Content uploaded by Joakim Sturup
Author content
All content in this area was uploaded by Joakim Sturup on Jul 27, 2018
Content may be subject to copyright.
RESEARCH ARTICLE
Comparing serial homicides to single homicides: A
study of prevalence, offender, and offence
characteristics in Sweden
Joakim Sturup
1,2,3
1
Swedish Police Authority, Region Stockholm,
Sweden
2
National Board of Forensic Medicine,
Department of Forensic Psychiatry, Sweden
3
Department of Clinical Neuroscience,
Karolinska Institutet, Sweden
Correspondence
National Board of Forensic Medicine,
Department of Forensic Psychiatry, PO‐Box
4044, 141 04, Huddinge, Sweden.
Email: joakim.sturup@ki.se
Abstract
Multiple and serial homicides constitute around 1.6% of all homi-
cides but have unusually tragic outcomes. The aims of this study
were to report the prevalence serial‐homicide offences and the
characteristics of serial‐homicide offenders in a non‐North
American context. The study included all convicted serial‐homicide
offenders in Sweden from 1973 to 2012 (n= 25), as well as a pop-
ulation‐based control sample of single‐homicide offenders collected
during 2007 and 2009 (n= 201). Similar to U.S. findings, approxi-
mately 1.6% of homicides could be attributed to serial‐homicide
offenders. Serial offenders were more often diagnosed with person-
ality disorders and autism‐spectrum disorders compared with single
offenders; and serial offences more often involved victims who
were female and unacquainted with the offender, as well as plan-
ning, sexual motives, and asphyxia as method of violence. One third
of serial offences involved actions aimed at disguising the manner of
death; such actions were considerably less common in single‐
homicide offences. Few offenders displayed the traditional profile
of a serial‐homicide offender; however, one third displayed sexual
motives, a rate roughly 10 times higher than that in the single‐
homicide group. The study concludes that serial‐homicide offenders
should be subjected to forensic‐psychiatric examinations and that
research involving cross‐contextual differences and similarities in
serial‐homicide offenders is needed.
KEYWORDS
serial homicide, serial killer, serial murder, sexual murder
--------------------------------- --- -- -- --- -- --- -- -- --- -- --- -- -- --- -- --- -- --- -- -- --- -- --- -- -- --- -- --- --
Copyright © 2018 John Wiley & Sons, Ltd.
Received: 11 November 2016 Revised: 11 January 2018 Accepted: 7 February 2018
DOI: 10.1002/jip.1500
J Investig Psychol Offender Profil. 2018;1–15. wileyonlinelibrary.com/journal/jip 1
1|BACKGROUND
1.1 |Introduction
Multiple and serial homicides are unusual events with unusually tragic outcomes (FBI, 2008; Fox & Levin, 1998;
Hickey, 2013). Häkkänen (2008) has pointed out that most research on multiple homicides has been influenced by
investigations of serial killers in the United States and that multiple‐and serial‐homicide offenders in Europe may dif-
fer from their American counterparts. Other scholars have also emphasised that more research from non‐American
settings is needed (Harbort & Mokros, 2001; Jenkins, 1988; Salfati, Horning, Sorochinski, & Labuschagne, 2015;
Sturup & Lindqvist, 2014). Research has shown that offenders who kill two or more victims are different from
offenders who kill only one victim (Fox & Levin, 1998; Harbort & Mokros, 2001; Kraemer, Lord, & Heilbrun, 2004;
Pakkanen, Zappalà, Bosco, Berti, & Santtila, 2015; Trojan & Salfati, 2011). There has been a lively debate concerning
the classification and definition of multiple‐homicide offenders (DeLisi & Scherer, 2006; Reid, 2017a), especially with
regard to serial killers (Adjorlolo & Chan, 2014; Fridel & Fox, 2017). A serial‐homicide offender will, in this study, be
operationalised as an individual who commits two or more homicides during a period of more than 24 hr (in accor-
dance with Adjorlolo & Chan, 2014; FBI, 2008; Pakkanen et al., 2015; Santtila et al., 2008; Yaksic, 2015). Research
has reported increasing trends of serial‐homicide offenders both in the United States and internationally until the
1980s and 1990s, respectively, and after that decreasing trends (Aamodt, 2015; Fox & Levin, 2014; Hickey, 2013;
Quinet, 2011; Schlesinger, 2001).
There have been several attempts to classify serial‐homicide offenders into typologies (Canter & Youngs, 2009;
Miller, 2014). Classification approaches to serial killers are a first step of understanding the underlying mechanism
of serial killers and are therefore of importance (Canter & Wentink, 2004). The most well‐known typologies are the
organised/disorganised typology (Ressler, Burgess, & Douglas, 1988), the Holmes and Holmes's typology (Holmes &
Holmes, 1998), Keppel and Walter's system (Keppel & Walter, 1999), and Kelleher and Kelleher's typology for female
serial murderers (Kelleher & Kelleher, 1998). However, replication studies of the different typologies are discouraging,
resulting in conclusions such as: “… throw considerable doubt of this dichotomy …(Canter, Alison, Alison, & Wentink,
2004, p. 313),”“…there is only mixed support for either model (Sewall, Krupp, & Lalumière, 2013, p. 82),”“…could not
be used to distinguish between offenses or support the proposed types (Canter & Wentink, 2004, p. 512),”and “…
potentially invalid (Bennell, Bloomfield, Emeno, & Musolino, 2013, p. 5).”The reason why typologies are not
supported in replication studies might be that it is not possible to classify serial‐homicide offenders or that there is
a need for more knowledge, and research before reliable and valid typologies can be proposed. For a closer descrip-
tion of serial‐homicide offender aetiology and typologies, please see Hickey (2013), Fox and Levin (2014), FBI (2008),
and Wilson (2007).
1.2 |Prevalence and characteristics
The share of all homicides that can be attributed to serial offenders has been proposed to range from less than 1%
(FBI, 2008; Kiger, 1990) to almost 2% (Jenkins, 1988). The Radford University Serial Killer Database (2015) includes
10 serial killers in Sweden; however, the database does not indicate during which years the offenders were active.
To date, there is no peer‐reviewed study on serial‐homicide offenders from Sweden.
