any particular combination
of victim characteristics leading to a specific weapon choice, were identified.
A couple of key and meaningful observations have emerged out of these
analyses. First, an edged weapon is preferred by these male SHOs when their target
is a male victim (i.e. homosexual homicide), especially those who are from a different
racial background (i.e. an interracial killing). This preference of using an edged
C h a n e t a l .
P a g e | 18
weapon in male-on-male sexual homicides is somewhat consistent with Myers and
Chan’s (2012) recent empirical study on 93 juvenile homosexual homicides (i.e. 88
maleon- male and five female-on-female sexual murders), whereby 40% of their
sample used edged or contact (e.g. blunt objects) weapons to sexually kill their male
victim. However, the findings of this study contradict those of Myers and Chan (2012)
who found that intraracial killing was more apparent in juvenile homosexual
homicides (93% for White on White and 69% for Black on Black). Perhaps a different
victimoffender racial phenomenon has emerged when the offender’s age and sex
are taken into consideration, as the present study focused only on male offenders
regardless of their age.
The second key finding observed is that a personal weapon is preferred by
male SHOs when their victim is perceived to be physically weaker than they are, such
as non-adult victims (e.g. children and adolescents) of both sexes, especially those
from a different racial background. This is not a novel finding. In the study by Chan
and Heide (2008), personal weapons were commonly used by juvenile and adult
sexual murderers to kill child (64% by juvenile and 30% by adult SHOs) and
adolescent victims (61% by juvenile and 43% by adult SHOs). Heide’s (1993) physical
strength hypothesis was supported in this case, in which the offenders’ choice of
weapon is made possible on the basis of the discrepancy between victim and
offender as it pertains to their physical build and strength. In this study, personal
weapons that required close contact with the victims were preferred when the victims,
such as children and adolescents, were perceived by the offenders to be at a
disadvantage in terms of their physical ability to resist a deadly assault. Moreover,
personal weapons allow the sexual murderers to ‘enjoy’ their killing process by
fulfilling their deviant and/or sadistic sexual fantasies in very close proximity to their
victim. Pertaining to the interracial killing, this trend is also observed in Chan, Myers,
and Heide’s (2010) study of 3,868 sexual murderers. Although White SHOs
predominantly killed within their own race, Black SHOs, in their study, murdered both
intraracially and interracially. The present study is limited in that no information on
the offender’s racial profile was examined. Nevertheless, it is beyond the scope of
this study to explore the potential effect of the offender’s racial profile in his or her
weapon choice.
It should be noted that weapon choice in sexual homicides is a complex
phenomenon with many origins. Besides victim characteristics, weapon lethality, the
offender’s psychological and emotional state during the offense (e.g. predatory
versus affective killings), and he offender’s true motive (e.g. serial versus one-off
killing), there are several other potential factors that need to be considered to
determine the sexual killer’s ultimate choice of weapon (Adjorlolo & Chan, 2014;
Chan, Beauregard, & Myers, 2014; Drawdy & Myers, 2004; Myers, Chan, Vo, &
Lazarou, 2009). A case-by-case analysis is ultimately necessary to uncover the
offender’s rationale behind his specific weapon choice. In addition, this study has
several data limitations. First, the SHR dataset is compiled from arrests, and not
convictions. Hence, it is unclear whether these arrested individuals were subsequently
C h a n e t a l .
P a g e | 19
charged and/or convicted. In addition, findings from this study are also limited to
only known cases. The weapon choice of those who have successfully evaded
apprehension or avoided detection remains unclear. Furthermore, the SHR data are
restricted to basic offender, victim, and offense circumstance information. A more in-
depth examination into other potential victimology factors that affect the offender’s
choice of weapon, such as victim resistance and other victimoffender interactions
during the offense, is not made possible.When the depth of data permits, future
researchers may consider examining more victim characteristics, in conjunction with
other possible contributing factors (e.g. offending and circumstantial such as the
exact location where the offense occurred), to acquire a better understanding of
sexual murderers’ choice of weapon. In addition, different statistical methods (e.g.
