Content uploaded by Catherine Hamilton-Giachritsis
Author content
All content in this area was uploaded by Catherine Hamilton-Giachritsis on Jun 30, 2016
Content may be subject to copyright.
1
Can Serial Rapists be distinguished from One-Off Rapists?
Chelsea Slater, Jessica Woodhams and Catherine Hamilton-Giachritsis
University of Birmingham
‘The final, definitive version of this paper has been published online first in Behavioral
Sciences & The Law Vol. 32 p220–239, 2014 by John Wiley and Sons
DOI: 10.1002/bsl.2096
http://onlinelibrary.wiley.com/doi/10.1002/bsl.2096/abstract
Embargo period: April 2014 - March 2016
We would like to acknowledge the assistance of the Serious Crime Analysis Section
(Serious Organised Crime Agency) for allowing the sampling of the data in their care
for this research project.
Author Note
Chelsea Slater, School of Psychology, University of Birmingham, Birmingham, United
Kingdom; Jessica Woodhams, School of Psychology, University of Birmingham,
Birmingham, United Kingdom; Catherine Hamilton-Giachritsis, School of Psychology,
University of Birmingham, Birmingham, United Kingdom.
Correspondence concerning this article should be addressed to Chelsea Slater, School of
Psychology, Frankland Building, University of Birmingham, Edgbaston, Birmingham,
B15 2TT, UK. Email: CLS011@bham.ac.uk.
2
ABSTRACT
There are investigative advantages to being able to determine early in a police
investigation whether a rape has been committed by a serial or one-off rapist. Previous
research has found some differences in the crime-scene behaviours of serial and one-off
rapists, however, this research suffers from the limitation of utilising a sample of rapes
within which there was a mixture of victim-offender relationships. To address this
limitation, this study sampled 38 serial (two or more convictions) and 50 one-off (one
conviction) stranger rapists and compared their crime scene behaviour across four
domains (control, sex, escape and style behaviours). Serial and one-off rapists differed
in some control and sexual behaviours; in particular, in the type of victim targeted, the
offence locations, methods of control and the sexual acts forced upon the victim.
However, the results did not indicate a striking difference in the offending behaviour of
the two groups. The implications of these findings for criminal investigations are
discussed.
Keywords: serial, singleton, single, rape, crime scene behaviour
3
Can Serial Rapists be distinguished from One-Off Rapists?
In a recent Government report entitled Forging the links: Rape investigation and
prosecution, Her Majesty’s Inspectorate of Constabulary (HMIC) and Her Majesty’s
Inspectorate of the Crown Prosecution Service (HMICPS) (2012) reported on their
investigation of current police procedures and results in England and Wales. One of the
main recommendations of their investigations was to improve the collection of
intelligence, specifically with regard to serial rapists. The report noted confusion within
police forces regarding what constituted a serial rapist and adopted the definition
commonly used in academic research (Beauregard, Rossmo, & Proulx, 2007; Grubin,
Kelly, & Brunsdon, 2001; LeBeau, 1987; Park, Schlesinger, Pinizzotto, & Davis, 2008;
Santtila, Junkkila, & Sandnabba, 2005) and also used in the current article: Serial rapists
are those who commit more than two offences against different victims. A key
recommendation of this publication was that police forces in England and Wales should
treat every stranger rape that is reported to them as part of a potential series. This is a
potentially costly recommendation to implement, in terms of both time and resources
(Rainbow, in press).
In responding to such a recommendation in a cost-effective way, it would be
beneficial for police forces to be able to differentiate at the early stages of an
investigation whether they are indeed dealing with a serial rapist or a “one off” rapist.
One way of achieving this has been suggested by previous authors (Grubin et al., 2001)
and involves attempting to use crime scene behaviour (as reported by the victim) to
predict whether a rapist is likely to be a serial rapist or a one-off rapist. In such a
scenario, when a rape is reported to police, a crime analyst could analyse the behaviour
of the offender (as reported by the victim) to identify any key features, established by
4
empirical research, that would suggest the offence was part of a series. Where such
indicators are found, they could then be communicated to the investigative team. This
may result, for example, in the analyst being tasked to try and identify other crimes that
form part of the same series based on behavioural similarity (a practice termed crime
linkage). Crime linkage is not the focus of this article; therefore, interested parties are
referred to Bennell, Mugford, Ellingwood and Woodhams (in press) for a summary of
the research to date.
Differentiating one-off from serial offenders
To be able to accurately differentiate the crimes of serial versus one-off
offenders, and therefore avoid providing the police with erroneous recommendations,
there must be differences in behaviour that are consistently observed between these two
groups. In the scenario described above, a crime analyst presented with a given crime
would not know whether it was the work of a one-off offender, or, if the work of a serial
offender, at what stage in a series the crime was committed. In an ideal world, the
behaviour of one-off offenders would differ from all the crimes committed by a serial
offender. This would require consistency in the crime scene behaviour of serial
offenders, as well as all serial offenders differing from one-off offenders in the same
way. Research of crime linkage, which is concerned with the relative consistency and
distinctiveness of the behaviour of serial offenders, indicates that this ideal scenario
does not exist. In fact, it has been shown that while some serial offenders do show a
high degree of consistency in behaviour, not all do (e.g., see Woodhams &
Labuschagne, 2012). In addition, this research shows that different serial offenders
behave in different ways (Sorochinski & Salfati, 2010). Therefore, the ideal scenario
whereby only serial offenders (and not one-off offenders) show behaviour X and that
5
behaviour X is consistently observed within crime series, thereby enabling perfect
prediction of whether a crime is the work of a serial or one-off offender, will not be
realised. This does not, however, mean that the search for behaviours that might be
more suggestive of a serial offender is doomed. It is possible that the variability in
behaviour seen within series in the crime linkage literature reflects evolution of crime
scene behaviour and there may be some behaviours that are more likely to be produced
by serial offenders at the end of their series. Indeed there has been some evidence of this
(Grubin et al., 2001). Such behaviours could still be utilised to suggest the presence (or
not) of a serial offender but there will likely be a degree of error associated with such
predictions.