In a study from England, Jenkins (1988) reported that 124 victims were killed by serial murderers between 1940
and 1985. The three most common types of homicide committed by serial offenders were murders of female prosti-
tutes, murders of children, and killings for profit. Another British study reported that between 1960 and 2006, 19
serial murderers killed 326 victims, at an average of seven victims per year (Wilson, 2007). In an influential European
study of 61 offenders who had committed at least three premediated murders, Harbort and Mokros (2001) reported
that the clearance rate in German serial homicides between 1945 and 1995 was about 83%. In comparison with 750
single‐homicide offenders, serial‐homicide offenders had more often killed victims with whom they were
2STURUP
unacquainted (80% vs. 21%) and had more often displayed sexual motives (41% vs. 3%); further, serial offenders
tended to be of above‐average intelligence, a finding that had been reported previously (Miller, 2014; Ressler et al.,
1988). However, there are also reports suggesting that serial killers rather are of average intelligence (Aamodt
et al., 2007). Of the 61 serial‐homicide offenders identified in Germany during the period, 22 were sexual serial
murderers (36%), and they were considerably younger, 22.5 years at mean, compared with the other serial murderers
(30.9 years; Harbort & Mokros, 2001).
Kraemer et al. (2004) have reported that among 157 U.S. serial killers and a reference group from the FBI's
Uniform Crime Report system, serial offenders more often killed out of sexual motivation (55%) and without any pre-
ceding contact with the victim, whereas single offenders more often killed out of anger and in the context of offences
preceded by interpersonal conflicts. In addition, serial offences more often involved planning, strangulation, and
dumping of the body. Pakkanen et al. (2015) have reported that the mean age of 23 male Italian serial‐homicide
offenders was 37.9 years; offenders had killed 116 victims (i.e., on average, about five victims each), and in 30% of
the offences, there was evidence of a sexual motive. After comparing the offences, the authors proposed a model
for differentiating serial homicides from single homicides. According to the model, which was able to effectively dis-
criminate between serial and single cases, serial homicides are characterised by seven crime scene features: Victims
are females and prostitutes; weapons are brought to the crime scene by the offender; the victim exhibits wounds
to the hands and has suffered multiple blunt injuries; the body has been found outdoors; and there is evidence of
forensic awareness on the part of the offender.
1.3 |Serial homicide and mental disorders
Studies on forensic psychiatry and mental disorders among serial‐homicide offenders are rare (Schlesinger, 2001).
There are a few studies that mainly focus on sexual serial killers. For example, Harbort and Mokros (2001) reported
that serial‐homicide offenders had more often been diagnosed with personality disorders compared with the
single‐homicide offenders (89% vs. 18%). Hill, Habermann, Berner, and Briker (2007) have reported that among 36
German sexual serial killers, 86% had at least one personality disorder; 72% had at least one paraphilia (the most com-
mon of which was sexual sadism, 68%); 58% had a substance‐abuse disorder; and 36% met the criteria of psychopathy
(a score of 25 or higher on the Hare Psychopathy Checklist—Revised). Similar findings have been described in non‐
European settings as well. Stone (2001) reported that of 89 male sexual serial killers, 87% met the Hare criteria for
psychopathy according to the Psychopathy Checklist‐Revised (PCL‐R); about half had schizoid personality disorder,
as defined in Diagnostic and Statistical Manual of Mental Disorders (DSM)‐IV; and 88% had a sadistic personality
disorder (as described in the appendix of DSM‐III‐R). In another study, all 13 sexual serial murderers in a Canadian
high‐security prison had at least one personality disorder (Chan, Beauregard, & Myers, 2015). Using the Millon Clinical
Multiaxial Inventory‐III self‐report questionnaire (Millon, Millon, Davis, & Grossman, 2009), Culhane, Hildebrand,
Mullings, and Klemm (2016) investigated psychiatric diagnoses among 60 male serial‐homicide offenders incarcerated
in the United States and who had been active between 1967 and 2007. The authors report that two thirds of the
serial offenders suffered from at least one personality disorder.
The association between autism‐spectrum disorder (ASD) and multiple homicides is in a novel state (Reid, 2017b).
The prevalence of ASD is about 1% in the general population (Lundström, Reichenberg, Anckarsäter, Lichtenstein, &
Gillberg, 2015), and clinical and population‐based studies indicate that people with ASD are no more likely than peo-
ple from the general population to commit violent crimes (Hippler, Viding, Klicpera, & Happé, 2010; Im, 2016;
Lundström et al., 2014). Nevertheless, people with ASDs have been reported to be overrepresented in forensic sam-
ples (Kristiansson & Sörman, 2008), and furthermore, several authors have speculated that an association may exist
between autism‐spectrum psychopathology and both multiple‐homicide behaviour (Allely, Minnis, Thompson, Wilson,
& Gillberg, 2014; Hickey, 2013; Katz & Zemishlany, 2005; Silva, Ferrari, & Leong, 2002, 2003; Silva, Leong, & Ferrari,
2004) and necrophilia (Fitzgerald, 2013; Lester & White, 2011). In their systematic review, including scientific, legal,
and media sources, Allely et al. (2014) estimated that over 10% of multiple homicide offenders have an ASD. The
STURUP 3
authors concluded that such a figure would be considerably higher than the frequency of ASD in the general popula-
tion but emphasised that the presence of an ASD is not a sufficient cause in multiple killings, because the majority of
offenders do not suffer from such a disorder.
1.4 |This study
In short, there are only a few European studies on the prevalence of serial‐homicide offences and the characteristics
of serial‐homicide offenders; thus, more research is needed. Further, research suggests that personality disorders are
common and ASDs overrepresented among serial‐homicides offenders. The aims of this study were to report the
prevalence in Sweden of serial‐homicide offences and the characteristics of Swedish serial‐homicide offenders, to
compare the results with findings regarding single‐homicide offences and offenders. A secondary aim was to evaluate
the utility of a model for differentiating serial‐homicide offences from single‐homicide offences developed by
Pakkanen et al. (2015). The study aims to improve understanding of the circumstances surrounding serial‐homicide
offences outside of the North American context and, further, to provide guidance for investigators concerning
differences and similarities between serial‐and single‐homicide offenders.
2|METHOD AND MATERIAL
The study includes all convicted serial‐homicide offenders in Sweden from January 1, 1973, to December 31, 2012
(n= 25), as well as a control group consisting of a population‐based sample of 201 single‐homicide offenders collected
from January 1, 2007, to December 31, 2009.