mediation and moderation models) that allow for an exploration of the different
interactions between variables may be considered in future studies. It is only by
comprehensively analysing the data that more complex relationships that potentially
influence the SHOs’ choice of weapon could be uncovered. Clearly, much remains to
be learned about this topic
Regardless of the noted data limitations, this study has nevertheless offered
an initial understanding of the role of victim characteristics in influencing the SHOs’
choice of weapon. On the basis of the two key observations in this study, practical
implications in the area of police prioritisation of suspects are proposed. If a non-
adult (i.e. a male or female child or adolescent) has been sexually murdered by a
personal weapon, or a male sexual homicide victim has been killed by an edged
weapon, and the police have reason to believe that this offense is committed by a
male offender, the probability that he is from a different racial background (i.e. an
interracial killing) is increased. Moreover, in cases where a non-adult victim is sexually
murdered by a personal weapon, the male offender is likely to be physically stronger
than the victim. These observations contradict the findings of most sexual homicide
studies where it has been found that intraracial sex killings are more prevalent (Chan,
et al., 2010). Therefore, it is hoped that these findings will provide the police with
some practical utility when prioritising their investigative efforts. Nonetheless, these
suggestions should be interpreted cautiously as they are derived from a sample of
single-victim SHOs. Sexual murderers with more than one victim or those who have
successfully avoided detection may present a different weapon choice profile.
C h a n e t a l .
P a g e | 20
References
Adjorlolo, S., & Chan, H. C. O. (2014). The controversy of defining serial murder:
Revisited. Aggression and Violent Behavior, 19(5), 486491.
Balemba, S., Beauregard, E., & Martineau, M. (2014). Getting away with murder: A
thematic approach to solved and unsolved sexual homicides using crime
scene factors. Police Practice and Research, 15(3), 221233.
Beauregard, E. &Martineau,M. (2013). A descriptive study of sexual homicide in
Canada: Implications for police investigation. International Journal of Offender
Therapy and Comparative Criminology, 57(12), 14541476.
Beauregard, E., & Mieczkowski, T. (2012). Risk estimations of the conjunction of
victim and crime event characteristics on the lethal outcome of sexual assaults.
Violence and Victims, 27(4), 470486.
Beauregard, E., & Proulx, J. (2007). A classification of sexual homicide against men.
International Journal of Offender Therapy and Comparative Criminology, 51(4),
420432.
Beauregard, E., & Proulx, J. (2002). Profiles in the offending process of nonserial
sexual murderers. International Journal of Offender Therapy and Comparative
Criminology, 46(4), 386399.
Beauregard, E., Proulx, J., & St-Yves, M. (2007). Angry or sadistic: Two types of sexual
murderers. In J. Proulx, E. Beauregard, M. Cusson, & A. Nicole (Eds), Sexual
murderers: A comparative analysis and new perspectives (pp. 123141).
Hoboken, NJ: Wiley.
Beauregard, E., Stone, M. R., Proulx, J., & Michaud, P. (2008). Sexual murderers of
children: Developmental, precrime, crime, and postcrime factors. International
Journal of Offender Therapy and Comparative Criminology, 52(3), 253269.
Biggs, D., DeVille, B., & Suen, E. (1991). A method of choosing multiway partitions for
classification and decision trees. Journal of Applied Statistics, 18, 4962.
Burgess, A. W., Hartman, C. R., Ressler, R. K., Douglas, J. E., & McCormack, A. (1986).
Sexual homicide: A motivational model. Journal of Interpersonal Violence, 1,
251272.
Chan, H. C. O. & Frei, A. (2013). Female sexual homicide offenders: An examination of
an underresearched offender population. Homicide Studies, 17(1), 96118.
Chan, H. C. O., & Heide, K. M. (2009). Sexual homicide: A synthesis of the literature.
Trauma, Violence, and Abuse, 10(1), 3154.
Chan, H. C. O., & Heide, K. M. (2008). Weapons used by juveniles and adult offenders
in sexual homicides: An empirical analysis of 29 years of US data. Journal of
Investigative Psychology and Offender Profiling, 5(3), 189208.
