ArticlePDF Available

Serial homicide: An investigation of behavioral consistency

Authors:

Abstract and Figures

To examine if serial homicide offenders are consistent across their crimes, a model was developed empirically that could be used to distinguish between crimes that were instrumental and those that were expressive. The first known three offences in each series of 69 US serial homicides committed by 23 offenders, were examined and the instrumental and expressive themes determined. Three models were then explored that test for consistency across these themes. The most liberal model was found to classify all of the offences effectively and to reveal complete consistency across the three crimes for all offenders. The implications of these results for offender profiling and further study of serial homicide are discussed. Copyright © 2005 John Wiley & Sons, Ltd.
Content may be subject to copyright.
Serial Homicide: An Investigation of Behavioural
Consistency
C. GABRIELLE SALFATI1* and ALICIA L. BATEMAN2
1Department of Psychology, John Jay College of Criminal Justice,
City University of New York, USA
2Ottowa, Canada
Abstract
To examine if serial homicide offenders are consistent across their crimes, a model was
developed empirically that could be used to distinguish between crimes that were instru-
mental and those that were expressive. The first known three offences in each series of 69
US serial homicides committed by 23 offenders, were examined and the instrumental and
expressive themes determined. Three models were then explored that test for consistency
across these themes. The most liberal model was found to classify all of the offences effec-
tively and to reveal complete consistency across the three crimes for all offenders. The
implications of these results for offender profiling and further study of serial homicide are
discussed. Copyright © 2005 John Wiley & Sons, Ltd.
Key words: serial homicide; linking; consistency; profiling; crime scene
LINKING SERIAL HOMICIDE
Serial homicide may be defined as the unlawful killing of three or more human beings
over a period of time (Bartol, 1995; Holmes & Holmes, 1998). For the offence to be con-
sidered serial homicide, a ‘cooling-off’ period must be present between each offence. The
cooling-off period is a distinguishable time interval between each homicide which can be
days, weeks, months, or even years (Bartol, 1995).
In order to establish that a single offender has committed more than one homicide, the
homicides must be linked together, whether it is by forensic or behavioural methods. While
forensic evidence such as DNA or fingerprints is the preferred means for linking offences,
it is often unavailable to the investigation team (Canter, 1994). This means that it is left
to the police to try to determine if the offences are behaviourally similar to one another.
Behavioural consistency requires an offender’s behaviour to be similar across their
series of offences. In addition for the purposes of profiling, it is also of value for the behav-
iours to be distinct from other offenders who are committing the same type of serial crime.
Copyright © 2005 John Wiley & Sons, Ltd.
Journal of Investigative Psychology and Offender Profiling
J. Investig. Psych. Offender Profil. 2: 121–144 (2005)
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jip.27
* Correspondence to: Dr C. Gabrielle Salfati, Department of Psychology, John Jay College of Criminal Justice,
City University of New York, New York, NY 10019, USA. E-mail: gsalfati@jjay.cuny.edu
Unfortunately, evidence to support the assumption of behavioural consistency is limited
(Bennell & Canter, 2002) and there appear to be no published studies to date on behav-
ioural consistency in serial homicide. The aim of the current study is therefore to deter-
mine if serial homicide offenders are consistent across their series of homicides in their
crime scene behaviours.
Before we can establish how offenders act across their series however, we must first
establish what behaviours occur in general in serial homicide as a whole. Second, we must
establish how offenders differ in how they employ these behaviours and if they can be
classified into different ‘types’. Once we have established such types, we can then test
whether they will consistently act according to type. This approach is very different and
does not rely on looking at the use of individual behaviours, but rather, using ‘pools’ of
behaviours, which all have the same underlying psychological meaning, as first outlined
by Canter (2000) and more recently developed by Salfati and Kucharski (2005). Using
this line of thinking, offenders may then use a number of different individual behaviours
each time they commit their crime, depending on the situation, but they will all be true to
type. By looking at behaviours in this way, it may help factor out external influences on
behaviour choice. An example of this is an offender gagging his victim in one crime, but
binding them in another. The offender is using different individual behaviours, but
both of them have the same underlying psychological meaning—control (see Canter &
Heritage, 1990 for more details on themes of rape). What changes is the offender’s choice,
possibly due to the situation. In the case of gagging, the offender may be controlling the
victim by not allowing her to scream to attract attention, and in the case of binding, the
offender is controlling the victim by not allowing her to escape. In both cases the offender
employs behaviours that are consistent with control.
Developing a classification model of serial homicide behaviours
There are a number of general characteristics of serial homicide that have been highlighted
by the literature. One of the key issues in most of the serial homicide literature is the idea
that the majority of homicides are sexually motivated (Meloy, 2000). In the case of sexual
homicide, the sexual acts may occur at any stage of the homicide: before, during, or after
(Meloy, 2000; Grubin, 1994). Much research has looked at the role that sexual desires and
fulfilments play in serial homicide offences indicating that investigators feel that the sexual
component is of great importance for understanding the activities of serial homicide
(Keppel & Walter, 1999; Ressler et al., 1986). However, Hickey (2002) in a review of
serial homicide cases from 1800 to 1995, which included 337 male and 62 female appre-
hended serial homicide offenders, responsible for 2525–3860 victims provided an account
of basic demographics of this group, as well as characteristics of subgroups. This serial
helped dispel a number of myths regarding serial homicide offenders and their character-
istics which are often cited, such as that serial homicide is mainly sexual in nature, or that
all serial homicide offenders are white, male, insane, very intelligent or travel great dis-
tances to commit their crimes. None of these oversimplifications are supported by Hickey’s
(2002) results.
Other research has found extensive evidence of serial homicide offenders employing a
high level of planning and control in their offence behaviours. Furthermore, being foren-
sically aware, i.e. being aware of factors that will help to avoid detection by the investi-
gation team seems to be a feature that facilitates the possibility of the homicides being
122 C. G. Salfati and A. L. Bateman
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121–144 (2005)
serial (Canter et al., 2004). The literature also suggests that it is likely that serial homi-
cide offenders have previous convictions and are therefore aware of the types of evidence
that will lead to their capture (Bartol, 1995; Soothill et al., 2002). Serial homicide offend-
ers will therefore be likely to perform acts such as destroying evidence, moving the body
and hiding the body (Canter et al., 2004).
Although the above generalised characteristics help us understand serial homicide
offenders as a whole, for investigations it is of more pertinence to understand differences
between offenders so that different behavioural patterns observed at the crime scene can
be ascertained. The next step, as stated previously, is to then understand whether offend-
ers use similar patterns across their series.
In addition, in order for this information to be used in a more practical arena to
help link crime scenes, it is important to identify the exact crime scene behaviours
that are the most useful for linking. Most studies to date have used those variables origi-
nally outlined by the Special Agents of the Federal Bureau of Investigation (FBI) in
the USA (Ressler et al., 1986) who delineated two groups of crime scene variables;
Organised and Disorganised. This approach to modelling serial sexual homicide behav-
iour has provided researchers with the possibility of utilising crime scene behaviours as a
tool for classifying serial homicide activities (Salfati, 2000). Furthermore, it provided the
basis for empirical exploration of how police can use crime scene behaviours to make
inferences about the type of offender on which they should focus their investigation.
However, despite its benefits to serial sexual homicide research and potentially other inves-
tigations, this model has had many criticisms that have led to the need to develop alter-
native classification systems for this type of crime (for more details, see Salfati & Canter,
1999). Most notably, recent research by Canter et al. (2004) found that Organised and Dis-
organised are not two opposing dichotomous categories, but that Organised behaviours
seems to be a function of serial sexual homicide in general and what varies is the par-
ticular sub-variety of Disorganised behaviours they carry out. In addition, they point out
that there are a number of methodological problems with the original FBI reports, mainly
relating to adequate definitions, so although they provide a first step to understanding the
compositions of serial homicide crime scenes, they do not provide a good measure of
behaviours.
Another well quoted typology is that of Holmes and De Burger (1985) who suggest
another classification system of serial homicide offenders consisting of four categories
of killers; Visionary, Hedonistic, and Power/Control. Again the main problem with this
model of serial homicide behaviours is that the boundaries between each of the different
types of killers described are vague, and there is a great deal of overlap of similar behav-
iours across the different types. There is also the problem of lack of empirical support for
the different categories of serial homicide offered by this typology (Ainsworth, 2001).
Indeed, Canter and Wentink (2004) conducted an empirical evaluation of the model and
raised several concerns, again, most focusing on the difficulty, based on the lack of defi-
nition of behaviours, of defining precisely the conditions under which an offence or
offender will be assigned to one type or another.
The third widely cited model of serial homicide, is that of Keppel and Walter (1999)
who specifically looked at four types of serial sexual homicide: Power-Assertive Rape-
Murder; Power-Reassurance Rape-Murder; Anger-Retaliatory Rape-Murder; and Anger-
Excitation Rape-Murder. Although the Rape-Murder model does provide information
regarding how each type can be identified by crime scene behaviours and makes an attempt
Serial homicide 123
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121–144 (2005)
to explain theoretically how the motives and actions will develop from rape to sexual
homicide, through the series of homicides, it does not provide thorough empirical support
for the model’s assumption that the behaviours within each type will be correlated with
each other and therefore that they will actually occur in the same type of Rape-Murder.
Moreover, while Keppel and Walter (1999) did attempt to provide the investigators with
the frequencies with which the four types of Rape-Murder occur in a prison sample, like
other models, they also neglected to provide details of the criteria used to assign each
offence to each type of Rape-Murder.