2.1 |Data collection
Data were collected from the National Crime Register, which is administrated by the National Council for Crime
Prevention. The registry encompasses all convictions from January 1, 1973, and no data are ever erased, even in
the case of an individual's death or emigration. A search was conducted of all individuals who had been convicted
of two or more homicides between 1973 and 2012. Homicide was defined as either murder, infanticide, or voluntary
or involuntary manslaughter. After excluding individuals convicted of homicides committed outside of Sweden (who,
nonetheless, had been convicted in Sweden according to international laws), the search uncovered 638 convictions of
439 offenders. Because the registry does not differentiate between attempted and completed offences, all 638 court
verdicts from district courts were requested for review, in order to establish whether offences were completed or
merely attempted homicides, allowing identification of 611 homicide convictions (96%). Of the 27 instances in which
convictions could not be identified, 10 cases had been closed as the result of prosecutorial discretion, whereas the
remaining 17 court verdicts could not be procured for review owing to archiving procedures at closed or merged
courts. After exclusion of 359 subjects who did not meet inclusion criteria of two or more completed homicide
offences on two or more occasions, 80
1
subjects remained. Of these 80 offenders, 47 were recidivistic homicide
offenders (homicide, followed by conviction, followed by homicide; not included in this paper); eight were spree‐
homicide offenders (two or more homicides in a time period not exceeding 24 hr; not included in this paper); and
25 were serial‐homicide offenders (two or more homicides over a period exceeding 24 hr).
As a next step, an administrative forensic‐psychiatric registry administrated by the National Board of Forensic
Medicine was queried. The registry holds data on all forensic‐psychiatric examinations; 24 of 25 offenders in the
current study had undergone such an evaluation. All forensic‐psychiatric evaluations in Sweden are conducted by
the National Board of Forensic Medicine upon request by the courts. Evaluations are performed during a 4‐week
1
A Swedish man who was initially convicted of eight murders was excluded from the present study; the man had confessed to killing
about 30 people but later withdrew his confessions, after which, in the wake of a great deal of national and international media atten-
tion, the convictions were overturned.
4STURUP
assessment period by a multidisciplinary team consisting of a forensic psychiatrist, a forensic psychologist, and a
forensic social worker. As an aid in the diagnostic procedure, teams of evaluators at forensic‐psychiatric assess-
ment units in Sweden have since the late 1970s predominantly made use of the classification system presented
in American Psychiatric Association's DSM‐IV—since 1994, the fourth edition of the manual (American Psychiatric
Association, 2000). Reports from forensic‐psychiatric evaluations were collected and consisted of medical‐
psychiatric, psychological and social reports, and results of psychological testing. As part of the forensic‐psychiatric
evaluation, a psychological assessment is often conducted, during which intelligence functioning is measured
using instruments such as the Wechler Adult Intelligence Scale‐Revised (Wechsler, 2008) or Synonyms, Reasoning
and Blocks, which is a part of the larger Dureman–Sälde neuropsychological battery (Dureman, Kebbon, &
Österberg, 1971).
Data on the number of homicide victims between 1973 and 2012 were collected from Statistics Sweden for the
years 1973 to 1996 (www.scb.se) and from the National Board of Health and Welfare for the years 1997 to 2012
(www.socialstyrelsen.se).
2.2 |Reference group
The reference group had been collected in conjunction with another research project on homicide offences commit-
ted between 2007 and 2009 (Sturup, Karlberg, & Kristiansson, 2015). In total, 264 homicide incidents had been com-
mitted during the inclusion period; exclusion of unsolved (n= 36), homicide–suicide (n= 13), serial (n= 4; but included
in the serial sample), spree (n= 1), and recidivistic (n= 3) cases—as well as cases in which the offender had previously
been convicted of a homicide offence outside Sweden (n= 3) or had no Swedish social security number (n=3)—the
201 primary offenders of solved, single homicides who remained were designated the reference group. Data from
forensic‐psychiatric evaluations were collected for these offenders, in the same manner as for cases. Because far
fewer single‐homicide offenders (51%) than serial‐homicide offenders (96%) had undergone forensic‐psychiatric eval-
uation, data on mental disorders were also collected from the National Reception Unit at Kumla Prison of the Swedish
Prison and Probation Services, which rendered relevant information for an additional 17% of controls (at the begin-
ning of their sentences, all males convicted to a prison sentence of at least 4 years undergo a risk assessment that
includes examination by a psychologist [for more detailed description of the unit, see Skeem, Johansson, Andershed,
Kerr, & Louden, 2007]). Thus, in total, data on mental disorders were available for 68% of the controls. Substance
abuse was assessed using multiple sources (forensic‐psychiatric evaluations; data from the National Reception Unit;
personal investigation in criminal cases, conducted by the Prison and Probation Services; and other available informa-
tion the offender).
2.3 |The setting
Sweden is a Nordic country with a current population of over 10 million inhabitants, whereas the population in 1992
(in the middle of the observation period) was 8.5 million.
2
During the observation period 1973 to 2012, the annual
average number of homicides was 100 (1.2 per 100.000 inhabitants). The clearance rate in homicides was over
90% in the 1970s and the 1980s (Somander, 1981; Wikström, 1992), 85% in the 1990s (Rying, 2000), and 87% in
the 2000s (Sturup et al., 2015). The traditionally high clearance rate has decreased due to increases in firearm‐
perpetrated homicides in criminal milieus in the last years (Granath & Sturup, in review).
2
Https://www.scb.se/H/SOS%201911‐/Befolkningsstatistik/Befolkningsstatistik%20Del%204%20F%C3%B6dda%20d%C3%
B6da%20civilst%C3%A5nds%C3%A4ndringar%20(SOS)%201991‐2001/Befolkningsstatistik‐1992‐4‐Fodda‐doda‐
civilstandsandringar.pdf
STURUP 5
2.4 |Statistics and ethics
Differences between groups were examined using chi‐square tests or Fisher's exact test for categorical data, ttests
for continuous data, and odds ratios (ORs) from logistic regression. The predictive validity of the model by Pakkanen
et al. (2015) was evaluated, as in the original study, using results from receiver‐operating‐characteristic analysis,
reported as areas under the curve (AUCs). All tests of significance were two tailed, and when appropriate, 95% con-
fidence intervals (95% CI) are reported. The study was approved by the regional ethical board in Stockholm (study pro-
tocol 2012/1238‐31/5). The study only includes offences in which the offender had been convicted in court. Several
offenders had been suspected of, and sometimes even prosecuted for, additional homicide offences; however, in
accordance with the conditions of ethical approval, such offences were not included in the study.
3|RESULTS
3.1 |Prevalence and description
In total, 25 serial‐homicide offenders were identified; they were responsible for the deaths of 63 victims (with a range
of 2 to 11 offences), as a group, and were also convicted of an additional 25 attempted homicides (offences not
included in these analyses). Between 1973 and 2012, 3,933 individuals died as the result of homicide in Sweden. Thus,
the share of homicides attributed to serial‐homicide offenders was at least 1.6% (63 ÷ 3,933); as this figure only
includes victims from solved cases, the true figure is presumably somewhat higher. In order to not skew the analysis
towards the offender that committed 11 offences, only the three first offences in each series were included in the
analysis, resulting in 53 serial offences that was compared with the 201 single‐homicide offences.