Chan, H. C. O., Beauregard, E., & Myers, W. C. (2014). Single-victim and serial sexual
homicide offenders: Differences in crime, paraphilias, and personality traits.
Criminal Behaviour and Mental Health. Advance online publication. doi:
10.1002/cbm.1925.
C h a n e t a l .
P a g e | 21
Chan, H. C. O., Frei, A., & Myers, W. C. (2013a). Female sexual homicide offenders: An
analysis of the offender racial profiles in offending process. Forensic Science
International, 233(1-3), 265272.
Chan, H. C. O., Heide, K. M., & Beauregard, E. (2011). What propels sexual murderers:
A proposed integrated theory of social learning and routine activities theories.
International Journal of Offender Therapy and Comparative Criminology, 55(2),
228250.
Chan, H. C. O., Heide, K. M., & Myers, W. C. (2013b). Juvenile and adult offenders
arrested for sexual homicide: An analysis of victimoffender relationship and
weapon used by race. Journal of Forensic Sciences, 58(1), 8589.
Chan, H. C. O., Myers, W. C., & Heide, K. M. (2010). An empirical analysis of 30 years
of US juvenile and adult sexual homicide offender data: Race and age
differences in the victim offender relationship. Journal of Forensic Sciences,
55(5), 12821290.
Drawdy, S. M., & Myers, W. C. (2004). Homicide victim/offender relationship in Florida
Medical Examiner District 8. Journal of Forensic Sciences, 49(1), 150154.
Federal Bureau of Investigation (2010-2013). Uniform Crime Reporting Program
data [United States]: Supplementary Homicide Reports, 20082011
[Computer file]. Ann Arbor, MI: Inter-university Consortium for Political and
Social Research.
Fisher, D., & Beech, A.R. (2007). Identification of motivations for sexual murder. In J.
Proulx, E. Beauregard, M. Cusson, & A. Nicole (Eds.), Sexual murderers: A
comparative analysis and new perspectives (pp. 175190). Hoboken, NJ: Wiley.
Fox, J. A., & Swatt, M. L. (2008). Uniform Crime Reports [United States]: Supplemental
homicide reports with multiple imputation, cumulative files 19762005
[Computer file]. Ann Arbor, MI: Inter-university Consortium for Political and
Social Research.
Greenall, P. V. (2012). Understanding sexual homicide. Journal of Sexual Aggression,
18(3), 338354.
Greenall, P. V., & Richardson, C. (2014). Adult male-on-female stranger sexual
homicide: A descriptive (baseline) study from Great Britain. Homicide Studies.
Advance online publication. doi: 10.1177/ 1088767914530555.
Harbot, S., & Mokros, A. (2001). Serial murderers in Germany from 1945 to 1995: A
descriptive study. Homicide Studies, 5, 311334.
Hart, T. C., & Miethe, T. D. (2008). Exploring bystander presence and intervention in
nonfatal violent victimization: When does helping really help? Violence and
Victims, 23, 637651.
Heide, K. M. (1993). Weapons used by juveniles and adults to kill parents. Behavioral
Sciences and the Law, 11, 397405.
Kass, G. V. (1980). An exploratory technique for investigating large quantities of
categorical data. Applied Statistics, 29, 119127.
C h a n e t a l .
P a g e | 22
Kerr, K. J., Beech, A. R., & Murphy, D. (2013). Sexual homicide: Definition, motivation,
and comparison with other forms of sexual offending. Aggression and Violent
Behavior, 18(1), 110.
Lobato, A. (2000). Criminal weapon use in Brazil: A psychological analysis. In D.V.
Canter & L.J. Alison (Eds.), Profiling property crimes, offender profiling series,
Vol. IV (pp. 107145). Ashgate, UK: Dartmouth.
Mieczkowski, T., & Beauregard, E. (2012). Interactions between disinhibitors in sexual
crimes: Additive or counteracting effects? Journal of Crime and Justice, 35(3),
395411.