The studies of Canter et al. (2004) and Canter and Wentink (2004) reveal the value and
importance of developing reliable and valid classification systems for considering Serial
Homicide. These complement recent work on single homicide such as Salfati (2000),
Salfati and Heratsis (2001) and Santilla et al. (2001) by providing models which reveal
different subsets of behaviours, and so provide more methodologically and empirically
based classification systems of crime scenes. For this reason, this body of work will be
used as a basis for the current paper to test out some of the same ideas in a sample of
serial homicide, to investigate whether we can understand this subgroup of homicides as
part of homicide in general.
One of the most notable models for crime scene classification as it pertains to profil-
ing, has drawn upon Instrumental and Expressive aggression as thematic distinction
between crime scene behaviours in single homicide offences (Block et al., 2000; Salfati,
2003; Salfati & Haratsis, 2001; Salfati, 2000; Salfati & Canter, 1999; Santilla et al., 2001).
The instrumental pole consists of behaviours relating to the offender attacking in order to
gain something from the offence such as money or sex (Block et al., 2000; Salfati, 2000).
Here, the offender views the victim not as a person, but rather as a vehicle through which
they can gratify some need (Canter, 1994; Salfati, 2003). The primary goal of the expres-
sive homicide offender is to harm the victim (Block et al., 2000; Salfati, 2000). This theme
of aggression is often provoked through an emotional, interpersonal confrontation and is
most likely to occur against someone the victim knows intimately. Here, the victim as the
person is the target and the behaviours centre on an extreme physical attack against the
individual (Block et al., 2000; Salfati, 2003).
This study will attempt to explore some of these similarities and differences by identi-
fying if it is possible to utilise the methods of classifying homicide behaviour as instru-
mental and expressive as illustrated in single homicide research to classify serial homicide
behaviour. If the model holds, it can be used first, as a basis for exploring how offenders
can be differentiated by theme, and second, if they are consistent in using the same theme
across their series of crimes.
AIMS OF THE STUDY
(1) The first aim of the study is to determine if the aggressive themes of instrumentality
and expressiveness are evident in serial homicide offences, as is the are in single homi-
cide offences.
(2) The second aim is to establish if offences can be clearly assigned to the identified
behavioural themes.
(3) The third aim is to determine if serial homicide offenders are consistent across their
series of homicides in their crime scene theme, and thus to investigate whether the
model can be used to link cases of serial homicide.
124 C. G. Salfati and A. L. Bateman
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121–144 (2005)
THE DATA
The data was obtained from police case files of the Homicide Investigation and Tracking
System (HITS) database in Seattle, Washington, USA. All cases selected from the HITS
database were classified as solved serial homicide cases. Cases were classified as solved
if the offender had been arrested for the offence. The definition of serial homicide used
during the data collection was that the offender needed to have committed a minimum of
three or more homicides, at different times and locations. This conforms to the definition
most commonly found in serial homicide literature (Holmes & Holmes, 1998; Lundrigan
& Canter, 2001). Although it may be that offenders will change their behaviour due to
learning and experience as they move along in a series, the first three in the series were
used due to the fact that in investigations, it is important to be able to establish a link early
on in the series. In addition, although we may assume that offenders change their behav-
iour along a series, it is important that we understand how this behaviour changes, so that
we may understand the nature of what is consistent across a series, and what is not, and
when these changes occur.
The variables
From the selection of serial homicide behaviours obtained from HITS, 94 variables relat-
ing to serial homicide offences were selected for the current study. Of the 94 variables,
61 were crime scene behaviours and 33 were offender characteristics. To be included, the
behaviours were required to be those that could be objectively observed at the crime scene
by investigators, and which could less easily be prone to subjective interpretation. These
behaviours were chosen in order to make comparisons with previous models easier, and
to make the results from the study directly applicable to investigators, and to lend reli-
ability and validity to the model. All behaviours and characteristics with the exception of
age, body disposal and death sites were coded as dichotomous variables being labelled
either 1 referring to the behaviour being present during the offence or 0 referring to the
absence of the behaviour. It should however be clarified that due to the sometimes unclear
evidence in files, the absence of the behaviour could indicate both that the offender did
not perform the behaviour or that the information regarding that behaviour was missing
from the data set. The consequent statistical analysis used (Smallest Space Analysis, see
Results section for details) takes this into account in its computations.
Sample demographics
The age of the offenders during all of their offences ranged from 17 to 53 years with a
mean age of 33 and a standard deviation of 9. It is important to note that each offender’s
age at the time of each of the three homicides has been entered as the offender’s age could
change across the time period of their series of homicides.
The age of the offenders at the time of their first homicide in the series ranged from 17
to 53 years with a mean age of 32 and a standard deviation of 8. Of interest is that only
one offender in the sample committed their first and second homicide at different ages
and this offender committed the second homicide 8 years after he committed the first
homicide in the series. This raises questions about the length of any ‘cooling-off’ period
between their crimes, and how long this period may be, which has been unclear in the lit-
erature to date. Of further note therefore, is the fact that in the current sample, only five
Serial homicide 125
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121–144 (2005)
offenders committed their third homicide at a different age than their second homicide.
Two offenders committed the third homicide one year after the second, one committed
their third homicide 2 years after the second, one committed the third homicide 5 years
after the second and one committed the third homicide in the series 27 years after the
second homicide. It would seem therefore that, in this sample at least, there is a tendency
for any series of homicides to occur over a relatively short period of time, often less than
12 months. This may carry implications for consistency that need to be considered in the
future.
Other offender demographics included race, where 14 (61%) were white and 8 (35%)
were black. All of the offenders were male. Fifteen of the offenders were employed at the
time of the homicides (65%). In terms of marital status, 16 (70%) offenders were single
at the time of their offence, four (17.4%) were married and one was divorced (4.3%).
Eleven of the offenders were noted to engage in criminal activity (48%), six were known
as alcohol abusers (26%) and seven were drug users/abusers (30%). Five of the offenders
had non-offender status at the time of the offence (22%), nine offenders were on parole
(39%) and two were out on bail (9%).
Of the 69 victims in the sample, there was a large age distribution, ranging from under
4 years of age to 82 years of age with a mean age of 28 and a standard deviation of 16.
Thirty-six (52%) of the victims were female and 33 (48%) were male. Forty-nine
victims were white (71%), 12 victims were black (17%), five victims were Hispanic (7%),
one was American Indian (1%) and one was Asian (1%). Of the victims, 15 had a small
build (22%), 32 had a medium build (46%) and 17 had a large build (25%). Eleven of the
victims were noted to have outstanding physical features or something to attract attention
(16%). Thirty-three (48%) had an occupation, 11 (16%) were students and five (7%) were
transients. Furthermore 22 (32%) of the victims were prostitutes. Five of the victims
were noted to engage in criminal activity (7%), seven were alcohol abusers (10%) and 10
were drug users/abusers (15%). Of the victims, 10 lived alone (15%).
RESULTS
Frequency analysis
The first step to developing a model of serial homicide offending behaviours was to analyse
the frequencies of the 61 crime scene behaviours identified in the data. It is important to
examine these frequencies before choosing the behaviours to use in the analysis of the
themes of serial homicide behaviours. The importance relates to the discussion by Canter
(2000) Salfati (2003), and later by Canter et al. (2004) regarding problems with the model
developed by Ressler et al. (1986). Without considering the behaviours that occur in the
majority of serial homicides, higher frequency behaviours may be inappropriately used for
differentiating between offenders. Salfati (2003) and Canter et al. (2004) both refer to
behaviours with frequency values above 50% to be higher frequency behaviours. In the
current sample therefore, there are four behaviours that could be considered to be high-
frequency, non-differentiating behaviours. These are: the victim being white (71%), the
body recovery site being outdoors (65%), sexual assault (61%) and the victim being female
(52%) (See Table 1).
These findings are consistent with the general trends of serial homicide aforementioned
in that victims tend to be females and that serial homicide is more likely to be intra-racial
126 C. G. Salfati and A. L. Bateman
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121–144 (2005)
Serial homicide 127
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121–144 (2005)
Table 1. Frequency of occurrence of crime scene behaviors
Serial homicide behaviours Percentage Number of cases
Victim’s race was white 71.0 49
Body recovery site was outdoor 65.2 45
Sexual assault 60.9 42
Victim was female 52.2 36
Victim was male 47.8 33
Small personal items/trophies taken from victim 47.8 33
Victim had an occupation 47.8 33
Victim’s build was medium 46.4 32
Unconcerned with disposal of body 44.9 31
Victim found partially dressed 44.9 31
Weapon brought to the scene by offender 43.5 30
Sexual assault was vaginal 40.6 28
Items of victim’s clothing missing from recovery site 40.6 28
Body hidden or placed to prevent discovery 34.8 24
Victim was 15 years of age or less 34.8 24
Victim was a prostitute 31.9 22
Victim was bound 30.4 21
Body moved from death site to recovery site 27.5 19
Victim found nude 27.5 19
Stabbing or cutting weapon used 26.1 18
Victim’s build was large 24.6 17
Body recovery site was in or about victim’s residence 23.2 16
Victim was between 15 and 25 years of age 23.2 16
Victim was 35 years of age or older 21.7 15
Victim’s build was small 21.7 15
Body was staged or posed 21.7 15
Semen found in body cavities of victim 21.7 15
Victim was between 25 and 35 years of age 20.3 14
Victim’s race was black 17.4 12
Weapon used was firearm 17.4 12
Weapon used was ligature 17.4 12
Offender destroyed/attempted to destroy evidence 17.4 12
Body displayed openly or to ensure discovery 17.4 12
Restraining devices brought by offender to scene 15.9 11
Victim had outstanding features to attract attention 15.9 11
Victim was a student 15.9 11
Victim was a drug abuser 14.5 10
Victim lived alone 14.5 10
Weapon used was the offender ’s hands or feet 14.5 10
Sexual assault was anal 14.5 10
Victim was tortured 11.6 8
Weapon used was bludgeon 10.1 7
Victim was an alcohol abuser 10.1 7
Victim was Hispanic 7.2 5
Victim was a transient 7.2 5
Victim engaged in criminal activity 7.2 5
Offender disabled phone, security, other 7.2 5
Victim was gagged 7.2 5
Sexual assault was oral sex by offender 7.2 5
Unusual act, ritual, or thing performed 5.8 4
Victim’s body parts were removed 5.8 4
Victim performed oral sex on offender 5.8 4
Evidence of ejaculation on victim’s body 5.8 4
Victim was blindfolded 4.3 3
Foreign objects inserted into openings of victim’s body 4.3 3
Bite marks found on victim’s body 2.9 2
Evidence of ejaculation at scene other than on body 2.9 2
Writing or carving on the body 1.4 1
Victim’s race was American Indian 1.4 1
Victim’s race was Asian 1.4 1
and therefore, the finding that the majority of serial homicide offenders in this sample
were white corresponds to the finding that the majority of victims were white.