Several studies have tried to identify means of differentiating between serial and
one-off offenders. These studies have tended to focus on homicide offenders or rapists.
For example, Kraemer, Lord and Heilbrun (2004) studied a sample of homicide
offenders gathered from Federal Bureau of Investigation records. Their sample
consisted of 195 single (one-off) offenders (who accounted for 133 victims), and
compared them to the first offence of 147 serial offenders (who accounted for 133
victims). Using a chi squared analysis they looked at victim and offender characteristics,
intent, relationship between victim and offender, approach, locations, body disposal,
and different evidence types. Their three most significant findings were that serial
offenders were more often strangers to their victims, more likely to strangle their
victims, and more likely to leave the victim’s body in a remote location. Using a
Discriminant Function Analysis based on the crime scene variables they were able to
correctly classify the offences as either a single or serial homicide for 72.2% to 76% of
6
the cases (depending on the number of variables included), and when focusing on
female victims only the percentage increased to 78.6%.
Salfati and Bateman (2005) also investigated single and serial homicides. They
compared a sample of 23 serial murderers from the USA to a sample of 247 single
murderers from the UK from a previous study (Salfati, 2003). They examined 61 crime
scene behaviours and 33 offender characteristics. They found that serial murderers were
more likely to display behaviours that reflected a higher degree of planning and control,
compared to the single murderers whose crimes were more impulsive and emotional.
The serial offenders’ behaviours appeared to be more frequently motivated by delaying
detection, controlling the victim, theft of property and engaging in sexual acts with the
victim. The most common behaviours displayed by the single offenders were related
directly with the killing. Salfati and Bateman suggested single (one-off) murderers are
focused on the actual murder whereas serial murderers are influenced more by other
motives.
With regards to differentiating serial from one-off rapists, only three studies
have investigated ways of doing this. In 1987, LeBeau investigated the offending
patterns of serial rapists compared to “open” cases of rape (rapes where the identity of
the suspect remained unknown) and one-off rapists. The sample was comprised of all
612 incidences of rape perpetrated by a lone offender in San Diego, California, from
1971-1975; separated into 194 open cases, 80 single, and 151 serial offences. Using chi
square analyses, the relationship between the rapist and the victim, the approach, and
the number of scenes involved in the offence were compared across groups. Similar to
Kraemer et al.’s findings for serial murderers, LeBeau found that the serial rapists were
significantly more likely to be strangers to their victims. The serial offenders were also
7
more likely to use a blitz style approach and not move their victims very far. LeBeau
speculated that several behaviours commonly displayed by the serial rapists were
related to avoiding or delaying their apprehension. The serial rapists, therefore, were
similar in this respect to the cases that remained unsolved (the open cases) allowing the
serial offender to commit multiple offences before being apprehended.
It was not until 2001 that Grubin and colleagues conducted the next study to
investigate differences between serial and singleton (one-off) offences of serious sexual
assaults. From a UK database of sexual assaults they sampled the crimes of 129 one-off
offenders and 81 serial offenders who had committed 339 attacks. They arranged 30
crime scene behaviours present in their sample into four different domains and used
cluster analysis to develop distinct types within each domain. The four domains were
termed control (behaviours focused on gaining control of the victim), sex (behaviours
that are part of the sexual component of the attack), escape (behaviours associated with
leaving the crime scene or avoiding detection), and style (behaviours that are not
necessary for the attack that reflect the offender’s personality or style). Having grouped
the crime scene behaviours in this way, Grubin et al. initially conducted a cluster
analysis to determine whether “singleton” offences would cluster differently to serial
offenders’ “first”, “second”, etc. offences. From this analysis, a cluster was identified
that contained 61% of the singleton cases and first offences in the series, however it also
contained 42% of the subsequent crimes in the series meaning any differences between
this cluster and others would be of limited practical value.
The most recent research into serial and one-off rapists was conducted by Park
et al. (2008). The behaviours of 22 serial rapists from the USA were compared to 22
one-off rapists using chi-square analyses. For the serial rapists, two offences from each
8
series were randomly selected for comparison to the one-off rapists’ offences. Twenty-
eight different behavioural variables were studied which were divided into three themes:
violence, interpersonal involvement, and criminal sophistication. The violence theme
contained 11 variables that represented the nature of the offender’s violence towards the
victim (e.g., blitz-style attack, weapon use, vaginal penetration). The one-off rapists
were more likely to display these types of behaviours; specifically threatening the
victim and engaging in manual hitting and kicking, as well as vaginal and/or oral
penetration. The interpersonal involvement theme comprised seven variables which
described the type of interaction between the victim and offender (e.g., using a
confidence approach, making sexual comments, extending time with the victim). One-
off offenders were more likely to force the victim to participate in the sexual assault and
more often made sexual comments (e.g., “Do you like it?”) than the serial offenders.
The criminal sophistication theme consisted of 10 variables that focused on assisting the
offender in the commission of the offence (e.g., having forensic awareness, planning,
gagging the victim). Here the serial offenders were more likely to display forensic
awareness, deter the victim’s resistance, gag the victim, use a surprise attack, ask the
victim questions, and complete the act of rape.
Rationale
As can be seen above, the existing literature on whether there are differences in
crime scene behaviour between serial and one-off offenders is very limited, meaning
there is little guidance at present that the police could utilise in trying to determine early
in an investigation if they are dealing with a serial or a one-off offender. With regards
to the recommendations from the Forging the links report, there are only three existing
studies of serial versus one-off rapists that could give any indication as to what crime
9
scene behaviour might suggest a rape was committed by a serial offender, rather than a
one-off rapist. In addition, all three studies with rapists cited above suffer from the same
methodological flaw, that they have analysed samples of rapists who have a mixture of
relationship types with their victims, i.e., their samples contain offenders who were
acquainted with their victims as well as those who were strangers to their victims. The
relationship between offenders and victims will likely impact the behaviours displayed
during the offence, such as the approach style utilised or the means used to control the
victim. For example, associations between victim age and victim-offender relationship
have been reported in studies of rape with older victims more likely assaulted by
strangers (Muram, Hostetler, Jones, & Speck, 1995), as well as differences between
stranger and acquaintance rapes in terms of approach location, violent acts, weapon use,
sexual acts, use of intoxicants and post-rape offender behaviour (Bownes, O’Gorman, &
Sayers, 1991; Koss, Dinero, Seibel, & Cox, 1988). Previous findings of differences in
the offence behaviour displayed by serial versus one-off rapists could therefore be a
product of differences in the proportions of victim-offender relationships in the two
subsamples.