Mieczkowski, T., & Beauregard, E. (2010). Lethal outcome in sexual assault events: A
conjunctive analysis. Justice Quarterly, 27(3), 332361.
Miethe, T. D., Hart, T. C., & Regoeczi, W. C. (2008). The conjunctive analysis of case
configurations: An exploratory method for discrete multivariate analyses of
crime data. Journal of Quantitative Criminology, 24, 227241.
Myers, W. C. (2002). Juvenile sexual homicide. San Diego, CA: Academic Press.
Myers, W. C., & Chan, H. C. O. (2012). Juvenile homosexual homicide. Behavioral
Sciences and the Law, 30(2), 90102.
Myers, W. C., Chan, H. C. O., Vo, E. J., & Lazarou, E. (2009). Sexual sadism,
psychopathy, and recidivism in juvenile sexual murderers. Journal of
Investigative Psychology and Offender Profiling, 7(1), 4958.
Myers, W. C., Eggleston, C. F., & Smoak, P. (2003). A media violence-inspired juvenile
sexual homicide offender 13 years later. Journal of Forensic Sciences, 48, 15.
Proulx, J., Blais, E., & Beauregard, E. (2007). Sadistic sexual offenders. In J. Proulx, E.
Beauregard, M. Cusson, & A. Nicole (Eds.), Sexual murderers: A comparative
analysis and new perspectives (pp. 107122). Hoboken, NJ: Wiley.
Radojević, N., Radnić, B., Petković, S., Miljen, M., Čurović, I., Čukić, D., et al. (2013).
Multiple stabbing in sex-related homicides. Journal of Forensic and Legal
Medicine, 20, 502507.
Safarik, M. E. (2002). The elderly female victim. In W. C. Myers (Ed.), Juvenile sexual
homicide (pp. 147161). San Diego, CA: Academic Press.
Safarik, M. E., Jarvis, J. P., Nussbaum, K. E. (2002). Sexual homicide of elderly females:
Linking offender characteristics to victim and crime scene attributes. Journal of
Interpersonal Violence, 17, 500525.
Van Patten, I. T., & Delhauer, P. Q. (2007). Sexual homicide: A spatial analysis of 25
years of deaths in Los Angeles. Journal of Forensic Sciences, 52, 11291141.
... A large proportion of the victims were females, whose average ages were between 27 and 37 years, with most of them (70% to 80%) being at least 18 years of age (Chan & Beauregard, 2016a;Smith et al., 2011). Similarly for female offenders, the majority of their victims (74%) were males (Chan & Frei, 2013;Chan, Frei, & Myers, 2013a). ...
... However, the lethality of the outcome does not depend solely on the offender's motivation (e.g., sexual motive). There may be other contributing factors, such as victim characteristics (e.g., victim-offender interaction, victim resistance) or circumstantial conditions (e.g., the exact location where the offense occurred, substance-related intoxication) (Chan & Beauregard, 2016a;Chopin & Beauregard, 2019a;Mieczkowski & Beauregard, 2010). Thus, this study argues that despite the claims that sex worker homicides bear a certain resemblance to sexual homicides, the dynamics of these types of homicides show more differences than similarities. ...
... Multiple factors may influence the offenders' decisions. Additional potential contributing factors need to be taken into account when analyzing the offender's modus operandi, such as the offender's psychological and emotional state during the offense (e.g., instrumental vs. expressive violence), the offender's true motives (e.g., sexually motivated vs. nonsexually motivated), or the availability of a weapon (e.g., weapon of choice vs. weapon of opportunity) (Adjorlolo & Chan, 2017;Chan & Beauregard, 2016a;Chan, Li, Liu, & Lu, 2019). Despite the investigation-oriented lessons derived from this study, the timeless caveat stands undiminished, whereby the police investigators and other law enforcement professionals must maintain an objective, open-minded approach to all homicide cases, and avoid any development of tunnel vision. ...