Furthermore, the suggestion by previous authors that the majority of serial homicides
tend to be sexual in nature seems to be partly supported by the current findings, although
results show that there were still a significant proportion that did not display any sexual
behaviour during the crime. The finding that most victims are found outdoors may be
related to the offender attempting to distance themselves from the victim and the offence
by leaving the body in a neutral, non-identifiable location. This would be consistent with
the findings from Canter et al. (2004) where planning and delay detection behaviours are
core and perhaps necessary behaviours for serial homicides to continue and thus become
serial.
The next set of behaviours to be examined were those that fell between the 30 and 50%
range of frequencies (See Table 1). In this group many of the victim variables were found;
the victim was male (48%,), the victim had an occupation (48%), the victim’s build was
medium (46%), the victim was less than 15 years of age (35%) and the victim was a pros-
titute (32%). Again, these are consistent with previous findings that suggested that the
largest categories of serial homicide victims are prostitutes and children. Of interest here
is how similar the percentage of male victims (48%) is to female victims (52%). It appears
that the sex of the victim is not predominately female for serial homicide offences and
thus suggests that the sex of the victim may not be a good single indicator of whether a
homicide is like to be part of a series.
In this frequency range it can again be seen that planning and delaying detection behav-
iours are relatively important for serial homicide to occur as the behaviours of bringing a
weapon to the scene (44%) and hiding the body to prevent discovery (35%) are included
here. Behaviours involving controlling the victim (body bound) were also found in this
category (30%).
In the study by Hodge cited in Canter (2000) it was suggested that stealing from the
victim may be a sign of power or possession over the victim and therefore is a core behav-
iour of serial homicide offences. The current study found a relatively high element of theft
in this sample. Stealing small personal items or trophies from the victim was found to
occur in 48% of cases and stealing items of clothing from the victim occurred in 41%.
The reasons for the inconsistency between the Hodge study and the present study may be
that the Hodge study had both these theft categories combined into one. Of notice also, is
that these theft behaviours are not indicative of the actual act of theft being of importance
to the offender as these behaviours are not useful for monetary gain. Rather, they are per-
sonal items of the victims that may help the offender to remember the victim and/or the
offence.
Also in this group of behaviours, there is an element of sexual activity where the
offender committed vaginal sexual assault against the victim in 41% of cases. The victim
was also left partially dressed in 45% of cases, however it cannot be assumed that this
was a sexual variable without further evidence from the offenders’ statements themselves.
The next frequency range included behaviours between 10 and 30% (See Table 1). In
this frequency range the behaviours involved in the actual killing were found, i.e. all
behaviours involving the weapon used for the homicide are located here. Using a knife or
stabbing weapon occurred in 26% of cases, using a firearm or using a ligature occurred
in 17% of cases, the offender using their own hands or feet was found in 15% of cases
and using a bludgeon or club was found to occur in 10% of cases. Also found in this cat-
egory were overtly violent behaviours such as torturing the victim (12%).
128 C. G. Salfati and A. L. Bateman
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121–144 (2005)
More specific body disposal behaviours were also found to occur in this frequency
range. These behaviours include moving the body (28%), leaving the victim nude (28%),
posing the body (22%) and leaving the body openly displayed to ensure detection (17%).
The rarer victim characteristics were also found to occur in this range, such as the victim
being black (17%), the victim having unique or outstanding features (16%) and the victim
being a drug or alcohol abuser (15% and 10% respectively). These behaviours coupled
with the specific disposal behaviours and torturing of the victim may be hypothesised to
be related to the motivations or fantasies of the offender, as suggested by previous authors.
However, this would need to be verified and backed up with interviews with offenders in
order to test this.
The final frequency range is the behaviours that occurred in less than 10% of cases (See
Table 1). Less ‘stereotypically common’ sexual acts such as the victim performing oral
sex on the offender and the offender performing oral sex on the victim (6% and 7% respec-
tively) were found to occur in this range, indicating some degree of involvement between
the victim and the offender. Furthermore, ritualistic (6%) or abnormal acts such as biting
the victim (3%) and removing body parts (6%) were found to occur here. This again illus-
trates some level of involvement between the offender and the victim.
Delay detection behaviours such as disabling security, gagging the victim and blind-
folding the victim were also found to occur in this range (7%, 7% and 4% respectively).
Serial homicide frequencies compared with single homicide frequencies
It is important to examine how the behaviours found in the frequency ranges for serial
homicide compare with those found in single homicide to see if the similarities and dif-
ferences between the two can be relevant for behavioural modelling and investigations of
such crimes. For instance, if it is found that some behaviours occur in almost all serial
homicides, but not in single homicides, then the evidence of this behaviour occurring based
on the crime scene may signal to the investigation team that they are dealing with a serial
homicide case rather than a single homicide offence. If serial homicide and single homi-
cide are found to be highly similar in the behaviours that offenders perform during the
course of the event, then investigators may be able to utilise the same investigation pro-
cedures to apprehend either type of offender by relying on general behavioural models
relating to homicide. Furthermore, research that examines single homicide behaviour may
then directly relate to serial homicide behaviours and vice versa, which ultimately will
increase the overall empirical body of knowledge.
Unfortunately, there exist no studies to date that gives an in-depth analysis of the actual
behaviours that occur in single homicides in the US. However, in a recently published
study, Salfati (2003) conducted an in-depth analysis of the frequencies of crime scene
behaviours for single homicide in the UK. This study will therefore be used as an ap-
proximation for comparing single homicides to serial homicides as this study also relates
directly to the study examining expressive and instrumental behavioural themes in single
homicide, which will provide the basis for the thematic analysis in the current study.
Although the Salfati (2003) is a study of UK homicides, the findings from Salfati and
Haratsis (2001) on Greek homicides and Santilla et al. (2001) on Finnish homicides illus-
trate the cross-cultural similarities in homicide behaviours and therefore suggest that it is
appropriate to compare culturally different samples. All three studies imply that there is
an underlying structure to homicide, perhaps psychologically, that exists irrelevant of the
culture in which the offence is occurring. However, due to the fact that no studies have
Serial homicide 129
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121–144 (2005)
looked into US homicides, necessary caution needs to be taken in extrapolating from the
current analysis, and results should only be used as a primary indicator of some of the
factors that may be useful in comparing the two. Future studies need to be employed in
order to compare single and serial homicides within US cases.
The current and previous findings of serial homicides show that such an event involves
a high degree of planning and control and that these types of behaviours are higher fre-
quency behaviours. Salfati (2003) however found that the high frequency behaviours for
single homicides were suggestive of an impulsive, unplanned, emotional interaction
between the victim and the offender and that this may be related to previous findings that
homicide is often prompted by an emotional argument. There therefore seems to be a con-
trast in the role that emotions play in single and serial homicides. While emotions may
play a role in serial homicide, they may be more related to the psychological gratification
that the offender may gain from the event taking place and are therefore more active after
the offence rather than a precipitating factor. Once again, a more detailed investigation
into offender motivations may help in highlighting the key psychological issues relating
to the high frequency behaviours.
The 30–50% range of frequencies of serial homicide behaviours found in the current
study is related more to the goals of the offender through the event rather than the spe-
cific act of killing. In this range, behaviours relating to delaying detection and controlling
the victim were evident as well as the behaviours of gaining personal property of the victim
and engaging in sexual activities with the victim. In other words, the extra actions involved
in the event seem to be more important to the offender than the act of killing itself. In the
30–50% frequency range in the study of single homicide by Salfati (2003), the behaviours
largely revolved around the actual act of killing the victim, i.e. the methods employed to
ensure that the homicide took place in a particular way. In the current study however, the
behaviours surrounding the actual act of killing occurred in the lower frequency range of
10–30%.
Salfati (2003) found that the planning and organised behaviours in single homicide
were found to be less frequent and occurred in the 10–30% range. Furthermore, stealing
property from the victim and sexual acts were found to occur in this range. This is in stark
contrast to serial homicide behaviours where it was found that all of these behaviours,
planning, forensic awareness, theft of personal property, and sexual assault, occurred in
the higher frequency ranges.