In addition, previous studies comparing the behaviour of one-off and serial
offenders have failed to include cross-validation of their findings of differences between
the two. Further, it has been common practice to compare the offences of one-off
offenders to a randomly-selected crime from each series of the serial offenders. The
difficulty with this approach is that, as alluded to above, the behaviour of serial
offenders is not always consistent from crime-to-crime. The current study aimed to
develop the literature in this area by overcoming this fundamental limitation and not
only comparing the behaviour of one-off rapists to that displayed in a randomly selected
10
rapes committed by each serial rapists, but by also making comparisons to the behaviour
displayed in the first (known) rape in each series and the last (known) rape in each
series. This allowed for assessment of whether apparent differences in behaviour
between the two groups of rapist would generalise across these three subsets of offences
and therefore whether they were robust.
METHOD
Sample
A non-random national sample of rapes was obtained from the Serious Crime
Analysis Section (SCAS), of the Serious Organised Crime Agency, UK. SCAS is an
analytical unit with national responsibility to carry out analytical work on behalf of all
police forces. SCAS collates and analyses information on serious crimes that fulfil its
criteria, predominately stranger murders, and serious sexual assaults and/or rapes.
SCAS hold a database called the Violent Crime Linkage Analysis System (ViCLAS)
which contains information about the location of sexual crimes that meet their criteria
and the behaviours displayed during each offence by the offender. They hold the most
comprehensive dataset of stranger rapes in the UK.
A sample of rapes and attempted rapes committed by serial and one-off adult
male, sexual offenders was requested from SCAS. The rapes all met the definition of
rape as stated in the Sexual Offences Act of 2003. This defines rape as where “A
person commits an offence if he intentionally penetrates the vagina, anus or mouth of
another person with his penis” and that person does not consent (Sexual Offences Act,
2003, p. 1). A sample of 38 serial and 50 one-off rapists and their offence behaviours
were provided for analysis. The cases only included those with a lone, female victim
11
and a lone, stranger, male offender. All crimes included in the sample were from cases
that were closed with the offender having been convicted of the offence.
Serial Rapists
The sample of serial rapists represented 38 males with a mean age at the time of
the offence of 31.4 years (range 18-57 years). Seventy-four per cent of the offenders
(n=28) were of White European ethnicity, 8% (n=3) were Dark European, 13% (n=5)
were African or Caribbean, and 5% (n=2) identified as Other.
The serial offenders committed a combined 147 known sexual offences, of
which 120 were rapes or attempted rapes. Only the latter 120 offences were utilised
within the analysis comparing serial versus one-off rapists to ensure consistency in the
offence types under comparison. The offenders’ series ranged from two to 10 offences,
and the mean series length was four offences (the mode was three offences per
offender).
One-Off Rapists
The one-off rapists had a mean age at the time of the offence of 30.9 years
(range 18-55 years). All offenders were male. Seventy per cent of the offenders (n=35)
were of White European ethnicity, 2% (n=1) were Dark European, 12% (n=6) were
African or Caribbean, 10% (n=5) were Asian, 4% (n=2) were Arabic, and 2% (n=1)
were identified as Other. Of the 50 offences they had committed, 10 were attempted
rapes, while 40 were completed rapes.
Serial Rapists’ Victims
All 38 of the victims, from the studied sample, were female, and their mean age
at the time of the offence was 30.0 years (range 18-76 years). Seventy-six per cent
12
(n=29) of the victims were of White European ethnicity, 3% (n=1) were Dark European,
3% (n=1) were Arabic, and for 18% (n=7) their ethnicity was not recorded on the police
database.
One-Off Rapists’ Victims
All the victims were female, with a mean age at the time of the offence of 28.2
years (range 18-59 years). Eighty-four per cent (n=42) of the victims were of White
European ethnicity, 2% (n=1) were African or Caribbean, 4% (n=2) identified as Other,
and for 10% (n=5) their ethnicity was not recorded.
Procedure
Based on the victim’s account of the crime, SCAS codes each offence that
comes to their attention in a standard manner onto their ViCLAS database. The data
regarding the offenders’ behaviours during the offence were provided to the authors as a
spreadsheet of numerical codes representing this standardised coding. In total, 217
different behaviours were included in this spreadsheet ranging from the type of location
at which the offence was committed, to forms of violence used against the victim to
forensic precautions and sexual acts. All the variables had been coded dichotomously,
where 1 represented the presence of an action or behaviour during the offence, while 0
represented an absence or unknown data for a given behaviour.
Inter-rater reliability assessments of this coding were not possible as the data
were already coded when it was provided to the authors. However, all data inputted onto
the SCAS database is completed by a team of highly trained individuals, and is done in-
house in a controlled environment. Prior to employment, applicants are tested for their
attention to detail and ability to identify relevant information. Upon recruitment, staff
undergo a rigorous and lengthy training programme, which requires them to evidence a
13
clear understanding of behavioural vagaries, as well as detailed knowledge of the
system. From recruitment to working unsupported takes several months, in order to
ensure accuracy and knowledge. Initial training is not undertaken on a live database,
and staff will not begin working on the live database until they have evidenced their
capability to complete inputs accurately. A ‘Quality Control Guide’ is utilised by
everyone inputting data on the database, which ensures consistency in decision making
in relation to difficult issues. Where an unusual aspect is encountered, for which no
precedent has been set, a dedicated, experienced team meets to review the situation and
make a decision. This decision is then recorded for future reference to ensure future
consistency. In addition to having undertaken the above, each inputted case goes
through a detailed quality assurance process prior to any analysis taking place. This
involves a review of the inputted information in comparison to case details, by an
analyst from within the team and anomalies or errors are fed back to the inputter and
amended on the database.