Only one offender had communicated with the police by way of anonymous letters and telephone calls. Almost
one third of the serial offenders staged the crime scene to simulate manners of death other than homicide, including
natural causes (n= 5), accidents (n= 5), sudden‐infant‐death syndrome (n= 3), and suicide (n= 3). Most of these cases
involved a first, successful staging action, after which the offender was identified and apprehended following a second
offence; however, not all offenders who staged the crime scene did so in connection to both offences. As up today, 13
of the offenders are still incarcerated (52%), seven offenders have left the prison or forensic‐psychiatric hospital
(28%), and five offenders have deceased (20%).
3.2 |Characteristics and comparison with single‐homicide offenders
The 25 offenders were, on average, slightly younger than 30 years of age; the majority were males, although over 15%
were females. As seen in Table 1, the duration of the homicidal series was, on average, about 2 years, and the mean
time from the first offence to apprehension was over 3 years. A majority of offenders had a previous conviction, and
over half had been previously convicted of a violent offence; yet, in neither, this respect nor with respect to
sociodemographic factors did serial offenders differ from single offenders (Table 1).
The proportion of subjects with a personality disorder was higher in serial offenders than in single offenders;
a similar, yet more pronounced difference, was seen with regard to ASDs. Few of the offenders in either group
had a psychotic disorder (Table 1). With regards to overall prevalence, with both primary and secondary diagnoses
taken into consideration, serial offenders were more likely to have either a personality disorder or an ASD
compared with single offenders; in which the difference was statistically significant for the ASDs (OR = 7.9;
95% CI [2.6, 24.1]; n= 159). Regarding personality disorders, the 95% confidence interval just overlapped 1
(OR = 2.3; 95% CI [1.0, 5.6]; n= 159). Among the 25 subjects, 14 offenders had a personality disorder, compared
with 11 subjects who had personality disorders as the main diagnosis. Among the 14 subjects, six had antisocial
personality disorder (24%), three had borderline personality disorder (12%), and five had personality disorder not
otherwise specified (20%).
6STURUP
TABLE 1 Characteristics of serial‐homicide offenders (n= 25) and single‐homicide offenders in Sweden (n= 201)
Serial cases (n= 25) Single cases (n= 201) pvalue (χ
2
;df)
Sociodemographic factors
Mean age at first homicide, in years (SD) 29.0 (8.2) 33.7 (13.8) .1
a
(5.9; 224)
Gender, % (n) .12
b
(2.5; 1)
Male 84% (21) 93% (187)
Female 16% (4) 7% (14)
Criminal history
Previous conviction, % (n) .14 (2.2; 1)
No 40% (10) 26% (52)
Yes 60% (15) 74% (149)
Prior conviction for violent offence, % (n) .94 (0.05; 1)
No 56% (14) 55% (111)
Yes 44% (11) 45% (90)
Clinical factors
Substance abuse .72 (0.13; 1)
No 52% (13) 48% (97)
Yes 48% (12) 52% (104)
Diagnosis
cd
<.001 (35.4; 4)
Personality disorder 46% (11) 24% (33)
Autism‐spectrum disorder 33% (8) 4% (5)
Psychotic disorder 13% (3) 13% (18)
Other 8% (2) 45% (61)
None 0% (0) 13% (18)
Intelligence quotient
ce
.16 (6.6; 4)
Extremely low (<70) 4% (1) 2% (1)
Borderline (70–79) 29% (7) 20% (13)
Average (80–119) 42% (10) 69% (45)
Superior (120–129) 17% (4) 6% (4)
Very superior (>130) 8% (2) 3% (2)
Offence factors
Category n/a n/a
Sexual 32% (8)
Substance abuse‐related 24% (6)
Medical 8% (2)
Domestic 16% (4)
Contract killers 8% (2)
Others 12% (3)
Number of homicide victims, mean (SD) 2.7 (1.9) n/a n/a
Weeks between first and last offence, mean (SD) 98.4 (136.7) n/a n/a
Weeks between first offence and apprehension, mean
(SD)
163.9 (249.1) n/a n/a
a
ttest (t‐value; df).
b
Fisher's exact test.
c
Based on 24 offenders who had undergone forensic‐psychiatric evaluation.
d
Based on 135 offenders who had undergone forensic‐psychiatric evaluation.
e
Based on 65 offenders who had undergone intelligence testing; n/a = not applicable.
STURUP 7
About half of the offenders in both groups were diagnosed with a substance‐abuse disorder. Regarding intellec-
tual functioning, the sample seemed to be normally distributed, although the distribution was rather flat, with a small
share of subjects in the group adjudged to be of average intelligence. Compared with single offenders, serial offenders
were more likely to be of either below‐average (33% vs. 22%) or above‐average (25% vs. 9%) intelligence, presumably
reflecting the heterogeneity of serial‐homicide offenders with respect to intellectual functioning.
Planning and forensic awareness were also considerably more common in serial offences than in single offences
(Table 2). Victims who worked as prostitutes were, in general, uncommon, compared with what has been reported in
international research; nonetheless, they were more common in the serial‐homicide group. As seen in Table 2, serial
offences were more often sexually motivated (OR = 10.1; 95% CI [4.2, 24.4]; n= 254), whereas knife was considerably
more common in single cases.
3.3 |Differentiating serial offences from single offences
Four of the seven factors (female victim, victim prostitute, wounds to the hands, and forensic awareness) from the
model developed by Pakkanen et al. (2015) were significantly associated with serial cases, and a model incorporating
the four factors was able to differentiate serial cases from single cases significantly better than chance (AUC = 0.76;
SE = 0.35; 95% CI [0.70, 0.83]; n= 254), as could the full model incorporating all seven factors (AUC = 0.69; SE = 0.35;
95% CI [0.62, 0.76]; n= 254).