Here again, there is evidence of a different focus in single homicides compared with
serial homicides. Perhaps due to the impulsive nature of the single homicide attack, behav-
iours revolving around planning and avoiding detection are not as relevant as it is only
after the event that the offender comes to realise their importance. Furthermore, the emo-
tions that the single homicide offender is going through after such an attack may prevent
them from being fully aware of all the behaviours necessary to avoid apprehension. On
the other hand, the serial homicide offender appears to have ulterior motives other than
just the act of killing and these motives press them to want to commit the act multiple
times and hence, knowledge of delaying detection is pertinent to the act occurring.
Based on the examination of the low frequency behaviours from the current study, it
may be possible to hypothesise that the low frequency behaviours, i.e. those that occur in
less than 10% of cases, are also related to the specific agenda of the individual offender.
These include behaviours such as ritualistic acts being performed, foreign object used for
sexual assault and bite marks found on the victim’s body. Salfati (2003) also found that
behaviours occurring in less than 10% of cases may be more related to the specific agenda
130 C. G. Salfati and A. L. Bateman
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121–144 (2005)
of the offender rather than relating to the actual offence of homicide. Furthermore, they
may be associated with the offender attempting to distance themselves psychologically
from the event. Also consistent with the findings from the single homicide study by Salfati
(2003) is the result that the serial offender was also found to attempt to distance them-
selves psychologically in a similar fashion to the single homicide offender through anal
penetration of the victim and blindfolding the victim.
Based on the initial comparisons here, it appears then that in serial homicide, the ulte-
rior motives such as theft, sex and the behaviours that will allow the offender to continue
to kill are the more frequent and perhaps more important behaviours. Alternatively in
single homicide, the most frequent behaviours revolve around the act of killing and here
the ulterior motivators such as theft, sex, fantasy fulfilment as well as apprehension pre-
vention measures are less common and more specific to a certain subgroup of offenders.
The fantasy fulfilment behaviours also appear more frequent in serial homicides than in
single homicides.
Future research needs to further examine differences between single and serial homi-
cides in order to identify how investigators can better distinguish between the two types
when investigating such a crime. Of interest in the current study is whether or not the dif-
ferences in frequencies of offending behaviour for serial homicide and single homicide
will affect the instrumental and expressive aggressive themes predicted for serial homi-
cide behaviours. Perhaps it will be found that such themes cannot be found or that the
themes will contain different types of behaviours than those found in single homicide
research. Alternatively, the themes may be present with the same types of behaviours, but
the focal behaviours of each theme may be found to be different from those for single
homicides. It may be predicted then that if the instrumental and expressive aggressive
themes are found for serial homicide offences, the focal aspects of the behaviours will be
related to the extra acts involved in the offences such as sexual acts, stealing and power
and control type behaviours rather than behaviours that ensure the act of killing occurs
and in a specific manner.
AIM 1: CLASSIFYING SERIAL HOMICIDE CRIME SCENE BEHAVIOURS
In order to develop a model of serial homicide offending behaviour, the 61 crime scene
behaviours were examined to assess if serial homicide offending behaviours could be clas-
sified into instrumental or expressive aggressive themes.
Examining behavioural themes in serial homicide
In order to to develop a new model of serial homicide offending behaviours that is based
on the theme of aggression utilised by the offender during the commission of the crime,
a Smallest Space Analysis (SSA) of a subset of 25 serial homicide crime scene behaviours
was conducted to investigate if the themes of expressive and instrumental aggressive styles
of offending could be identified in serial homicide. There were two main reasons for
employing an SSA for this study. The first is that many previous studies examining the
themes of instrumentality and expressivity in homicide have utilised this method of analy-
sis (Salfati & Haratsis, 2001; Salfati, 2000; Salfati & Canter, 1999; Santilla et al., 2001).
The second reason is that these studies have demonstrated that SSA is effective at reliably
analysing such noisy data, and illustrating themes of behaviour if they exist in the data.
Serial homicide 131
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121–144 (2005)
SSA is a non-metric multidimensional scaling procedure. It produces a representation
of behaviours in a Cartesian-space (Shye, Elizur & Hoffman, 1994). The three-
dimensional result was used for the current examination (see Figure 1) as it provided
a clear reflection of the relationships between the behaviours. The SSA represents the
relationship between behaviours, represented by points in the geometrical space, such that
the greater the correlation or degree of co-occurrence between two behaviours across all
the cases examined, the closer together the points will be located within the space. It there-
fore facilitates the examination of each behaviour in relation to all other behaviours under
investigation.
The organisation of the points in the geometric space shows the structure of the data
and hence, the themes predicted should be visible if they are evident in the serial homi-
cide behaviours being assessed. The regional hypothesis for SSA assumes that those
behaviours that make up a common theme will be located in the same region of the SSA
plot allowing the theme to be easily identifiable. Therefore, if the themes of instrumental
and expressive aggression do exist in serial homicide offending behaviour, they should be
identifiable by examining how the crime scene behaviours fall on the SSA plot.
The coefficient of alienation gives a value for how well the spatial representation of the
relationship of the behaviours fits the correlations between the behaviours identified by
the SSA analysis. The smaller the alienation coefficient, the better the fit, i.e. the better
the representation of the behaviours in the geometric space is to the actual relationships
132 C. G. Salfati and A. L. Bateman
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121–144 (2005)
destroy
ritual
vicles15
v15to25
v25to35
vicold35
vicmale
bodresid
bodopen
bodhidd
bodstage
bodmove
bodbound
gagged
vicnude
partdres
trophies
sexassau tortured
firearm
cutting
bludgeon
ligature
handfeet
weapbrou
Expressive
Instrumental
Expressive
-- Frequency boundary
Instrumental
Figure 1. Smallest Space Analysis of the expressiveness and instrumentality of serial homicide crime scene
behaviours.
in the data. In general, a coefcient below 0.2 is considered, for this type of noisy data, a
reasonable t.
The behaviours used for this analysis were selected based on previous literature regard-
ing what behaviours would represent the predicted themes. It is generally agreed upon
that the themes of crime scene behaviours can best be identied by examining the low
frequency behaviours, i.e. those below 50%; therefore the high frequency behaviours of
victim was female, body left outside and victim was white were omitted from the SSA
analysis. The behaviour of sexual assault was included due to the importance of this behav-
iour in serial homicide offending as aforementioned, and also due to the fact that it was
very close to the 50% cut-off point.
As illustrated by Figure 1 and Table 2, the aggressive themes of instrumentality and
expressiveness are evident in serial homicide crime scene behaviours and are distinct from
each other. The upper half of the spatial plot displays the instrumental behaviours. Here
there is support for the idea that the instrumental act is primarily about something other
than the homicide itself as suggested by Salfati (2000). The idea that the victim was utilised
as a vehicle through which the offender could satisfy an ulterior aim such as sex or ful-
lling a ritualistic fantasy is supported by the ndings for behaviours of sexual assault,
leaving the victim nude or partially dressed, ritualistic acts performed, staging and posing
the body and displaying openly to ensure discovery. Also consistent with the results from
Salfati (2000) is the nding for the body being left in the victims home and manually
killing the victim.
Expressive homicide behaviours are found in the lower half of the spatial plot. The idea
that the expressive homicide is an emotional attack (Salfati, 2000) is supported by the
ndings of the behaviours of torturing the victim, using a cutting instrument to kill the
victim and taking personal items or trophies from the victim. There is also evidence for
the offender trying to separate themselves from the victim and the offence (Salfati, 2000)
by destroying evidence and hiding the body. This theme shows some evidence of plan-
ning and controlling the situation by the offender bringing the weapon to the scene and
binding the body. Interestingly, almost all of the victim characteristics fell within the
expressive theme which may be indicative of the offender targeting a specic victim,
Serial homicide 133
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121144 (2005)
Table 2. Instrumental and expressive behaviours in serial homicide
Instrumental behaviours Expressive behaviours
Sexual assault Victim was male
Victim found partially dressed Victim was 15 years of age or less
Body moved from discovery site Weapon brought to scene by the offender
Victim found nude Body hidden
Body recovery site was in or about victims Offender destroyed/attempted to destroy
residence evidence
Victim was 35 years of age or older Victim was bound
Body was staged or posed Stabbing or cutting weapon used
Weapon used was ligature Victim was between 15 and 25 years of age
Weapon used was offenders hands or feet Victim was between 25 and 35 years of age
Weapon used was a bludgeon Weapon used was rearm
Body displayed openly or to ensure discovery Small personal items/trophies taken from victim
Victim was gagged Victim was tortured
Unusual act, ritual or thing performed
which again supports the idea that the victim is important in the denition of these types
of offenders (see Salfati, 2000).
The frequency bands of serial homicide behaviours can be further super-imposed onto
the SSA plot to provide a clearer understanding of the behaviours comprising both the
instrumental and expressive themes of serial homicide (See Figure 1 and Table 1). The
high frequency behaviour of sexual assault, as predicted, sit around the centre of the plot.
Although the suggestion is that this behaviour, having a high frequency, should not be
used to distinguish between offending behaviours, it is still found to co-occur with the
instrumental theme as previously shown by Salfati (2000) and not the expressive theme
and therefore should be included as a behaviour under consideration when investigators
are examining a crime scene.
As previously mentioned, the behaviours falling under the 3050% range for serial
homicide were often related to controlling the situation and facilitating the offence. Behav-
iours included controlling the victim, and destroying evidence, moving the body, and
hiding the body. According to Figure 1, these types of behaviours are important for those
committing expressive serial homicide acts but not for those committing instrumental acts.