Article
Full-text available
Sex workers are commonly claimed to be at heightened risk of fatal victimization. Although prior research indicates that the dynamics of sex worker homicides resemble sexual homicides more than nonsexual homicides, little is known about how these types of homicides compare in terms of offending patterns. This study considers a sample of 2,851 single‐victim, single‐offender homicide cases extracted from a 37‐year (1976–2012) US Supplementary Homicide Reports database, and compares the offender, victim, and offender characteristics of 243 sex worker homicides (189 males and 54 females) with those of 2608 sexual homicides (2474 males and 134 females). The findings suggest that the offender, victim, and offense characteristics of general, male‐offender, and female‐offender sex worker homicides are essentially different from the characteristics of sexual homicides. Logistic regressions further indicate that most offender, victim, and offense characteristics on the occurrence of general and male‐offender sex worker homicides were significantly associated with the perpetration of general and male‐offender sex worker homicides, with reference to sexual homicides. These findings offer insights relevant to the prioritization of criminal investigative practices.
... A second limitation in past research has been choice of statistical procedure. Past research in this area have used simple cross-tabulations, frequentist methods such as chi-square (as well as more advanced chi-square methods, see Chan & Beauregard, 2016) and multinomial logistic regression (e.g., Reynolds et al., 2019), as well as correspondence analysis (e.g., Fox & Allen, 2014). These are all reasonable statistical approaches to understand the relationship between weapon type and victim-perpetrator relationship. ...
... Alternatively, poison generally requires some level of access to the victim, which would be easiest to exploit for family members. In other words, the relationship could be driven either by the nature of the relationship between victim and perpetrator, or the nature of the weapon itself (Chan & Beauregard, 2016). Regardless, we would expect this effect to replicate in other countries, as there are many available sources of poison in homes around the world. ...
Article
The weapon type used in a homicide predicts the victim-perpetrator relationship. However, there are some limitations in this past research including the common data analytic strategies. Our purpose was to build a model of weapon type, predicting relationship type, and to address previous limitations. We examined 363,927 homicides and used Bayesian multilevel categorical regression. In addition to analyzing weapon type (final model consisted of 16 weapon categories), we examined the victims' sex, age, and race as covariates and modeled the data across states and counties. Results indicate that weapon type is highly informative, however, the age of the victim and sex of the victim interact in important ways.
... These outcomes are common. Uncommon is any pathway to SH, an exceedingly rare event (Chan & Beauregard, 2016;Chan & Heide, 2009) even when considering the multitude of active offenders who have developmental histories that are conducive to lethal sexual violence. In our model, the only significant path toward SH was from antisocial problematic behaviors; thus it is very difficult to predict SH even among persons with numerous risk factors for it. ...
Article
Research focusing on the role of adverse childhood experience (ACE) of individuals involved in sexual homicide (SH) is scarce. Theoretical models of SH have postulated a connection between these adverse experiences and the development of internal risk factors. However, such assumptions have never been empirically tested. Therefore, the current research aims to identify how ACEs affect the development of personality disorders and problematic behaviors during adolescence, which constitute internal risk factors for the commission of SH. The sample comes from a database including 613 individuals involved in sexual crimes in Canada among which 60 committed a SH. Bivariate and multiple regression analyses were conducted to identify personality disorders and problematic behaviors during adolescence associated with the presence of ACEs. Next, path analysis was used to identify the direct and indirect relationships between ACEs, internal risk factors, and the commission of SH. Results showed that individuals who experienced ACEs were more likely to develop internal risk factors involved in the commission of SH. Moreover, findings suggest that the impact of these adverse experiences will differ, depending on whether the child has been victim of violence or if he/she has witnessed it.
... These outcomes are common. Uncommon is any pathway to SH, an exceedingly rare event (Chan & Beauregard, 2016;Chan & Heide, 2009) even when considering the multitude of active offenders who have developmental histories that are conducive to lethal sexual violence. In our model, the only significant path toward SH was from antisocial problematic behaviors; thus it is very difficult to predict SH even among persons with numerous risk factors for it. ...