No information that could be used to identify the offender, victim, or location
was present within the spreadsheet given to the authors thereby protecting the identity
of all parties. In accordance with previous research on the behaviour of serial rapists
(Santtila et al., 2005), behaviours that did not occur in at least five per cent of the total
offences were not included in the analysis. These behaviours were determined to be too
uncommon to be of pragmatic use to the police. This resulted in a total of 80 offence
behaviours for comparison across serial and one-off rapists.
As noted above, the number of offences committed by each serial rapist varied
from two to 10. In order to prevent the more prolific of the serial offenders from
unnecessarily biasing the results, only one crime from each offender was included in
14
any analysis. It is standard practice in research on serial criminals to control for
potential bias in this way (Bennell & Canter, 2002; Park et al., 2008). Several different
methods have been used to select such cases from a larger pool including using the last
known offences committed (e.g., Woodhams & Toye, 2007), or randomly selecting one
offence from each series (e.g., Park et al., 2008). While the random selection method
has been used in studies comparing one-off versus serial rapists in the past (Park et al.,
2008), three sampling methods were used to create three serial subsets for comparison
with the one-off rapes; a) randomly sampling one offence per series, b) selecting the
first offence from each series, and c) selecting the last offence from each series. In each
analysis there were 38 serial rapes/attempted rapes and 50 one-off rapes/attempted
rapes. Of the 38 serial offences randomly chosen, six were attempted rapes, while the
other 32 were completed rapes, and for the 50 one-off offences, 10 were attempted
rapes, while the other 40 were completed rapes. In the first serial offences vs. one-off
offences sample, five of the serial offences were attempted rapes while the other 33
were completed rapes, and in the last serial offence vs. one-off offences sample, six of
the serial offences were attempted rapes with the other 32 being completed rapes.
RESULTS
Comparisons were made between the frequencies of behaviours exhibited by
serial vs. one-off offenders for the 80 offence behaviours. For ease of presentation, these
behaviours have been grouped into four behavioural domains (control, sex, escape, and
style behaviours) commonly used to describe sexual offending behaviour (e.g., Grubin
et al., 2001; Woodhams, Grant, & Price, 2007).
For each behaviour, the relative frequency of occurrence in the offences of serial
vs. one-off offenders was assessed using chi-square analyses. Significant associations
15
are highlighted in bold in Tables 1-4 and effect sizes (phi) are reported for all
comparisons. A Phi coefficient between .2 and .3 is considered a small effect size, .3
and .5 is a medium effect size, and a value great than .5 is considered a large effect size
(Field, 2009). Each table also displays the frequencies of behaviours for all three subsets
of the serial rapes alongside the frequencies for the one-off rapes. In addition, for each
subset the frequency of each behaviour in the serial and one-off rapes combined is
reported.
When multiple chi-square analyses are run it is generally advisable to calculate a
correction to adjust for possible statistical errors. While a Bonferroni correction can be
used to reduce Type 1 errors (Pallant, 2007), it is argued to be too conservative,
resulting in Type 2 errors (Hochberg & Benjamini, 1990). The Benjamini-Hochberg has
been shown to have greater statistical power than the Bonferroni (Williams, Jones, &
Tukey, 1999) and is less conservative, therefore the Benjamini-Hochberg correction
(1995) was used to correct for Type 1 errors. The findings pre- and post-correction are
reported below.
Control Domain
The Control domain contained any behaviour that is deemed “necessary to
create and maintain an environment in which the crime can take place” (Grubin et al.,
2001, p. 14). This includes actions such as where the offence occurred (e.g., Alleyway),
how the offender approached the victim (e.g., Asked Victim for Help), and how the
offender kept the victim incapacitated (e.g., Gagged Victim).
A number of significant associations were found as well as several small and a
medium effect sizes (see Table 1). Three associations were significant across all three
subsets: serial rapists were significantly more likely to use solicitation as a method to
16
procure a victim (31.58% - 34.21%) than one-off rapists (2.00%) across all three
subsets. This finding remained after Benjamini-Hochberg correction and had an
associated medium sized effect. In contrast, across all three subsets, the serial rapists
were significantly less likely (5.26% - 7.89%) to obtain a victim by engaging them in
conversation than the one-off rapists (24.00%) representing a small effect size. The third
consistent, significant association was between the victim being bound and the rapists
being a serial offender, again representing a small effect size. Serial rapists were
significantly more likely to bind their victims (10.53% - 13.16%) than one-off rapists
(0.00%). The latter two findings did not remain following a Benjamini-Hochberg
correction.
In addition to these more consistent findings, there were other significant
associations within one or two subsets. The serial rapists were significantly more likely
than one-off rapists to commit the offence in a retail area (63.16 or 65.79% vs. 42.00%),
a parking area (21.05 or 23.68% vs. 6.00%), a street (68.42% or 76.32% vs. 46.00%) or
a wooded area (13.16% vs. 2.00%); however, the effect sizes for these associations were
small. In terms of other control behaviours used within the offence itself, serial rapists
were significantly more likely to gag their victims in the last offence subset than one-off
rapists (15.79% vs. 2.00%) but, again, only representing a small effect size. None of
these findings were significant following Benjamini-Hochberg correction.
Sex Domain
The Sexual behaviour domain contains the sexual acts that were part of the
offence, including physical acts and verbalisations about sexual acts (Woodhams, Grant
et al., 2007). For example, located within this domain are behaviours such as forms of
penetration (e.g. Offender Used Penis to Penetrate Vagina), levels of undress (e.g.,
17
Victim was Naked), who undressed whom (e.g., Offender Disrobed Victim), and sexual
comments made by the offender (e.g., Offender Discussed Sex Acts).
As can be seen from Table 2, the only behaviour consistent across all three
offence subsets was the finding that serial rapists more often forced their victims to
masturbate them than the one-off rapists (13.16 - 18.42% vs. 2.00%), however, this
represented a small effect size. This finding did not remain after statistical correction.
There were, again, some significant associations that were present in either one
or two subsets: serial offenders were more likely to fondle their victims (52.63% vs.