This is consistent with the ndings by Salfati (2000).
The behaviours found to fall under the 1030% range were related to how the offender
chose to kill the victim, the amount of violence used in the attack, how the offender dis-
posed of the body and specic victim characteristics. These combined were deemed to be
related to the motivations, goals and fantasies of the offender. Again, the ndings from
the current study are consistent with previous literature on the instrumental and expres-
sive themes of homicide offences. It seems that for serial homicide offenders as well as
single homicide offenders, the amount of violence is greater for the expressive theme, as
evidenced by the ndings of torture and stabbing or cutting occurring in this theme. This
nding supports the idea that the goal of expressive violence is to harm the victim.
According to Figure 1, the predicted goal of instrumental homicideto use the victim
to satisfy the offenders needs or desiresis supported for serial homicide offences. Under
this frequency band for the instrumental theme, behaviours relating to the offender ful-
lling their ritualistic fantasy were found such as staging and openly displaying the body.
Also found were sexual behaviours indicative of the offender utilising the victim to satisfy
sexual urges. Additionally under this frequency band was the nding for instrumental serial
homicide offenders to use their hands or feet, use a bludgeon or club or to use a ligature
to kill the victim. These too may be related to the offender coming unprepared for
such an attack as the use of a bludgeon or a ligature was found to be far apart on the plot
from bringing a weapon and therefore these methods of killing are not likely to co-occur
with the offender planning the attack ahead of time and coming prepared with the weapon
(in only four out of the 30 cases where a weapon was brought, did the offender use a
bludgeon/club).
Only two behaviours fell within the less than 10% frequency band and these were
both found to occur within the instrumental theme; these were ritualistic behaviours and
gagging the victim. The ritualistic behaviour is consistent with this theme as it represents
the ulterior goal of the offender for instigating the attack (see Salfati, 2000).
Summary
The rst aim of this study was successfully achieved and the themes of instrumental and
expressive aggression were found to be effective for use in understanding the behaviours
134 C. G. Salfati and A. L. Bateman
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121144 (2005)
involved in serial homicide offences. Furthermore, these themes are distinct from
each other indicating that they may be utilised to classify serial homicide offending
behaviours. In essence, what goals the offender wishes to achieve appears to drive what
types of offending behaviours the offender chooses to enact. The hypothesis here is
that those offenders who commit instrumental aggressive homicides are driven by
personal gain and utilise the victim as an object through which they can obtain their
desired objective, whereas offenders committing expressive aggressive homicides are
driven by the desire to harm the victim and see the victim as a person to interact with
(Canter,1994).
As the original idea for utilising instrumental and expressive aggression to explain homi-
cide was examined in single homicide offences, it is important to examine the similarities
and differences between the single homicide ndings and the serial homicide ndings. In
general, the behavioural indicators for instrumental and expressive homicides (i.e. per-
sonal gain and physical harm of the victim, respectively) were found to be similar for both
single and serial homicides. It was suggested though that perhaps due to the differences
in the types of behaviours that were found to occur in each of the frequency ranges for
single and serial homicides that the core behaviours of each of the two themes may also
differ between single and serial homicides. This difference was found to occur in both of
the aggressive themes.
In the expressive theme of homicide, serial homicides were shown to have core
behaviours related to controlling the victim and avoiding detection and the lower frequency
behaviours were related to killing the victim. In the single homicide studies however,
the core behaviours are more related to the actual act of killing the victim and delay
detection behaviours were found to occur in few cases. The higher frequency behaviours
in serial instrumental homicides also differed from single instrumental homicides in
that in serial homicides it was the sexual acts, or the potential gains of the offender
that were higher in frequency whereas in single homicides, the higher frequency instru-
mental behaviours were related to the act of killing the victim in an unplanned manner.
Future research is still needed to conrm and fully understand the structural differences
between the make up of the instrumental and expressive themes for single and serial
homicides.
The current model, unlike previous models is not a typology; it is a thematic model of
serial homicide offending behaviour. As previously mentioned, this means that each
offence is not expected to t solely into one category; rather the offence behaviours may
contain aspects of both of the themes but should primarily fall under only one main theme.
The ability to classify each offence into either of the themes will be addressed in the next
section of this paper.
An additional criticism of previous models was that the boundaries between the cat-
egories were vague. In the current study there is little overlap between the two themes.
According to the SSA in the current study, all behaviours relating to utilising the victim
for personal gain occurred in the instrumental theme and all behaviours related to the direct
goal of harming the victim and avoiding detection, with the exception of moving the body,
occurred in the expressive theme. The current model therefore has a more dened bounda-
ry between categories than previous models.
Overall, the current study has developed a model of serial homicide offending behav-
iour that addresses and overcomes many of the problems of previous models. This fulls
the requirements of providing a model which can be used to investigate thematic consis-
tency in the behaviours used in an offenders series.
Serial homicide 135
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121144 (2005)
AIM 2: ASSIGNING CRIMES TO THEMES; AND AIM 3: BEHAVIOURAL
CONSISTENCY IN SERIAL HOMICIDE
A major absence in all of the previous models of serial homicide to date has been the con-
sideration of the key question of whether or not the offenders consistently perform the
same type of offending behaviours across their series of homicides. The second aim of the
current study was therefore to examine if serial homicide offenders are consistent in their
employment of either instrumental aggression or expressive aggression throughout their
series of homicides.
Background
Although there are no studies to date in the linking and serial homicide literature that has
discussed the issue of behavioural consistency, there are a number of other studies that
shed light on the evidence for this. Funder and Colvin (1991) studied the idea of behav-
ioural consistency using university students. The participants were involved in three
situations, two similar to each other but each took place several weeks apart and a third
situation that was different than the previous two but was presented to participants within
hours of the second situation. The results indicated that although presented with different
situations, participants brought their own set of behaviours into each event and remained
relatively consistent with their behaviours across the three situations. It appeared that the
participants used their set of behaviours to adapt to each situation.
Shoda, Mischel and Wright (1994) examined consistency in behaviours in a different
way. They used the psychological features of an event or encounter to observe consisten-
cies in individual differences in behaviours. Individual differences in behaviours are
referred to as those behaviours that an individual exhibits that are different from behav-
iours that are normative in such situations. Furthermore, they describe psychological fea-
tures as active features of situations(p. 685). To provide an example, in their experiment,
the participants were at a camp (the event/situation) and they experienced being teased,
threatened, praised, etc. (the psychological features). The results of the study showed that
the participants exhibited a higher degree of cross-situational consistency in events that
were psychologically similar than in events that contained less psychologically similar
features.
Shoda, LeeTiernan and Mischel (2002) looked at stability in how individuals arrive at
the behaviours they exhibit. The authors identied a cognitive-affective processing system
(CAPS) that guides how one processes the features of a situation to produce the resulting
behaviours. The authors used a computer simulation and discovered that the CAPS
network used by an individual is the stable aspect directing behaviours (Shoda et al., 2002).
This means, therefore, that when one is presented with a particular set of situational fea-
tures, one will react with a particular set of behaviours. Furthermore, if the individual is
presented with a similar set of features at another time, they should react in a similar
manner to the initial contact with the situation (Shoda et al., 2002).
Although these studies were not investigations of consistency amongst criminal activ-
ities, they can be applied to the current discussion. Despite the notion that each crime the
offender partakes in will be somewhat different in terms of the situational features, the
hypothesis would be that the offender should still exhibit some consistency in their behav-
iours across all the events. Moreover, one would expect to nd this consistency in the
136 C. G. Salfati and A. L. Bateman
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121144 (2005)
behaviours that an offender exhibits across the series because, for the offender, the crimes
would contain psychologically similar features.
Researchers are beginning to provide evidence that supports the idea that there is
consistency in a criminals behaviours across a series of crimes (Bennell & Canter,
2002; Canter, 1995, 2000; Grubin, Kelly & Brunsdon, 2001; Lundrigan & Canter, 2001).
Bennell and Canter (2002) looked at behaviours across serial burglary offenders that
would allow investigators to link the crimes together. They determined that the geo-
graphical distance between offences is a stable and consistent behaviour. The authors
further noted that how applicable behaviours are to linking crimes together and how con-
sistently the behaviours are performed may be related to how situation-dependent the
behaviours in question are. In their study, those behaviours that were more dependent on
what the situation was were less likely to be consistent across the series of offences by
the same offender. Extrapolating this research into serial homicide research, it may be
predicted that not all the behaviours in a series of homicides by an offender will be con-
sistent due to the fact that the situation the offenders are encountering will be different for
each offence.
Grubin, Kelly and Ayis (1997) addressed the issue of studying consistency in serial
sexual assault by using a sample of British sexual assault offenders. They developed a
model consisting of four behavioural domains each consisting of four types of behaviours.
They used this model to classify sexual assault offenders based on what type of behav-
iour they performed in each of the four domains. The researchers then assessed how con-
sistently the offenders were performing these types of behaviours across the series of
sexual offences that the offenders were committing. In 2001, Grubin, Kelly and Brunsdon
continued to assess the model of serial sexual assault behaviours and they replicated their
initial study using data on serial sexual assault from the Canadian ViCLAS database. The
analyses looked at single-domain consistency as well as multi-domain consistency. The
overall results of both of these studies were that offenders do show consistency in some
of their offence behaviours. The most stable behaviour across all offendersseries was the
methods of controlling the victim that were employed. Furthermore, the results indicated
that the consistencies evident were not due to chance.