Article
Full-text available
Research focusing on the role of adverse childhood experience (ACE) of individuals involved in sexual homicide (SH) is scarce. Theoretical models of SH have postulated a connection between these adverse experiences and the development of internal risk factors. However, such assumptions have never been empirically tested. Therefore, the current research aims to identify how ACEs affect the development of personality disorders and problematic behaviors during adolescence, which constitute internal risk factors for the commission of SH. The sample comes from a database including 613 individuals involved in sexual crimes in Canada among which 60 committed a SH. Bivariate and multiple regression analyses were conducted to identify personality disorders and problematic behaviors during adolescence associated with the presence of ACEs. Next, path analysis was used to identify the direct and indirect relationships between ACEs, internal risk factors, and the commission of SH. Results showed that individuals who experienced ACEs were more likely to develop internal risk factors involved in the commission of SH. Moreover, findings suggest that the impact of these adverse experiences will differ, depending on whether the child has been victim of violence or if he/she has witnessed it.
... Alternatively, it is also possible that the combination of multiple killing methods was due to situational factors. For example, studies have shown that personal weapons (e.g., strangulation, beating) were commonly used when the victim was perceived to be physically weaker in sexual homicide cases (Chan & Beauregard, 2016). The choice of weapon was also found to be associated with the offender's primary motivation for the crime (e.g., sexual, financial) (Chan & Li, 2020), and whether it was a weapon of opportunity (e.g., weapon found at the scene) (Salfati & Park, 2007). ...
... Alternatively, it is also possible that the combination of multiple killing methods was due to situational factors. For example, studies have shown that personal weapons (e.g., strangulation, beating) were commonly used when the victim was perceived to be physically weaker in sexual homicide cases (Chan & Beauregard, 2016). The choice of weapon was also found to be associated with the offender's primary motivation for the crime (e.g., sexual, financial) (Chan & Li, 2020), and whether it was a weapon of opportunity (e.g., weapon found at the scene) (Salfati & Park, 2007). ...
Article
Purpose The duration of time that the serial offender remains free in the community to commit murders may be seen as a direct measure of their longevity; a sign of their success. The aim of this study is to predict the duration of the serial homicide series by examining the factors that contribute to the length of time a serial murderer is able to remain free of police detection. Methods Generalized estimating equations with a negative binomial link function were used to examine factors predicting the duration of series in a sample of 1258 serial murder cases. Results Results showed that offenders' criminal history, race (i.e., White and Hispanic), and victims of minority backgrounds significantly predicted longer duration in their murder series. A combination of multiple killing methods and atypical methods also predicted longer murder series, while the moving of the victim's body predicted shorter duration in the series. Conclusions This study builds upon the serial homicide literature, particularly the duration of the series. Results from this study help inform investigative efforts in serial homicide cases.
... SHR have been largely used in the criminological literature to study homicides in the United States (Fox and Levin, 1991;Browne and Williams, 1993;Gallup-Black, 2005;Roman, 2013;Chan and Beauregard, 2016). This also applies to homicide clearance in particular (Regoeczi et al., 2000;Ousey and Lee, 2010;Mancik and Parker, 2019;Chalfin et al., 2020). ...
Preprint
Full-text available
Purpose: To explore the potential of Explainable Machine Learning in the prediction and detection of drivers of cleared homicides at the national- and state-levels in the United States. Methods: First, nine algorithmic approaches are compared to assess the best performance in predicting cleared homicides country-wise, using data from the Murder Accountability Project. The most accurate algorithm among all (XGBoost) is then used for predicting clearance outcomes state-wise. Second, SHAP, a framework for Explainable Artificial Intelligence, is employed to capture the most important features in explaining clearance patterns both at the national and state levels. Results: At the national level, XGBoost demonstrates to achieve the best performance overall. Substantial predictive variability is detected state-wise. In terms of explainability, SHAP highlights the relevance of several features in consistently predicting investigation outcomes. These include homicide circumstances, weapons, victims' sex and race, as well as number of involved offenders and victims. Conclusions: Explainable Machine Learning demonstrates to be a helpful framework for predicting homicide clearance. SHAP outcomes suggest a more organic integration of the two theoretical perspectives emerged in the literature. Furthermore, jurisdictional heterogeneity highlights the importance of developing ad hoc state-level strategies to improve police performance in clearing homicides.