4|DISCUSSION
4.1 |Overall findings
It has previously been noted that some high‐profile multiple‐homicide offenders suffer from ASD (Allely et al., 2017;
Silva et al., 2002; Silva et al., 2004), and Allely et al. (2014) have suggested that the proportion is at least 10% among
serial‐homicide offenders. The present study, in which about one third of serial‐homicide offenders had been
diagnosed with an ASD, is the first population‐based study to provide evidence of such a notion. This finding has
potential implications within clinical forensic psychiatry. Because ASDs are associated with repetitive thoughts and
obsessive–compulsive behaviour (Russell, Mataix‐Cols, Anson, & Murphy, 2005), there is a need to examine whether
a serial violent offender suffers from an ASD and, if so, to examine the degree to which the disorder has contributed to
the offences. The issue is especially important in the context of questions of diminished responsibility (Haskins & Silva,
2006) and sentences of compulsory psychiatric care instead of prison, as stipulated by Swedish law (Kristiansson &
Sörman, 2008).
The second finding that has implications for police investigations is that the model by Pakkanen et al. (2015) was
able to differentiate serial homicide from single‐homicide offences in Sweden as well, although the predictive validity
was lower than in the original cohort (AUC = 0.69 vs. AUC = 0.88). This discrepancy is probably due to the reduced
effect size that is often observed in replication studies when a model is tested in a sample other than the sample in
which it was originally developed. Nonetheless, the results of the replication study do support the assumption that
serial‐homicide offences can be differentiated from single‐homicide offences; and on the whole, there were differ-
ences between the serial‐and single‐homicide offenders, though not with regard to sociodemographic or criminolog-
ical factors.
Discussion of the current findings in the light of research on serial killers from the North American context, per-
formed before and after publication of the FBI's Crime Classification Manual (Ressler, Douglas, Burgess, & Burgess,
1992), is unavoidable. The proportion of homicides attributed to serial offenders in Sweden is in line with what has
been proposed by international research (FBI, 2008; Jenkins, 1988; Kiger, 1990; Mouzos & West, 2007). According
to the Radford University Serial Killer Database, there were 10 identified serial killers in Sweden; however, the pres-
ent study found a markedly higher number. It is worth noting that the majority of subjects in this study were not serial
8STURUP
TABLE 2 Crime scene behaviour in Swedish serial‐homicide offences (n= 53) and single‐homicide offences (n= 201)
Serial cases (n= 53) Single cases (n= 201) pvalue (χ
2
;df)
Offender factors
Intoxication, % (n)
a
0.02 (5.4; 1)
No 57% (30) 39% (77)
Yes 43% (23) 61% (121)
Relationship, % (n) 0.002 (10.0; 1)
Known victim 57% (30) 78% (157)
Stranger victim 43% (23) 22% (44)
Co‐offender, % (n) 0.65 (0.2; 1)
No 87% (46) 89% (179)
Yes 13% (7) 11% (22)
Victimology
Gender, % (n) 0.02 (5.9; 1)
Male 49% (26) 67% (135)
Female 51% (27) 33% (66)
Age, % (n) <0.001 (29.4; 5)
0–15 years 17% (9) 2% (4)
16–19 years 8% (4) 9% (17)
20–29 years 11% (6) 26% (53)
30–44 years 32% (17) 20% (40)
45–64 years 15% (8) 30% (61)
>65 years 17% (9) 13% (26)
Victim intoxication, % (n) 0.004 (8.2; 1)
No 62% (33) 40% (81)
Yes 38% (20) 60% (120)
Prostitute victim, % (n) 0.007
c
(7.2; 1)
No 94% (50) >99% (200)
Yes 6% (3) <1% (1)
Modus operandi
Crime scene, % (n) 0.38 (0.78; 1)
Other than public 51% (27) 58% (116)
Public 49% (26) 42% (85)
Planning, % (n) 0.004 (8.1; 1)
No 28% (15) 50% (101)
Yes 72% (38) 50% (100)
Weapon brought to crime scene, % (n) 0.33 (0.94; 1)
No 53% (28) 60% (121)
Yes 47% (25) 40% (80)
Method of violence, % (n) <0.001 (43.8; 4)
Blunt force 11% (6) 19% (38)
Asphyxia 45% (24) 10% (19)
Knife/axe 19% (10) 51% (103)
Firearms 13% (7) 13% (26)
Other 11% (6) 8% (15)
(Continues)
STURUP 9
killers in the traditional sense, that is, male sexual‐sadistic offenders killing stranger females; nevertheless, over one
third of the offenders did display sexual motives, a rate considerably higher than in single‐homicide offences. It has
previously been recognised that not all serial‐homicide offenders are sexually motivated (FBI, 2008). The findings from
this study, in which sexual motives were identified in 32% of cases, are comparable to figures from Italy (Pakkanen
et al., 2015) and Germany (Harbort & Mokros, 2001)—where, respectively, 30% and 41% of serial‐homicides offences
were reported to be sexually motivated—yet lower than the figure of 55% reported from the United States (Kraemer
et al., 2004). In the present study, sexual serial offenders had sometimes been convicted of other, nonlethal
offences, including rape and attempted homicide; and in a few cases, offenders were tried but not convicted of
additional crimes.
There are at least three possible explanations for the apparently lower rate of traditional sexual serial killers in
Sweden than in the United States. First, Sweden may have a higher rate of unsolved sexual serial homicides: that
is, the lower rate may be attributable to the fact that many offenders have not been identified and convicted. How-
ever, unsolved Swedish homicide cases are typically characterised by male victims shot to death in public places,
whereas unsolved sexual homicides are unusual (Granath & Sturup, in review; Sturup, Karlberg & Kristiansson
2015; Sturup et al., 2018); therefore, this explanation lacks empirical support, though, admittedly, not all sexual
serial‐homicide offenders in Sweden are apprehended. Second, it may be the case that given the high clearance rate
in Sweden, potential serial‐homicide offenders are identified after commission of a single homicide; thus, some poten-
tial sexual serial‐homicide offenders are probably apprehended before they can commit a second sexual homicide.
Further, that prostitutes were seldom victims in either single‐or serial‐homicide offences in Sweden. This is very strik-
ing compared with serial‐homicide victims in, for example, the United States (which is over 30%; Quinet, 2011) and
may be attributable to the low rate (Danna, 2012) and character of prostitution in Sweden. Prostitution in Sweden
has more or less disappeared from the streets
3
and probably turned to more controlled contexts, where organised
crime is controlling the activity of selling and buying sex. One might speculate that the prostitutes are more vulnerable
to violence from these members of the organisation but in a controlled area where they are less vulnerable to aggra-
vated violence by males buying sex. A third possible explanation to the lower rate of traditional sexual serial killers in
Sweden than in the United States, of course, might be that there exists a true difference in the prevalence of serial‐
homicide offences between North American and Sweden.