The overall nding of the aforementioned research is that there is the support for con-
sistency in behaviours performed by serial homicide offenders. It may be possible then to
nd consistency in the aggressive theme, either instrumental or expressive, that the serial
homicide offender employs throughout their series of offences.
While evidence has been presented to support the idea that certain behaviours will be
consistent across a series of offences, one must keep in mind that not all behaviours will
be consistent. Abrief discussion regarding the potential reasons for inconsistency in behav-
iours is therefore required.
The rst major reason for non-consistent behaviours is related to the discussion by
Salfati (2000, 2003) regarding how situation dependent the behaviours are, and how inu-
enced they are by the victimoffender interaction. Salfati explains how homicide is essen-
tially an interaction between the victim and the offender and that the events that occur
during this interaction may dictate how the homicide is eventually played out. Said another
way, the behaviours of the victim may determine the behaviours of the offender. For
example, for offenders where control is highly important, the victim may be highly resist-
ant and therefore restraints, which are not normally used by the offender, must be placed
on the victim in order for the homicide to proceed. The behavioural manifestation of this
restraint however depends again on the situation. If the victim screams, gagging may be
Serial homicide 137
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121144 (2005)
used, whereas if the victim attempts to escape, binding may be used (e.g. see Canter &
Heritage, 1990).
Grubin et al. (2001) suggest a second major reason for inconsistency, which is that the
aims of the offence may change during the series of crimes. The offender may have accom-
plished a set of goals in a previous offence and will therefore move on to achieve new
objectives. Grubin et al. (2001) indicate that this may be related to the offenders evolv-
ing fantasy, which they are fullling by committing the crime. Prentky et al. (1989) provide
evidence to support the importance of fantasy in serial sexual homicide but do not indi-
cate if the fantasy changes or remains consistent.
A third major reason for non-consistency in behaviours is the learning that takes place
during every offence an offender is involved in (Canter, 1994; Grubin et al., 2001). It has
been found that most homicide offenders have previous criminal convictions for offences
other than the intended homicide and/or violence (Bartol, 1995; Soothill et al., 2002). With
each crime, an offender learns what they have to do in order to succeed at not being caught.
While incarcerated, offenders will be in contact with and learn effective strategies from
other criminals. Furthermore, offenders may become more aware of what techniques
the police use to catch them via the media coverage of both their offences and other
criminalsoffences and therefore become more adept at not leaving the relevant clues. With
time and practice however, being forensically aware may become a consistent behaviour.
Anal potential reason for inconsistency is related to the mental state of the offender.
The offenders mental state may change due to the ingestion of drugs and alcohol through-
out their series of offences and this may make the offender less competent to commit the
act without detection and will alter their offence behaviours. Additionally the offender may
become over-condent in their abilities and may try new, perhaps riskier tactics in order
to prove to him or herself that they will not get caught. This, while changing the behav-
iours that the offender is performing, may make the offender more careless in their actions
and will lead to apprehension.
The current study
The same sample of 69 homicides committed by 23 offenders (rst known three offences
per offender) was used for the analysis. In order to establish consistency, each offence was
rst classied into instrumental or expressive homicides. In order to do this, the propor-
tion (%) of behaviours in each offence that are either instrumental or expressive were
calculated and then compared in order to identify if one theme dominates over the other.
This was done using the 25 behaviours utilised in the SSA analysis to establish the
existence of the two themes. Three strategies were employed to determine which aggres-
sive theme dominates each offence (see Salfati, 2000 for a discussion of this). These vary
in how conservative is the criterion for assigning a crime to a theme. Table 3 displays the
results of classifying each offence using the three different strategies. Strategy 1 is com-
pleted by looking to see which of the two themes displays a higher proportion of behav-
iours and then assigning the offence as being predominately that aggressive theme. For
example, Offender 1 Case 1 has 42% of their behaviours falling under the expressive theme
and 31% of their behaviours falling under the instrumental theme. Therefore, this homi-
cide case will be deemed to be an expressive homicide. Strategy 2 involves examining the
proportions of instrumental and expressive behaviours for each offence and to only clas-
sify those offences where one theme contains a proportion of behaviours that is 1.5 times
higher than the other theme. Finally, Strategy 3 requires one theme to have twice as many
138 C. G. Salfati and A. L. Bateman
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121144 (2005)
behaviours performed by the offender than the alternative theme. Depending on how strin-
gent a category is used, different results will be obtained. This again highlights the impor-
tance of establishing a consistent methodology of classication (see Salfati 2000 for a
discussion of this issue).
Strategy 1
According to the ndings displayed in Table 3, using Strategy 1 allows for all of the
offences to be classied as being either instrumental or expressive. The results indicate
that 67% of all of the offences were classied as expressive homicides (46/69 cases) and
that 33% of the offences were classied as instrumental homicides (23/69 cases). These
results also indicate that serial homicide, like single homicide, in general is predominately
an expressive crime.
Now that the offences have been classied, the series of three offences by each offender
must be examined to determine consistency (see Table 4). By using Strategy 1, 13 of the
23 serial homicide offenders (56%) consistently employed the same aggressive theme
when committing all of their three homicides. Of those 13 offenders who consistently per-
formed the same theme of aggression across all three offences, 10 of the offenders (77%)
consistently performed expressive homicides and three of the offenders (23%) consistently
performed instrumental homicides. While this nding does support the notion that serial
homicide offenders are consistent in the aggressive theme they employ during their
offence, the percentage of consistent offenders is only slightly higher than half signifying
only moderate support for consistency in serial homicide offending behaviour.
There is a problem with this classication strategy in that it does not discriminate well
between the two themes. For example, Offender 5 Case 2 has two behaviours (17%) falling
under expressive and two behaviours (15%) that fall under instrumental aggression. This
Serial homicide 139
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121144 (2005)
Table 3. Thematic classication using three different strategies
Classication Total % of cases % of cases that could be % of cases that could be
strategy classiable classied as Expressive classied as Instrumental
1 100.00 66.67 33.33
2 63.77 68.18 31.82
3 40.58 75.00 25.00
Table 4. Detailed examination of cases where all three cases followed same crime scene theme
Percentage theme allocation of
those cases where all three
Percentage of cases where all cases belonged to the same
three cases belonged to the dominant theme
Classication strategy same dominant theme Expressive Instrumental
1 56.22 76.92 23.08
2 85.70 83.33 16.67
3 100.00 66.67 33.33
proportion is so close that an investigator could easily classify the offence as being either
of the two themes or even as a blend of the two themes. More specically, there does not
appear to be a signicant difference between the two themes that would warrant a reliable
classication of the offence as being either instrumental or expressive. It may therefore
be necessary to employ a more stringent classication strategy to examine if the model
does in fact reliably classify offences into respective themes, if serial homicide is pre-
dominately an expressive crime, and if there is support for consistency in serial homicide
offending behaviour.
Strategy 2
As previously mentioned, for each homicide to be classied using Strategy 2, one aggres-
sive theme needs to have 1.5 times as many behaviours performed by the offender com-
pared with the alternative theme. This is a more stringent classication strategy that may
be more appropriate for examining classication and consistency as this strategy has more
discriminating power than Strategy 1. The results from classifying serial homicide offend-
ing behaviour using Strategy 2 show that 44 of the 69 cases (64%) are classiable as either
instrumental or expressive (see Table 3). This indicates that the model is effective at clas-
sifying serial homicide offending behaviour as it accounts for a substantial proportion of
cases. Of the cases classied, 30 (68%) were expressive and 14 (32%) were instrumental.
This conrms the ndings from the classication using Strategy 1 that serial homicide
seems to be more of an expressive crime than an instrumental one.
When using Strategy 2 to classify offences, seven offenders had all three cases in their
series that were classiable (30% of all cases in the sample) and hence of use for the con-
sistency analysis (see Table 4). Examining the 21 cases by the seven offenders revealed a
high degree of consistency across the series of homicides by the offenders. The results
indicated that the series of three homicides by six of the seven offenders (86%) whose
cases were applicable for the analysis were consistent across their series in the theme of
aggression utilised during their homicides. Of the six offenders who maintained consist-
ency across their series, ve performed expressive homicides and one performed instru-
mental homicides (83% and 17% respectively).
The results from the analysis using Strategy 2 to classify serial homicide offending
behaviours as either instrumental or expressive indicate that this model is effective at
explaining serial homicide behaviour. It also shows that most serial homicide offenders
commit expressive homicides and that serial homicide offenders are highly consistent in
the aggressive theme that they enact during their offence.
Strategy 3
Strategy 3 was the most conservative method of classication and involved the require-
ment of one theme having twice the percentage of behaviours of the alternative theme.
This strategy has been utilised in previous research that classies homicide behaviour
using the instrumental/expressive aggression model (see Salfati & Haratsis, 2001; Salfati,
2000). Using this strategy, 28 of the 69 cases (41%) in the sample can be classied as
being either instrumental or expressive (see Table 3). While this percentage is lower than
the percentage classiable by Strategies 1 and 2, there is still a substantial proportion of
offences that are classiable with this strategy, conrming the notion that the instrumen-
tal and expressive model of aggression is effective for explaining and examining serial
homicide offending behaviour. Of the cases classied, 21 (75%) were deemed to be expres-
140 C. G. Salfati and A. L. Bateman
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121144 (2005)
sive and seven (25%) were deemed to be instrumental which is consistent with the nding
of the previous two strategies.