30.00%) and discuss the sex acts that they wanted the victim to perform (55.26% in
both the random and first subset vs. 30.00%). Also, they were more likely to remove the
victim’s clothing without damage in the first offence subset (21.05% vs. 6.00%). The
one-off rapists were more likely than the serial rapists to vaginally penetrate their
victims with their penises in the last offence subset (70.00% vs. 39.47%), representing a
medium effect size. Finally, they were also more likely than the serial rapists to disrobe
the victim themselves in the last offence subset (72.00% vs. 44.74%) (a small effect
size). However, none of these significant associations remained following statistical
correction.
Escape Domain
The Escape domain contains behaviours where the main function was to aid in
the offender’s escape from the scene and avoiding detection (Woodhams, Grant et al.,
2007). These include certain precautions (e.g., Wore Gloves) and verbal themes used by
the offenders (e.g., Warned Victim Not to Report Offence). As can be seen from Table
3, there was only one significant association found; within the first offence subset, the
serial rapists were more likely to wear gloves than the one-off offenders (15.79% vs.
18
2.00%) representing a small effect size. This was no longer significant following
Benjamini-Hochberg correction.
Style Domain
The Style domain contains behaviours that had no role in aiding the commission
of the offence, but were instead something that the offender chose to do (Grubin et al.,
2001). All of the behaviours in this domain were verbal themes (e.g., Apologises to
Victim). As can be seen from Table 4, there were no statistically significant associations
in this domain, nor were there any notable effect sizes.
DISCUSSION
This study aimed to address several of the limitations of existing research
contrasting the crime scene behaviour of serial and one-off rapists by focusing solely on
rapists who were strangers to their victims. Previous studies (LeBeau, 1987; Park et al.,
2008) had found differences in the offence behaviours of serial versus one-off rapists;
however it was unclear if these were due to inherent differences between serial or one-
off rapists or a result of the mixture of victim-offender relationships in the samples. In
this study, by holding the type of victim-offender relationship constant, these two
competing explanations could be disentangled. A further improvement on past study
designs was the extension of the sampling frame from just randomly sampling one
offence per series for comparison, to also comparing one-off offences with both the first
and last (known) offence from each series. This enabled the robustness of significant
associations to be assessed in terms of determining whether they were present across the
three subsets.
As with earlier studies, the majority of the behaviours included in the analysis
did not differ significantly in terms of frequency of occurrence between the serial and
19
the one-off rapists (Park et al., 2008). This poses a significant challenge to using crime
scene behaviour for the purpose of differentiating serial from one-off offences. As
explained above, police forces in England and Wales are expected to treat each stranger
rape reported to them as a potential serial offence. Information about crime scene
behaviour is available to the police early in an investigation and so would be an
effective means of supporting such investigative strategies if one-off and serial rapes
could be discriminated accurately. The findings reported here indicate that there are few
crime scene behaviours that could be used for this purpose.
Indeed, after conducting a Benjamini-Hochberg (1995) correction, the only
statistically significant association remaining was the more frequent use of solicitation
as a form of con-approach by serial rapists. As well as remaining significant following
statistical correction, this finding was also replicated across all three subsets of serial
offences. By implication, the victims of serial rapists were therefore significantly more
likely to be sex workers. There is evidence from other research that serial sex offenders
target sex workers, and that the offences against them are more violent (Silbert & Pines,
1982, 1984). Similarly, research on serial homicide has reported a tendency for serial
murderers to target prostitutes (Fox & Levin, 1998). The serial rapists’ apparent
preference for targeting sex workers has another advantage for their continued
offending: research shows that sex workers are reluctant to report rape and sexual
assaults to the Police (Silbert, 1981, as cited in Barnard, 1993; Sullivan, 2007), therefore
the offender can continue his offending without attracting police attention and hence is
able to become a serial rapist.
Closely tied to the use of the con of soliciting were the locations that were
chosen by the serial offenders for the offence, namely parking areas or on the street.
20
One of the most striking advantages for the offender of targeting a sex worker as a
victim is that he/she is likely to go alone with the offender to a more remote location.
There are also certain locations that sex workers choose, specifically to facilitate their
businesses, which are then sought by the offender (Douglas, Ressler, Burgess, &
Hartman, 1986). Most of the scene locations were places where there would not have
been many people around at the time of the offence. Several studies have shown that
sexual offenders weigh up the costs and benefits of where and when they commit their
offences, and that there is a rationale behind their actions (Beauregard & Leclerc, 2007;!
Beauregard, Rossmo, & Proulx, 2007).
The one-off offenders, in contrast to the serial offenders, were more likely to try
and con their victim by engaging her in a conversation. There was also a trend for one-
off rapists to more often offer the victim a ride. This was a very different style of
approach to the serial rapists, and would tend to be associated with a different type of
location than those frequented by sex workers and their clients. As such, the one-off
offenders were subsequently more likely to commit their offence indoors. Given that the
one-off offender would have had to talk face-to-face with the victim to either engage in
a conversation or offer a ride, the victim has more time during which to observe the
offender’s appearance and also possibly note other identifying information, such as their
type of vehicle or registration plate. It is possible that such an approach, while
successful in facilitating a completed or attempted rape, also aids in the apprehension of
the offender, which might curtail the offending of an individual before a series can
emerge.
Another behaviour within the control domain that might aid in the continued
offending of serial rapists was their more frequent use of binding the victim. Binding
21
the victim inhibits his/her ability to seek help from potential witnesses through physical
means and could potentially buy the offender more time in which to escape safely from
the scene before the victim could raise the alarm. Previous studies have suggested that
serial offenders are more “criminally sophisticated” and that this is what aids them in
avoiding detection (Park et al., 2008). Besides binding the victim, no other statistical
differences in the control behaviours of serial or one-off offenders were found, although
similar trends to Park et al.’s study were noted. For example, the elevated frequency of
gagging the victims by serial offenders could also not only facilitate the commission of
the offence itself by preventing disturbance by a third party, but also prolong the period
of time in which the offender can make a safe departure.