3
Https://www.thelocal.se/20080110/9621
TABLE 2 (Continued)
Serial cases (n= 53) Single cases (n= 201) pvalue (χ
2
;df)
Victim hit several times, % (n) 0.08 (3.1; 1)
No 77% (41) 65% (130)
Yes 23% (12) 35% (70)
Forensic awareness, % (n) <0.001 (18.8; 1)
No 9% (5) 41% (83)
Yes 91% (48) 59% (118)
Staging, % (n) <0.01 (54.3; 1)
No 70% (37) 99% (199)
Yes 30% (16) 1% (2)
Body found outside, % (n) 0.94 (0.006; 1)
No 62% (33) 62% (124)
Yes 38% (20) 38% (77)
a
Data unavailable for three single‐homicide offenders.
b
Fisher's exact test.
10 STURUP
Another important factor is that in the present study, the rate of females among serial‐homicide offenders (16%)
was higher than the rate reported for Swedish single‐homicide offenders (7%; see also Trägårdh, Nilsson, Granath &
Sturup, 2016)—a finding consistent with previous reports that male preponderance is less pronounced in serial homi-
cides (Fox & Levin, 1998; Hickey, 2013). However, three of four female offenders in the present study were mothers
convicted of killing biological or adopted offspring; such offenders are not traditionally viewed as serial‐homicide
offenders but fulfilled the present study's inclusion criteria and was thereby included, well in line with what recently
has been suggested by serial‐homicide researchers (Aamodt, 2015; Yaksic, 2015). Another interesting circumstance is
the difference between single‐and serial‐homicide offences in relation to the instrumental/expressive dichotomy of
crime scene behaviour. Adjorlolo and Chan (2017) has in their review on the instrumental/expressive dichotomy con-
cluded that there is strong evidence for the typology from different homicide samples but also different settings. As
found in Table 2, it is evident that the single‐homicide offences in this study to a higher extent involved expressive
crime scene behaviour such as intoxication, blunt violence and knife, acquainted victims, defence wounds, and that
the victim was hit several times. However, the serial cases did to a higher extent display instrumental crime scene
behaviour such as stranger and female victim, planning, strangulation, forensic awareness, and staging. Even though
no statistical testing was conducted regarding the instrumental/expressive dichotomy, this study lends support to
the evidence reported by Adjorlolo and Chan (2017).
4.2 |Methodological considerations
An important factor that may have influenced the results is the fact that among the serial cases, 96% had undergone
forensic‐psychiatric or forensic‐psychological assessment, whereas the figure among the single offenders was only
68%. Forensic‐psychiatric evaluations can be performed for several reasons; one is that the offender has committed
an offence that is hard to understand or explain. Therefore, it is unsurprising that mental disorders were diagnosed
more often in serial offenders than single offenders; this circumstance is unlikely to have affected the prevalence
of mental disorders in serial cases but may contributed to an overestimation of differences between serial and single
offenders. Even though official data from the forensic‐psychiatric evaluations were used in the present study, diagno-
ses are made retrospectively, after the individual has committed a serious offence. This circumstance introduces a risk
that the assessor may make use of a psychiatric diagnosis to make sense of a heinous, otherwise hard‐to‐understand
violent act, in which case the mental disorder may become part of a circular argument whereby the offence confirms
the mental disorder and the mental disorder confirms the offence. Indeed, forensic‐psychiatric evaluations are per-
formed by a state‐run board of multidisciplinary teams of experienced clinicians, who, on average, each conduct about
30 forensic‐psychiatric evaluations per year; thus, rates of mental disorders among offenders who had undergone
forensic‐psychiatric evaluations may be inflated.
A second factor that may have influenced the results is the fact that the reference group was selected from a
short time period near the end of the inclusion period for cases; this major methodological drawback was difficult
to avoid, because collection of cases required such a long time period. It would be possible to start collecting serial‐
homicide offender and a control cohort of single‐homicide offenders prospectively, but in a Swedish setting, such a
process would have taken 20 to 30 years. On the other hand, both ascertainment of both cases and controls was pop-
ulation‐based. Another important notice is the uneven size of case sample and control sample.
As a final methodological consideration, it is worth pointing out that it is always difficult to draw firm conclusions
regarding unusual phenomena; thus, the major drawback of the study is the small number of subjects (N= 25).
Another way of handling the situation would have been to refrain from conducting the study, based on the argument
that too few subjects could be observed; however, as is the case regarding unusual but life‐threatening diseases or
terrorist attacks, the ramifications of serial homicides are very serious, thereby justifying research despite the very
low prevalence of the phenomenon. On the other hand, this study was population‐based, which increases the
generalisability of the findings. All in all, the author would contend that the study fairly reflects the prevalence of
serial‐homicide offences and characteristics of serial‐homicide offenders in a Northern European setting.
STURUP 11
5|CONCLUSIONS
The main finding of the present study is that serial‐homicide offenders more often suffer from personality disorders or
ASDs than single‐homicide offenders. With regard to personality disorders, this study merely replicates previous find-
ings; yet the high rate of serial‐homicide offenders suffering from ASDs is a novel finding and warrants reexamination
in future research. The present study largely reflects the findings of international research that indicate that serial
offences can be differentiated from single offences; it seems to be the case that the former offences more often
involve planning, victims unacquainted with the offender and strangulation, and forensic awareness and a sexual
motive on the part of the offender, whereas the latter offences involve knife and nonplanned violence preceded by
interpersonal conflicts. Overall, serial offences appear to be instrumental in nature, whereas single offences appear
to be expressive. Indeed, a final conclusion of the study is that serial offenders do differ from single offenders,
facilitating differentiation in contexts in which research findings are put into practice.
ACKNOWLEDGEMENTS
The author is grateful to Thomas Masterman, MD, PhD, at the Karolinska Institutet, Shilan Caman, PhD, also at the
Karolinska Institutet, and Detective Superintendent Eva von Vogelsang at the National Offender Profiling Unit at
the Swedish Police Authority for comments on previous versions of this paper.
ORCID
Joakim Sturup http://orcid.org/0000-0002-0404-8197
REFERENCES
Aamodt, M. G. (2015) Serial killer statistics. Retrieved from http://maamodt.asp.radford.edu/Serial%20Killer%20Informa-
tion%20Center/Serial%20Killer%20Statistics.pdf.
Aamodt, M. G., Moberg, J. L., Nash, T. M., Pendleton, G. B., Hommema, D. A., & Walker, C. E. (2007). Profiling serial killers:
When science and the media collide. Poster session presented at the annual meeting of the Society for Police and
Criminal Psychology, Springfield, MA.
Adjorlolo, S., & Chan, H. C. (2017). The nature of instrumentality and expressiveness of homicide crime scene behaviors: A
review. Trauma, Violence & Abuse,18(2), 119–133.