In terms of consistency, only three of the offenders however had all three of their
offences classied in Strategy 3 and therefore only these three sets of offences were applic-
able to the analysis (see Table 4). Of these three series of homicides, two offenders had
expressive series of homicides and one offender had an instrumental series of homicides.
Here again, consistent with the ndings from Strategy 1 and 2, there is support for con-
sistency in serial homicide offendersbehaviours.
Summary
Three strategies for classifying crimes were examined for their effectiveness. Strategy 1
used the least stringent criteria and was found not to discriminate enough to be effective
in practice. This strategy creates considerable ambiguity and therefore is like to be of little
utility. Strategy 3 was very stringent and while it provided a high level of discrimination
and made for large and more obvious boundaries between the two themes, it accounts for
a considerably smaller proportion of the offences brought to investigatorsattention and
therefore, like Strategy 1, may not be effective in practice.
Strategy 2 may therefore, currently, be the most suitable to be employed for classify-
ing serial homicide offending behaviour into either of the two aggressive themes. This
strategy was found to reduce ambiguity when classifying offence behaviours by allowing
for a discernable boundary between the two themes to be developed that would enable the
investigator to identify which of the two themes dominates the offence in question.
Furthermore, the results from this study illustrated that this strategy allowed for a sub-
stantial proportion of offences to be classied, making it appropriate for use by both inves-
tigators and researchers.
The general trend observed from classifying the offences is that the majority of serial
homicides were deemed to be expressive. This is surprising considering the suggestions
in the literature that serial homicide is a very different type of homicide offence. It is
however consistent with the ndings from studies on single homicides where the major-
ity of cases were classied as expressive (Salfati & Canter, 1999; Salfati, 2000; Salfati &
Haratsis, 2001). This may suggest that a major objective for both single and serial homi-
cide offenders is to harm the victim rather than for personal gain by the offender at the
victims expense. This may indicate that serial homicide is highly interpersonal in nature,
similar to single homicides. The potential for the meaning of any interpersonal transac-
tion to the offender needs to be explored in more depth as previous literature has found
that serial homicide typically occurs between strangers or only brief acquaintances (see
Bartol, 1995 and Holmes & Holmes, 1998).
The aim of this second set of analyses was to examine the one aspect of what Canter
(2000) calls the proling equationin order to assess if serial homicide offenders are con-
sistent in the aggressive theme that they employ across their series of homicides. Although
the numbers in the study are small, the initial indication is that serial offenders are con-
sistent in the aggressive theme they enact when they commit each of their homicides in
the series. This indicates that the behavioural theme that the serial offender during one
homicide uses, continues in general throughout their homicides.
In terms of linking two or more homicides together and to a common offender, inves-
tigators may be able to utilise the current instrumental/expressive model of behaviour to
classify each offence, then determine if the homicides in question are of the same aggres-
Serial homicide 141
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121144 (2005)
sive theme. If this is the case, it would provide some support for the assumption that the
homicides are linked together. Because this model was not found to be able to classify all
homicides examined to a dominant theme, and a dominant theme throughout the series,
the model should, until further more in-depth studies have been done, be used only as
a guideline to research, and be used with caution, and certainly not stand-alone, in any
investigation.
CONCLUSIONS
The main aim of the current study was to examine whether offenders behave thematically
consistently across their series. A method of classifying serial homicide offending behav-
iour into either of the two aggressive themes was determined and once each offence was
classied, it was found that serial homicide offenders are relatively consistent in the
aggressive theme that they perform during their series of homicides.
LIMITATIONS OF THE CURRENT STUDY AND DIRECTIONS
FOR FUTURE RESEARCH
There are two important limitations in the current study that need to be discussed in order
for recommendations for future research to be made. The rst limitation is the very small
sample size examined in the current study. The sample only contained 23 offenders who
were charged with three or more homicides. Increasing the number of offenders who
commit and are charged with three or more homicides in a series will more accurately
identify if serial homicide offenders are consistent across their series by allowing for more
series of homicides to be compared. Furthermore, examining series of homicides with
more offences in the series (for example ve homicides in the series) may provide a more
denitive look at consistency in the series by leaving a smaller possibility of the ndings
for consistency to be due to chance rather than by the choice of behaviours by the offender.
Increasing the number of offenders and increasing the number of offences in a series may
thus provide additional support for consistency in the aggressive theme employed by the
serial homicide offender.
As has previously been highlighted, researchers also need to conduct interviews with
convicted serial homicide offenders to understand the offence from the offenders point
of view. Previous studies have made attempts at using interviews for classication,
however, using behavioural and empirical models at the rst stage of the analysis provides
a much more reliable basis for model building. Behavioural models, such as the ones
included in the current study, allows researchers to identify what role more qualitative
analysis can provide in rening the models at hand. As such, and unlike previous research,
they can form the basis for more structured interviews, so that theory driven questions can
be identied and explored in an interview setting. The current study demonstrated that the
goal of the homicide can be distinguished based on the crime scene behaviours. An inter-
view could go further to help to identify what made the offender want to inict the victim
with such a high level of violence or why the offender used the victim for sexual grati-
cation. Furthermore, interviewing the offender can help researchers to appreciate why
offenders with particular characteristics are driven to satisfy specic goals through the act
of serial homicide.
142 C. G. Salfati and A. L. Bateman
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121144 (2005)
Finally, now that it has been established that a model of serial homicide offending
behaviour can be created, which is effective at classifying serial homicide cases and which
provides support for consistency, it is important to explore the issue of the types of offender
characteristics that can be predicted from knowledge of the instrumental or expressive
aggressive theme employed by the offender during their offence. In other words, it is essen-
tial to examine the characteristics of the offender associated with each behavioural type,
and how different types of offenders engage in either thematically consistent or inconsist-
ent series.
ACKNOWLEDGEMENTS
The authors would like to sincerely thank the Homicide Investigation and Tracking System
(HITS) team in Seattle, Washington, USA, who helpfully provided the data for this paper,
and the anonymous reviewers and Professor Canter for their guidance on the revision of
this paper.
REFERENCES
Ainsworth, P. B. (2001). Offender profiling and crime analysis. Cullompton, Devon: Willan
Publishing.
Bartol, C. R. (1995). Criminal behavior: A psychosocial approach. New Jersey: Prentice Hall.
Bennell, C., & Canter, D. (2002). Linking commercial burglaries by modus operandi: tests using
regression and ROC analysis. Science and Justice 42, 153164.
Block, C. R., Devitt, C. O., Donoghue, E. R., Dames, R. J., & Block, R. L. (2000). In P. H. Blackman,
V. C. Leggett, & J. P. Jarvis (Eds.), Diversity of homicide: Proceedings of the 2000 Homicide
Research Working Group (pp. 0211). Washington, DC: Federal Bureau of Investigation.
Canter, D. V. (2000). Offender proling and criminal differentiation. Legal and Criminological
Psychology,5, 2346.
Canter, D. (1995). Psychology of offender proling. In R. Bull, & D. Carson (Eds.), Handbook of
psychology in legal contexts. Chichester, UK: John Wiley & Sons.
Canter, D. V. (1994). Criminal shadows. London: Harper Collins.
Canter, D. V., Alison, L. J., Alison, E., & Wentink, N. (2004). The organized/disorganized typology
of serial murder: Myth or model? Psychology, Public Policy, & Law, 10, 293320.
Canter, D., & Heritage, R. (1990). A multivariate model of sexual offence behaviour: Developments
in offender proling. Journal of Forensic Psychiatry, 1, 185212.
Canter, D. V., & Wentink, N. (2004). An empirical test of Holmes and Holmess serial murder typol-
ogy. Criminal Justice & Behavior, 31, 489515.
Funder, D. C., & Colvin, C. R. (1991). Explorations in behavioral consistency: Properties of person,
situations, and behaviors. Journal of Personality and Social Psychology, 60, 773794.
Grubin, D. (1994). Sexual murder. British Journal of Psychiatry,165, 624629.
Grubin, D., Kelly, P., & Ayis, S. (1997). Linking serious sexual assaults. London: Home Ofce.
Grubin, D., Kelly, P., & Brunsdon, C. (2001). Linking serious sexual assaults through behaviour.
London: Home Ofce.
Hickey, E. (2002). Serial killers and their victims (3rd ed.). California: Wadsworth.
Holmes, R. M., & Holmes, S. T. (1998). Serial murder (2nd ed.). California: Sage Publications.
Holmes, R. M., & De Burger, J. E. (1985). Proles in terror: The serial murderer. Federal Proba-
tion, 43, 2934.
Keppel, R. D., & Walter, R. (1999). Proling killers: A revised classication model for understand-
ing sexual murder. International Journal of Offender Therapy and Comparative Criminology, 43,
417437.
Lundrigan, S., & Canter, D. (2001). A multivariate analysis of serial murderersdisposal site loca-
tion choice. Journal of Environmental Psychology, 21, 423432.
Serial homicide 143
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121144 (2005)
Meloy, J. R. (2000). The nature and dynamics of sexual homicide: An integrative review. Aggres-
sion and Violent Behavior,5, 122.
Prentky, R. A., Burgess, A. W., Rokous, F., Hartman, C., Ressler, R., & Douglas, J. (1989). Pre-
sumptive role of fantasy in serial sexual homicide. American Journal of Psychiatry, 146, 887891.
Ressler, R. K., Burgess, A. W., Douglas, J. E., Hartman, C. R., & DAgostino, R. B. (1986). Serial
killers and their victims: Identifying patterns through crime scene analysis. Journal of Interper-
sonal Violence, 1, 288308.