... SHR have been largely used in the criminological literature to study homicides in the United States (Fox and Levin, 1991;Browne and Williams, 1993;Gallup-Black, 2005;Roman, 2013;Chan and Beauregard, 2016). This also applies to homicide clearance in particular (Regoeczi et al., 2000;Ousey and Lee, 2010;Mancik and Parker, 2019;Chalfin et al., 2020). ...
Article
Full-text available
Purpose To explore the potential of Explainable Machine Learning in the prediction and detection of drivers of cleared homicides at the national- and state-levels in the United States. Methods First, nine algorithmic approaches are compared to assess the best performance in predicting cleared homicides country-wise, using data from the Murder Accountability Project. The most accurate algorithm among all (XGBoost) is then used for predicting clearance outcomes state-wise. Second, SHAP, a framework for Explainable Artificial Intelligence, is employed to capture the most important features in explaining clearance patterns both at the national and state levels. Results At the national level, XGBoost demonstrates to achieve the best performance overall. Substantial predictive variability is detected state-wise. In terms of explainability, SHAP highlights the relevance of several features in consistently predicting investigation outcomes. These include homicide circumstances, weapons, victims' sex and race, as well as number of involved offenders and victims. Conclusions Explainable Machine Learning demonstrates to be a helpful framework for predicting homicide clearance. SHAP outcomes suggest a more organic integration of the two theoretical perspectives emerged in the literature. Furthermore, jurisdictional heterogeneity highlights the importance of developing ad hoc state-level strategies to improve police performance in clearing homicides.
Article
Purpose This research uses rational choice theory to analyze the effects of motivation, premeditation and offender characteristics on offenders' weapons during decision-making processes when they are violent towards on-duty police officers. The paper aims to discuss the aforementioned issues. Design/methodology/approach The researchers examined 597 cases ( n = 597) of violence against the police in the China Judgments Online (CJO) database, and analyzed the data using multinomial logistic regression methods. Rational choice theory was used to explore the offenders' weapons decision-making process. Findings The research results showed that offenders with premeditation were more likely to use a weapon, and tended to choose sharp weapons; offenders motivated to “escape arrest” were more likely to use a weapon, and tended to choose a vehicle as a weapon; and offenders motivated by “conflict resolution” were more likely to choose a sharp or blunt weapon. Research limitations/implications These findings have limited applicability to other countries and must be considered in the local background of violence against police. Practical implications Through the rational choice theory analytical framework, this study clarifies how motivation and premeditation influence offenders' weapons decision-making processes. Social implications Also, this study may provide support for frontline police officers' law enforcement. Originality/value The research identified some specific connections between offenders' weapon choice preferences, their motivation for the violence and whether or not there was premeditation. The findings provide guidance for police agencies developing preventive policies, and for frontline officers in interpreting and managing the situations they face.
Article
Purpose The analysis of previous studies showed that research pertaining to the examination of the crime scene and Modus Operandi variables in intimate partner homicide (IPH) is scarce. Additionally, to our knowledge, there are no studies investigating sexual homicide perpetrated by intimate partners. This study aims to address that void. Thus, the study examined various components of the crime event and as such, it was exploratory in nature. Design/methodology/approach The study consisted of male sexual killers, who perpetrated against pubescent female victims (14 years old and over) and served a custodial sentence within Her Majesty’s Prison Service in England and Wales. Variables for the study were chosen on the basis of previous research examining IPH and sexual homicide. Descriptive analyses were used in this exploratory study. Findings Descriptive analyses indicated that the most prevalent aggravating circumstances in the lead up to the killing included conflict with the victim before the offence and substance use by the perpetrator around the time of the killing. Stalking was present in approximately a quarter of cases. The results of the analyses of sexual behaviours showed that in 54.9% of the sample the act of killing was purely instrumental whereas in 39.4% of the sample the underlying drive of the act of killing was closely related to the sexual aspect. Originality/value This is a unique study on a topic not yet explored.