Adjorlolo, S., & Chan, H. C. O. (2014). The controversy of defining serial murder: Revisited. Aggression and Violent Behavior,
19(5), 486–491.
Allely, C. S., Minnis, H., Thompson, L., Wilson, P., & Gillberg, C. (2014). Neurodevelopmental and psychosocial risk factors in
serial killers and mass murderers. Aggression and Violent Behavior,19(3), 288–301.
Allely, C. S., Wilson, P., Minnis, H., Thompson, L., Yaksic, E., & Gillberg, C. (2017). Violence is rare in autism: When it does
occur, is it sometimes extreme? The Journal of Psychology,151(1), 49–68.
American Psychiatric Association (2000). In 4th ed (Ed.), text revisionDiagnostic and statistical manual of mental disorders.
Washington, DC: American Psychiatric Association.
Bennell, C., Bloomfield, S., Emeno, K., & Musolino, E. (2013). Classifying serial sexual murder/murderers: An attempt to
validate Keppel and Walter's (1999) model. Criminal Justice and Behavior,40(1), 5–25.
Canter, D. V., Alison, L. J., Alison, E., & Wentink, N. (2004). The organized/disorganized typology of serial murder: Myth or
model? Psychology, Public Policy, and Law,10(3), 293–320.
Canter, D. V., & Wentink, N. (2004). An empirical test of Holmes and Holmes's serial murder typology. Criminal Justice and
Behavior,31(4), 489–515.
Canter, D. V., & Youngs, D. (2009). Investigative psychology: Offender profiling and the analysis of criminal action. Chichester:
John Wiley & Sons.
Chan, H. C. O., Beauregard, E., & Myers, W. C. (2015). Single victim and serial sexual homicide offenders: Differences in crime,
paraphilias and personality traits. Criminal Behaviour and Mental Health,25(1), 66–78.
Culhane, S. E., Hildebrand, M. M., Mullings, A., & Klemm, J. (2016). Self‐reported disorders among serial homicide offenders:
Data from the millon clinical multiaxial inventory—III. Journal of Forensic Psychology Practice,16(4), 268–286.
12 STURUP
Danna, D. (2012). Client‐only criminalization in the city of Stockholm: A local research on the application of the “Swedish
model”of prostitution policy. Sexuality Research & Social Policy,9(1), 80–93.
DeLisi, M., & Scherer, A. M. (2006). Multiple homicide offenders: Offense characteristics, social correlates, and criminal
careers. Criminal Justice and Behavior,33(3), 367–391.
Dureman, I., Kebbon, L., & Österberg, E. (1971). Manual for the DS battery (in Swedish). Stockholm: Skandinaviska Testforlaget,
Sweden.
FBI (2008). Serial murder: Multi‐disciplinary perspectives for investigators. Washington DC: Behavioral Science Unit, National
Center for the Analysis of Violent Crime, U.S. Department of Justice, U.S.A.
Fitzgerald, M. (2013). Violent and Dangerous to Know. New York: Nova Science Publishers, Inc., USA.
Fox, J. A., & Levin, J. (1998). Multiple homicide: Patterns of serial and mass murder. Crime and Justice,23, 407–455.
Fox, J. A., & Levin, J. (2014). Extreme killing: Understanding serial and mass murder. USA: Sage Publications.
Fridel, E. E., & Fox, J. A. (2017). Too few victims: Finding the optimal minimum victim threshold for defining serial murder.
Psychology of Violence. http://doi.org/10.1037/vio0000138
Granath, S., & Sturup, J. (in review). Homicide clearance in Sweden 1990 to 2013 with special reference to firearm‐perpetrated
homicides.
Häkkänen, H. (2008). A study of offense patterns and psychopathological characteristics among recidivistic Finnish homicide
offenders. In Serial murder and the psychology of violent crimes (pp. 51–62). Totowa, NJ: Humana Press.
Harbort, S., & Mokros, A. (2001). Serial murderers in Germany from 1945 to 1995: A descriptive study. Homicide Studies,5(4),
311–334.
Haskins, B. G., & Silva, J. A. (2006). Asperger's disorder and criminal behavior: Forensic‐psychiatric considerations. Journal of
the American Academy of Psychiatry and the Law Online,34(3), 374–384.
Hickey, E. W. (2013). Serial murderers and their victims. USA: Cengage Learning.
Hill, A., Habermann, N., Berner, W., & Briker, P. (2007). Psychiatric disorders in single and multiple sexual murderers.
Psychopathology,40(1), 22–28.
Hippler, K., Viding, E., Klicpera, C., & Happé, F. (2010). No increase in criminal convictions in Hans Asperger's original cohort.
Journal of Autism and Developmental Disorders,40, 774–7780.
Holmes, R. M., & Holmes, S. T. (Eds.) (1998). Contemporary perspectives on serial murder. Thousands Oak: Sage.
Im, D. S. (2016). Template to perpetrate: An update on violence in autism spectrum disorder. Harvard Review of Psychiatry,
24(1), 14–35.
Jenkins, P. (1988). Serial murder in England 1940–1985. Journal of Criminal Justice,16(1), 1–15.
Katz, N., & Zemishlany, Z. (2005). Criminal responsibility in Asperger's syndrome. The Israel Journal of Psychiatry and Related
Sciences,43(3), 166–173.
Kelleher, M. D., & Kelleher, C. L. (1998). Murder most rare: The female serial killer. Westport, CT: Praeger.
Keppel, R. D., & Walter, R. (1999). Profiling killers: A revised classification model for understanding sexual murder.
International Journal of Offender Therapy and Comparative Criminology,43(4), 417–437.
Kiger, K. (1990). The darker figure of crime: The serial murder enigma. In S. Eggers (Ed.), Serial murder: An elusive phenomenon
(pp. 35–52). New York: Praeger.
Kraemer, G. W., Lord, W. D., & Heilbrun, K. (2004). Comparing single and serial homicide offenses. Behavioral Sciences & the
Law,22(3), 325–343.
Kristiansson, M., & Sörman, K. (2008). Autism spectrum disorders: Legal and forensic psychiatric aspects and reflections.
Clinical Neuropsychiatry: Journal of Treatment Evaluation,5(1), 55–61.
Lester, D., & White, J. (2011). Necrophilia in serial killers: Is there evidence for Asperger's syndrome? Clinical Neuropsychiatry,
8(6), 367–368.