Salfati, C. G. (2003). Offender interaction with victims in homicide: A multidimensional analysis
of frequencies in crime scene behaviours. Journal of Interpersonal Violence, 18, 490512.
Salfati, C. G. (2000). The nature of expressiveness and instrumentality in homicide. Homicide
Studies,4, 265293.
Salfati, C. G., & Canter, D. V. (1999). Differentiating stranger murders: Proling offender charac-
teristics from behavioral styles. Behavioral Sciences and the Law, 17, 391406.
Salfati, C. G., & Haratsis, E. (2001). Greek homicide: A behavioral examination of offender crime-
scene actions. Homicide Studies, 5, 335362.
Salfati, C. G., and Kucharski, T. (2005). The Psychology of criminal conduct. In J. Trevino, &
S. Guarino (Eds.), The common subject of crime: A multi-disciplinary approach. Anderson
Publishing.
Santilla, P., Canter, D., Elfgren, T., & Hakkanen, H. (2001). The structure of crime-scene actions in
Finnish homicides. Homicide Studies, 5, 363387.
Shoda, Y., LeeTiernan, S., & Mischel, W. (2002). Personality as a dynamical system: Emergence
of stability and distinctiveness from intra- and interpersonal interactions. Personality and Social
Psychology Review, 6, 316325.
Shoda, Y., Mischel, W., & Wright, J. C. (1994). Intraindividual stability in the organization and
patterning of behaviour incorporating psychological situations in the idiographic analysis of
personality. Journal of Personality and Social Psychology, 67, 674687.
Shye, S., Elizur, D., & Hoffman, M. (1994). Introduction to facet theory: Content design and intrin-
sic data analysis in behavioural research. Newbury Park, CA: Sage.
Soothill, K., Francis, B., Ackerley, E., & Fligelstone, R. (2002). Murder and serious sexual assault:
What criminal histories can reveal about future serious offending. London: Home Ofce.
144 C. G. Salfati and A. L. Bateman
Copyright © 2005 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 2: 121144 (2005)
... Early studies in IP research on this issue outlined how hybrid crime scenes are related to a lower level of consistency in series. Salfati and Bateman (2005) published the first study in the field aiming to provide a baseline of knowledge on how consistent offenders are over their series, reporting that 70 percent of series included hybrids. This highlighted the need to untangle the nature of hybrid crime scenes, to understand whether their classification was due to models that required further refining or whether they were due to the presence of a specific subtype of crime scene that was a hybrid. ...
Chapter
Investigative Psychology (IP) as a field of study started in 1990. The domain of IP covers all aspects of psychology that are relevant to the conduct of criminal (and civil) investigations. Its focus is on providing an empirical basis to the ways in which criminal activities are examined and understood in order for the detection of crime to be effective and legal proceedings to be appropriate. As such IP is concerned with psychological input to the full range of issues that relate to the investigation and prosecution of crime. This chapter provides a full review of the area of Investigative psychology.
Article
Purpose Knife-enabled crime in England and Wales has increased by 7% in the year ending December 2023. Such increases in incidents are cause for concern due to the potential for significant injury and loss of life. This study aims to propose a model of differentiation of offending across 70 cases of violent knife crime (VKC) committed in England and Wales between 2015 and 2020 to inform preventative strategies. Design/methodology/approach Incident data was collected from online news articles and offender, victim and offence characteristics were recorded. Characteristics were subject to a non-metric multi-dimensional scaling procedure, smallest space analysis (SSA) to differentiate behavioural styles. Findings Regional interpretation of the SSA identified three distinct themes (Intimate Partner Violence, Gang-Influenced and Reactive Violence) that related to distinct styles of offending. Classification according to identified offence theme was possible for 69% of cases. Findings are discussed with reference to Social Identity Approach and Criminal Narrative Theory. Potential future research is discussed with recognition of the need to further differentiate offending behaviour in cases of reactive violence. Practical implications The theorical and practical implications are discussed with consideration of how the findings may inform preventative strategies as part of a public health approach. Originality/value While SSA has been used to differentiate behavioural styles across several types of crime, to the best of the authors’ knowledge, this is the first instance in which the method has been used to differentiate behavioural styles across cases of VKC.
Chapter
This chapter on criminal profiling and femicide provides an analysis of the methodologies and challenges involved in profiling perpetrators of gender-based homicide. It begins by defining femicide and distinguishing it from other forms of homicide, emphasising its roots in gender inequality and violence against women. The chapter then delves into the role of criminal profiling as a tool for law enforcement, outlining its historical development, theoretical foundations, and practical applications in solving femicide cases. Central to the discussion is the profiling process, which includes crime scene analysis, victimology, and behavioural patterns of offenders. By integrating psychological theories with forensic science, the chapter demonstrates how a multidisciplinary approach enhances the accuracy and effectiveness of criminal profiling in femicide cases. The conclusion underscores the need for ongoing research and training to refine profiling techniques and ensure they are used ethically and effectively.
Article
Full-text available
p style="text-align: justify;">The purpose of our study was the features of criminal behavior of the accused in the commission of particularly serious crimes in the context of the formation functions of programming and control. The sample consisted of 59 men aged 18—60 years, of those accused of committing particularly serious crimes aimed at a forensic psychiatric examination, the average age was 33.7 years. The methods of neuropsychological examination and psychological analysis of criminal cases were used. Syndrome of defeat of the basal divisions of the frontal lobes, prefrontal syndrome, syndrome of defeat of the medial divisions of the frontal lobes, Postfrontal (premotor) syndrome is most often seen among persons accused of particularly serious crimes. The criminal behavior of the accused in the Commission of particularly serious crimes was characterized by uncritical damage to the basal parts of the frontal lobes. The impulsivity is the main characteristic of the criminal behavior of the accused in especially serious crimes with the defeat of the prefrontal frontal lobe (prefrontal syndrome). Subjects with the defeat of the kinetic (dynamic) factor differed greater rigidity of criminal behavior. The behavior of those accused of committing particularly serious crimes was passive (energy-saving) in violation of the energy factor in the case of damage to the medial parts of the frontal lobes. The obtained results can be used to solve the issues of drawing a portrait of an unknown criminal, as well as in the course of correctional work with persons prone to repeat illegal behavior.</p
Article
Full-text available
In nomothetic analyses, the cross-situational consistency of individual differences in social behavior, assessed in vivo in a camp setting, depended on the similarity in the psychological features of situations. As predicted by the social–cognitive theory of personality, idiographic analyses revealed that individuals were characterized by stable profiles of if…then…, situation–behavior relationships that formed “behavioral signatures” of personality (e.g., he aggresses when warned by adults but complies when threatened by peers. Thus, the intraindividual organization of behavior variation across situations was enduring but discriminatively patterned, visible as distinctive profiles of situation–behavior relationships. Implications were examined for an idiographic reconceptualization of personality coherence and its behavioral expressions in relation to the psychological ingredients of situations.
Article
Full-text available
This article presents the results of an empirical test of Holmes and Holmes’s serial murder classification scheme. Crime scene evidence from 100 U.S. serial murders, each the third in a distinct series, was content analyzed. The co-occurrence of content categories derived from the crime scene material was submitted to smallest space analysis. The features characteristic of the category of “power or control” killings were found to be typical of the sample as a whole, occurring in more than 50% of cases, and thus did not form a distinct type. Limited support was found for aspects of the lust, thrill, and mission styles of killing, but this support drew attention to differences in the ways victims were dealt with, through mutilation, restraints, or ransacking their property rather than the motivations implicitly inferred in Holmes and Holmes’s typology. The current results are presented as an empirical basis for the classification of serial killings on which more detailed models can be built.
Article
Full-text available
The structure of crime-scene actions in Finnish homicides was analyzed using the theoretical distinction between instrumental and expressive aggression. Thirty-eight crime-scene actions were derived from the files of single-offender, single-victim Finnish homicides (N = 502). A multidimensional scaling procedure was used to investigate their structure. An expressive core indicating an emotional and impulsive attack without any indication of planning or post-mortem dealing with the body was found. Over 40% of the cases had no additional variables present. A distinction between expressive and instrumental types of aggression was observable in the rest of the crime-scene actions. The instrumental actions were further divided into subthemes indicating either a sexual or a resource-acquirement theme. Likewise, the expressive actions were divided into three subthemes, where the main emphasis had to do with the use of firearms, hiding and moving the body, and removing parts of the body.
Article
Homicide grows out of a transaction between individuals. This transaction is a product of the individuals and their relationship. More may therefore be understood about different "styles" of homicide by examining how an offender acts toward a victim during the crime. The analysis of the actual behaviors used by offenders at 247 single offender-single victim homicide crime scenes indicated a pattern of frequencies that follow a continuum from where the offender reacts in an impulsive way toward the emotions engendered through the conflicted interpersonal relationship with the victim, to where the offender interacts with the victim much more at a removed level, both physically and emotionally. Behaviors can thus be seen to follow a thematic pattern that can be related to underlying psychological principles, which closely relate to how an offender interacts with the victim at the crime scene.
Article
The study of crime scene profiling efforts elicits two important patterns of sexual murders: organized and disorganized. These law enforcement categories have been derived from evidence and patterns of evidence at the site of sexual murders. The study then explores victim information and its relationship to the two categories. In particular, we explored victim response to the offender in terms of no resistance and active resistance to the assault. We found that regardless of type of resistance, active or passive, and category of offender, death ensued. When we examined nine victims who survived, the category of offender was not the predictor, rather, “chance happenings” preserved life.