Article
Full-text available
Background Information on psychopathological characteristics of sexual homicide offenders is scarce.AimsTo investigate criminal, paraphilic and personality trait differences between serial and single-victim sexual homicide offenders.Methods All 73 single-victim and 13 serial sexual homicide offenders presenting within a cohort of 671 men sentenced for sexual crimes between 1994 and 2005 and serving their sentence in one high-security Canadian prison and who consented to interview were assessed and compared on their offending patterns, personality pathology and paraphilic behaviours.ResultsSerial sexual homicide offenders were more likely than the single offenders to report deviant sexual fantasies, having selected victims with distinctive characteristics, to have targeted strangers, structured premeditation and/or verbal humiliation of their victims during the offences. Personality pathology, defined by at least two Diagnostic and Statistical Manual of Mental Disorders IV criteria for personality disorder, was common in both groups, but the serial offenders were more likely to have narcissistic, schizoid and/or obsessive–compulsive traits; they were also more likely to engage in sexual masochism, partialism, homosexual paedophilia, exhibitionism and/or voyeurism.Implications for practiceSamples of serial sexual homicide offenders will, fortunately, always be small, and it may be that more could be learned to assist in preventing such crimes if data from several studies or centres were pooled. Our findings suggest that an investigation of sexual homicide offenders should include strategies for evaluating premeditation as well as personality and paraphilic characteristics. Crime scene features that should alert investigators should include similar characteristics between victims and particular aspects of body exposure or organisation. Copyright © 2014 John Wiley & Sons, Ltd.
Article
The technique set out in the paper, CHAID, is and offshoot of AID (Automatic Interaction Detection) designed for a categorized dependent variable. Some important modifications which are relevant to standard AID include: built-in significance testing with the consequence of using the most significant predictor (rather than the most explanatory), multi-way splits (in contrast to binary) and a new type of predictor which is especially useful in handling missing information.
Article
Youths who kill, particularly those who kill their parents, have been a matter of increasing concern in the United States in recent years. Empirical analyses of homicide and parricide cases, however, have been limited. The FBI Supplementary Homicide Report (SHR) Data Base for the 10-year period 1977–1986 was utilized to investigate which weapons were used by children in the United States to kill male and female parents and stepparents in single victim, single offender situations. Biological parents were analyzed separately from stepparents. Significant differences were found in the weapons used to kill male and female parents, the weapons used by juveniles and adults when killing parents and stepfathers, and the weapons used by both juveniles and adults when killing male as opposed to female victims. The age and gender differences found in relation to weapons used to kill parents were consistent with a physical strength hypothesis. The relationship found between offender age and weapon used strongly suggests that the number of parricides committed by juveniles could be somewhat curtailed if access to firearms, particularly rifles and shotguns, was severely restricted.
Article
Over the years, researchers and law enforcement professionals have tried to achieve uniformity in the use of the term “serial murder,” but such efforts have rather proved futile. Endeavors aimed at achieving a consensus on the definition of serial murder have potentiated its precarious applications. Though the term continues to remain elusive, ambiguous and amorphous in the literature, recent efforts are suggestive of a trend towards the attainment of a standardized definition of serial murder. However, a review of both current research and legal definitions of serial murder revealed some discrepancies in how the term is defined. Built upon Skrapec’s (2001) definition, the current paper aims to revisit several key definitional issues and to propose a more comprehensive serial murder definition with three key elements: (1) Two or more forensic linked murders with or without a revealed intention of committing additional murder, (2) the murders are committed as discrete event(s) by the same person(s) over a period of time, and (3) where the primary motive is personal gratification.
Article
Research into stranger sexual homicide remains limited as is our understanding of this crime, and this can hinder criminal investigations and the assessment and management of offenders. This study aims to address this. Using data gathered by various British police forces, this study presents a descriptive profile of adult male-on-female stranger sexual homicide in Great Britain in recent years. Along with demographic and occupational data on offenders and victims, the criminal histories of offenders are illustrated as are their offense behaviors. The results are discussed in light of similar research, and future directions and implications are considered.