The other main area of difference between serial and one-off rapists that was
seen in this study was regarding the actions involved in the sexual component of the
crime. The serial rapists engaged in more sexual acts than the one-off rapists;
specifically fondling the victim, forcing the victim to masturbate the offender, and
discussing sex acts with the victim during the crime. These are very different findings
than those reported by Park et al. (2008) where similar behaviours to these were more
often associated with the one-off rapists. Park et al. suggested that these verbal themes,
especially communications with the victim about the offender’s fantasy and sex acts,
helped investigators apprehend the offender. However, this study found that the serial
offenders were more likely to talk about sex acts during the offence.
In their study of stranger rapists, Canter, Bennell, Alison and Reddy (2003)
found four styles of behaviour within stranger rapes; control, theft, involvement and
hostility, which have been previously reported in other studies of sexual offences (see
Canter et al., 2003 for a review). These styles are proposed to affect the way a rapist
22
will relate to his victim, for example, rapists adopting an involvement style treat their
victim as a reactive individual (as a person), whereas more controlling rapists are
proposed to treat their victim as an object. Canter et al. suggest that the style adopted
will result in different themes of verbal communication between rapist and victim. This
seems to be reflected in the sample for this study, with the serial rapists appearing to
adopt a more involved style of verbal communication, although it should be noted that
some of their other behaviours, such as forcing the victim to masturbate them and
binding the victim, would represent a more hostile or controlling style, respectively.
Limitations
There were some limitations to the study that would necessitate caution before
applying these findings to all stranger rapists. It cannot be guaranteed that the one-off
rapists included in this study have only committed the one offence, since it is not
possible to be certain that the offences included in a study are the only ones the offender
has committed. As such, some one-off rapists in this study may instead by serial rapists.
In addition, in the absence of definitive DNA evidence, we cannot be completely certain
of the “serial” status given to some offenders, due to the possibility of miscarriages of
justice. These are limitations common to studies of this nature (Mokros & Alison, 2002;
Santtila et al., 2005) which must rely on conviction to categorise the offenders in this
way. Such errors in classification could mask potential differences in behaviour between
one-off and serial rapists.
Similar to other studies in the area, we were unable to run a statistical cross-
validation of the findings. Due to the size of the sample and the number of variables
being considered, it was impossible to run a leave-one out logistic regression without
violating the case: variable assumption (Peduzzi, Concato, Kemper, Holford, &
23
Feinstein, 1996). However, this study improved on previous study designs by
comparing the findings across three different sub-samples (i.e., constituting one-off
offences vs. first serial offence, one-off offences vs. last serial offence, and one-off
offences vs. a randomly selected offence from each series) to determine if each finding
was consistent.
The data that were utilised in this study were based only on offences for which
there was a conviction; therefore, the sample cannot be considered representative of all
stranger rapes. This is because it is well established that rapes which are prosecuted and
result in a conviction more closely reflect rape myths in our societies (Frazier & Haney,
1996; Harris & Grace, 1999) and may contain different offence behaviours to those
committed by offenders that have not been apprehended (Bennell & Canter 2002;
Woodhams, Hollin et al., 2007). We cannot, therefore, be sure that our findings will
transfer to crimes that remain unsolved, the type of crime to which investigators would
apply the findings in practice. However, as noted above, the methodology required to
compare apparent one-off with serial rapists necessitates it being “known” which
offenders have committed just one offence or several; therefore it would be very
difficult to overcome this limitation.
Conclusion
It has long been a policing priority to target prolific offenders; however, in the
current fiscal climate it is even more advantageous to be able to target limited police
resources in this way. In 2012, in England and Wales, police forces were advised to
initially consider every stranger rape part of a potential series (HMIC & HMICPS,
2012). This is potentially a costly and time-consuming exercise which could be aided if
it were possible to distinguish serial from one-off stranger rapists on the basis of crime
24
scene behaviour, reports of which are often available in the initial stages of a police
investigation. This study aimed to contribute to a very small set of existing studies
which have tried to empirically establish means of differentiating serial from one-off
rapists using crime scene behaviour by addressing limitations in past study design. The
findings of this present study suggest that there may be a limited number of differences
in the offence behaviour displayed by a one-off stranger rapist and a serial stranger
rapist, particularly in terms of the type of victim targeted, the locations chosen for the
offence, methods of control and the sexual acts that they force upon the victim. The
only key distinction between the two types of offenders was the choice of sex workers
as potential targets by serial offenders, which supports previous studies. However, what
was most striking was the similarity in crime scene behaviour between these two types
of rapist, meaning that it would be difficult in practice to accurately differentiate serial
from one-off rapes, at least on the basis of the crime scene behaviors investigated here.
25
REFERENCES
Barnard, M. (1993). Violence and vulnerability: Conditions of work for streetworking
prostitutes. Sociology of Health and Illness, 15, 683-705. doi:10.1111/1467-
9566.ep11434434
Beauregard, E., & Leclerc, B. (2007). An application of the rational choice approach to
the offending process of sex offenders: A closer look at the decision-making.
Sexual Abuse: A Journal of Research and Treatment, 19, 115-133.
doi:10.1177/0093854807300851
Beauregard, E., Rossmo, K., & Proulx, J. (2007). A descriptive model of the hunting
process of serial sex offenders: A rational choice perspective. Journal of Family
Violence, 22, 449-463. doi:10.1177/0093854807300851
Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical
and powerful approach to multiple testing. Journal of the Royal Statistical
Society. Series B (Methodological), 289-300. Retrieved from
www.math.tau.ac.il/~ybenja/MyPapers/benjamini_hochberg1995.pdf
Bennell, C., & Canter, D. (2002). Linking commercial burglaries by modus operandi:
Tests using regression and ROC analysis. Science and Justice, 42, 1-12.
doi:10.1016/S1355-0306(02)71820-0.
Bennell, C., Mugford, R., Ellingwood, H., & Woodhams, J. (in press). Linking crimes
using behavioural clues: Current levels of linking accuracy and strategies for
moving forward. Journal of Investigative Psychology and Offender Profiling.