Lundström, S., Forsman, M., Larsson, H., Kerekes, N., Serlachius, E., Långström, N., & Lichtenstein, P. (2014). Childhood
neurodevelopmental disorders and violent criminality: A sibling control study. Journal of Autism and Developmental
Disorders,44(11), 2707–2716.
Lundström, S., Reichenberg, A., Anckarsäter, H., Lichtenstein, P., & Gillberg, C. (2015). Autism phenotype versus
registered diagnosis in Swedish children: Prevalence trends over 10 years in general population samples. BMJ,350,
h1961.
Miller, L. (2014). Serial killers: I. Subtypes, patterns, and motives. Aggression and Violent Behavior,19(1), 1–11.
Millon, T., Millon, C., Davis, R., & Grossman, S. (2009). MCMI‐III: Millon clinical multiaxial inventory. Pearson.
Mouzos, J., & West, D. (2007). An examination of serial murder in Australia. Canberra: Australian Institute of Criminology.
STURUP 13
Pakkanen, T., Zappalà, A., Bosco, D., Berti, A., & Santtila, P. (2015). Can hard‐to‐solve one‐off homicides be distinguished
from serial homicides? Differences in offense behaviours and victim characteristics. Journal of Criminal Psychology,5(3),
216–232.
Quinet, K. (2011). Prostitutes as victims of serial homicide: Trends and case characteristics, 1970‐2009. Homicide Studies,
15(1), 74–100.
Radford University Serial Killer Database (2015). Serial killers statistics [downloaded 2016–01‐09] http://maamodt.asp.
radford.edu/Serial%20Killer%20Information%20Center/Serial%20Killer%20Statistics.pdf.
Reid, S. (2017a). Compulsive criminal homicide: A new nosology for serial murder. Aggression and Violent Behavior,34,
290–301.
Reid, S. (2017b). Developmental pathways to serial homicide: A critical review of the biological literature. Aggression and
Violent Behavior,35,52–61.
Ressler, R. K., Burgess, A. W., & Douglas, D. E. (1988). Sexual homicide: Patterns and motives. Lexington, MA: Lexington
Books.
Ressler, R. K., Douglas, J. E., Burgess, A. W., & Burgess, A. G. (1992). Crime classification manual: A standard system for
investigating and classifying violent crimes. New York: Lexington Books.
Russell, A. J., Mataix‐Cols, D., Anson, M., & Murphy, D. G. (2005). Obsessions and compulsions in Asperger syndrome and
high‐functioning autism. The British Journal of Psychiatry,186(6), 525–528.
Rying, M. (2000). Homicide in Sweden. A descriptive study [Dödligt våld i Sverige. En deskriptiv studie]. Stockholm: Department of
Criminology, Stockholm University.
Salfati, C. G., Horning, A. M., Sorochinski, M., & Labuschagne, G. N. (2015). South African serial homicide: Consistency
in victim types and crime scene actions across series. Journal of Investigative Psychology and Offender Profiling,12(1),
83–106.
Santtila, P., Pakkanen, T., Zappalà, A., Bosco, D., Valkama, M., & Mokros, A. (2008). Behavioural crime linking in serial
homicide. Psychology, Crime & Law,14(3), 245–265.
Schlesinger, L. B. (2001). Is serial homicide really increasing? The Journal of the American Academy of Psychiatry and the Law,
29(3), 294–297.
Sewall, L. A., Krupp, D. B., & Lalumière, M. L. (2013). A test of two typologies of sexual homicide. Sexual Abuse,25(1),
82–100.
Silva, J. A., Ferrari, M. M., & Leong, G. B. (2002). The case of Jeffrey Dahmer: Sexual serial homicide from a neuropsychiatric
developmental perspective. Journal of Forensic Sciences,47(6), 1347–1359.
Silva, J. A., Ferrari, M. M., & Leong, G. B. (2003). Asperger's disorder and the origins of the Unabomber. American Journal of
Forensic Psychiatry,24(2), 5–44.
Silva, J. A., Leong, G. B., & Ferrari, M. M. (2004). A neuropsychiatric developmental model of serial homicidal behavior.
Behavioral Sciences & the Law,22(6), 787–799.
Skeem, J., Johansson, P., Andershed, H., Kerr, M., & Louden, J. E. (2007). Two subtypes of psychopathic violent offenders that
parallel primary and secondary variants. Journal of Abnormal Psychology,116(2), 395–409.
Somander, L. (1981) Violent crime resulting in death in Sweden during the 1970's [Våldsbrott som lett till döden
i Sverige under 1970‐talet]. Paper presented at the Swedish medical associations conference in Stockholm
1981‐12‐04.
Stone, M. H. (2001). Serial sexual homicide: Biological, psychological, and sociological aspects. Journal of Personality Disorders,
15(1), 1–18.
Sturup, J., Karlberg, D., & Kristiansson, M. (2015). Unsolved homicides in Sweden: A population‐based study of 264
homicides. Forensic Science International,257, 106–113.
Sturup, J., & Lindqvist, P. (2014). Homicide offenders 32 years later—A Swedish population‐based study on recidivism.
Criminal Behaviour and Mental Health,24(1), 5–17.
Sturup, J., Rodre, S., Karlberg, D., von Vogelsang, E., Rying, M. & Caman, S. (in review) (2018). Male‐on‐female sexual homi-
cides in Sweden 1990 to 2013: A population‐based controlled study of incidents, victims and offenders.
Trägårdh, K., Nilsson, T., Granath, S., & Sturup, J. (2016). A time trend study of Swedish male and female homicide offenders
from 1990 to 2010. International Journal of Forensic Mental Health,15(2), 125–135.
Trojan, C., & Salfati, C. G. (2011). Linking criminal history to crime scene behavior in single‐victim and serial homicide:
Implications for offender profiling research. Homicide Studies,15(1), 3–31.
Wechsler, D. (2008). Wechsler adult intelligence scale—Fourth Edition (WAIS–IV). San Antonio, TX: NCS Pearson.
14 STURUP
Wikström, P. O. H. (1992). Context‐specific trends in criminal homicide in Stockholm 1951–1987. Studies on Crime & Crime
Prevention,1(1), 61–87.
Wilson, D. (2007). Serial killers: Hunting Britons and their victims, 1960 to 2006. Winchester: Waterside Press.
Yaksic, E. (2015). Addressing the challenges and limitations of utilizing data to study serial homicide. Crime Psychology Review,
1(1), 108–134.
How to cite this article: Sturup J. Comparing serial homicides to single homicides: A study of prevalence,
offender, and offence characteristics in Sweden. J Investig Psychol Offender Profil. 2018;1–15. https://doi.
org/10.1002/jip.1500
STURUP 15