Doi: 10.1002/jip.1395
26
Bownes, I. T., O’Gorman, E. C., & Sayers, A. (1991). Rape – a comparison of stranger
and acquaintance assaults. Medicine, Science and the Law, 31, 102-109.
doi:10.1111/j.1600-0447.1991.tb05507.x
Canter, D., Bennell, C., Alison, L. J., and Reddy, S. (2003). Differentiating sex
offences: A behaviorally based thematic classification of stranger rapes.
Behavioral Sciences and the Law, 21, 157-174. doi: 10.1002/bsl.526
Douglas, J. E., Ressler, R. K., Burgess, A. W., & Hartman, C. R (1986). Criminal
profiling from crime scene analysis. Behavioral Sciences and the Law, 4, 401-
421. Doi:10.1002/bsl.2370040405
Field, A. (2009). Discovering statistics using SPSS (3rd ed.). London, England: SAGE.
Fox, J. A., & Levin, J. (1998). Multiple homicide: Patterns of serial and mass murder.
Crime and Justice, 23, 407-455. Retrieved from
http://www.jstor.org/stable/1147545
Frazier, P. A., & Haney, B. (1996). Sexual assault cases in the legal system: Police,
prosecutor, and victim perspectives. Law and Human Behavior, 20, 607-628.
doi: 10.1007/BF01499234
Grubin, D., Kelly, P., & Brunsdon, C. (2001). Linking serious sexual assaults through
behaviour: HORS 215. Home Office Research Development and Statistics
Directorate. Retrieved from http://library.npia.police.uk/docs/hors/hors215.pdf
Harris, J., & Grace, S. (1999). A question of evidence? Investigating and prosecuting
rape in the 1990s. Home Office Research Study 196. Retrieved from
http://rds.homeoffice.gov.uk/rds/pdfs/hors196.pdf
Her Majesty’s Inspectorate of Constabulary (HMIC) and Her Majesty’s Inspectorate of
the Crown Prosecution Service (HMICPS) (2012). Forging the links: Rape
27
investigation and prosecution. Retrieved from
http://www.hmic.gov.uk/media/forging-the-links-rape-investigation-and-
prosecution-20120228.pdf
Hochberg, Y., & Benjamini, Y. (1990). More powerful procedures for multiple
significance testing. Statistics in medicine, 9(7), 811-818. doi:!
10.1002/sim.4780090710
Koss, M. P., Dinero, T. E., Siebel, C. A., & Cox, S. L. (1988). Stranger and
acquaintance rape: Are there differences in the victim’s experience? Psychology
of Women Quarterly, 12, 1-24. doi:10.1111/j.1471-6402.1988.tb00924.x
Kraemer, G. W., Lord, W. D., & Heilbrun, K. (2004). Comparing single and serial
homicide offenses. Behavioral Sciences and the Law, 22, 325-343. doi:
10.1002/bsl.581
LeBeau, J. L. (1987). The journey to rape: Geographic distance and the rapist's method
of approaching the victim. Journal of Police Science and Administration, 15,
129-136. Retrieved from
http://heinonline.org/HOL/Page?collection=journals&handle=hein.journals/jclc7
8&type=Image&id=321
Mokros, A., & Alison, L. J. (2002). Is offender profiling possible? Testing the predicted
homology of crime scene actions and background characteristics in a sample of
rapists. Legal and Criminological Psychology, 7, 25-43.
doi:10.1348/135532502168360
Muram, D., Hostetler, B. R., Jones, C. E., & Speck, P. M. (1995). Adolescent victims of
sexual assault. Journal of Adolescent Health, 17, 372–375.
Pallant, J. (2007). SPSS survival manual (3rd ed.). Maidenhead: Open University Press.
28
Park, J., Schlesinger, L. B., Pinizzotto, A. J., & Davis, E. F. (2008). Serial and single-
victim rapists: Differences in crime-scene violence, interpersonal involvement,
and criminal sophistication. Behavioral Sciences and the Law, 26, 227-237.
doi:10.1002/bsl.804
Peduzzi, P., Concato, J., Kemper, E., Holford, T. R., & Feinstein, A. R. (1996). A
simulation study of the number of events per variable in logistic regression
analysis. Journal of clinical epidemiology, 49(12), 1373-1379. doi:
10.1016/S0895-4356(96)00236-3
Rainbow, L. (in press). A practitioner’s perspective: Theory, practice and research. In J.
Woodhams & C. Bennell (Eds.), Crime linkage: Theory, practice and research.
Boca Raton, FL: CRC Press.
Salfati, C. G. (2003). Offender interaction with victims in homicide: A
multidimensional analysis of frequencies in crime scene behaviors. Journal of
Interpersonal Violence, 18, 490-512. doi:10.1177/0886260503251069
Salfati, C. G., & Bateman, A. L. (2005). Serial homicide: An investigation of
behavioural consistency. Journal of Investigative Psychology and Offender
Profiling, 2, 121-144. doi:10.1002/jip.27
Santtila, P., Junkkila, J., & Sandabba, N. K. (2005). Behavioural linking of stranger
rapes. Journal of Investigative Psychology and Offender Profiling, 2, 87-103.
doi:10.1002/jip.26
Sexual Offences Act (2003). Retrieved from
http://www.legislation.gov.uk/ukpga/2003/42/pdfs/ukpga_20030042_en.pdf
Silbert, M. H., & Pines, A. M. (1982). Entrance into prostitution. Youth and Society,
13, 471-500. doi:10.1177/0044118X82013004005
29
Silbert, M. H., & Pines, A. M. (1984). Pornography and sexual abuse of women. Sex
Roles, 10, 857-868. doi:10.1007/BF00288509
Sorochinski, M., & Salfati, C. G. (2010). The consistency of inconsistency in serial
homicide: patterns of behavioural change across series. Journal of Investigative
Psychology and Offender Profiling, 7, 109-136.
Sullivan, B. (2007). Rape, prostitution and consent. Australian and New Zealand
Journal of Criminology, 40, 127-142. doi: 10.1375/acri.40.2.127
Williams, V. S., Jones, L. V., & Tukey, J. W. (1999). Controlling error in multiple
comparisons, with examples from state-to-state differences in educational
achievement. Journal of Educational and Behavioral Statistics, 24(1), 42-69.
doi: 10.3102/10769986024001042
Woodhams, J., Grant, T. D., & Price, A. R. G. (2007). From marine ecology to crime
analysis: Improving the detection of serial sexual offences using a taxonomic
similarity measure. Journal of Investigative Psychology and Offender Profiling,
4, 17-27. doi: 10.1002/jip.55
Woodhams, J., Hollin, C. R., & Bull, R. (2007). The psychology of linking crimes: A
review of the evidence. Legal and Criminological Psychology, 12, 233-249. doi:
10.1348/135532506X118631
Woodhams, J., & Labuschagne, G. (2012). A test of case linkage principles with solved
and unsolved serial rapes. Journal of Police and Criminal Psychology, 27, 85-
98.
Woodhams, J., & Toye, K. (2007). An empirical test of the assumptions of case linkage
and offender profiling with serial commercial robberies. Psychology, Public
Policy, and Law, 13, 59-85. doi:10.1037/1076-8971.13.1.59
30
Table 1
Incidence of Behaviours in the Control Domain
Non-
Serial
% of
Offences
(N=50)
Random
First
Last
Behaviour
Serial
% of
Offences
(N=38)
Phi
All
Rapists
% of
Offences
(N=88)
Serial
% of
Offences
(N=38)
Phi
All
Rapists
% of
Offences
(N=88)
Serial
% of
Offences
(N=38)
Phi
All
Rapists
% of
Offences
(N=88)
Indoors
42.00
21.05
-0.22*
32.95
21.05
-0.22*
32.95
23.68
-0.19
34.09
Outdoors
76.00
86.84
0.14
80.68
89.47
0.17
81.82
86.84
0.14
80.68
Industrial Area
4.00
7.89
0.083
5.68
10.53
0.13
6.82
7.89
0.083
5.68
Retail Area
42.00
63.16
0.21*
51.14
65.79
0.24*
52.27
57.89
0.16
48.86
Residential Area
84.00
71.05
-0.16
78.41
65.79
-0.21*
76.14
81.58
-0.032
82.95
Rural Area
8.00
13.16
0.084
10.23
7.89
-0.002
7.95
10.53
0.044
9.09
Living Quarters
40.00
23.68
-0.17
32.95
26.32
-0.14
34.09
23.68
-0.17
32.95
In a Vehicle
14.00
15.79
0.025
14.77
15.79
0.025
14.77
18.42
0.060
15.91
Entertainment Area
20.00
7.89
-0.17
14.77
7.89
-0.17
14.77
2.63
-0.26*
12.50
Public Area
10.00
15.79
0.087
12.50
13.16
0.049
11.36
13.16
0.049
11.36
Parking Area
6.00
23.68
0.26*
13.64
15.79
0.16
10.23
21.05
0.23*
12.50
Alleyway
8.00
5.26
-0.054
6.82
10.53
0.044
9.09
7.89
-0.002
7.95
Wooded Area
2.00
13.16
0.22*
6.82
7.89
0.14
4.55
7.89
0.14
4.55
Access Path
14.00
18.42
0.060
15.91
15.79
0.025
14.77
15.79
0.025
14.77
Street
46.00
68.42
0.22*
55.68
76.32
0.31*
59.09
57.89
0.12
51.14
Main Road
24.00
18.42
-0.067
21.59
21.05
-0.035
22.73
18.42
-0.067
21.59
Park
10.00
15.79
0.087
12.50
10.53
0.009
10.23
10.53
0.009
10.23
Asked Victim for Help
8.00
2.63
-0.11
5.68
5.26
-0.054
6.82
5.26
-0.054
6.82
Solicited Victim
2.00
31.58
0.41**
14.77
31.58
0.41**
14.77
34.21
0.44**
15.91
Offered Ride to Victim
10.00
.00
-0.21*
5.68
.00
-0.21*
5.68
2.63
-0.14
6.82
Engaged Victim in Conversation
24.00
7.89
-0.21*
17.05
5.26
-0.25*
15.91
7.89
-0.21*
17.05
Threatened Victim upon Approach
8.00
2.63
-0.11
5.68
2.63
-0.11
5.68
2.63
-0.11
5.68
Snuck Up on Victim
28.00
44.74
0.17
35.23
42.11
0.15
34.09
36.84
0.094
31.82
Victim was Sleeping when Approached
12.00
7.89
-0.067
10.23
7.89
-0.067
10.23
7.89
-0.067
10.23
Gagged Victim
2.00
10.53
0.18
5.68
10.53
0.18
5.68
15.79
0.25*
7.95
Covered Victim’s Mouth
24.00
34.21
0.11
28.41
28.95
0.056
26.14
34.21
0.11
28.41
31
Bound the Victim
.00
10.53
0.25*
4.55
10.53
0.25*
4.55
13.16
.28*
5.68
Verbally Threatened Victim
54.00
55.26
0.013
54.55
52.63
-0.014
53.41
52.63
-0.014
53.41
Attempted to Reassure Victim
24.00
31.58
0.084
27.27
26.32
0.026
25.00
39.47
0.17
30.68
Upon Resistance used Some Violence
16.00
28.95
0.16
21.59
21.05
0.065
18.18
31.58
0.18
22.73
Without Resistance used Some Violence
28.00
28.95
0.010
28.41
23.68
-0.049
26.14
31.58
0.039
29.55
Threatened to Use Weapon, but Never Seen
10.00
13.16
0.049
11.36
10.53
0.009
10.23
10.53
0.009
10.23
Displayed Weapon but did not Use
20.00
10.53
-0.13
15.91
13.16
-0.090
17.05
21.05
0.013
20.45
Weapon was Used
6.00
7.89
0.037
6.82
7.89
0.037
6.82
10.53
0.083
7.95
Weapon was Brought By Rapist
20.00
21.05
0.013
20.45
26.32
0.075
22.73
28.95
0.10
23.86
Weapon was a Stabbing Instrument
30.00
28.95
-0.011
29.55
23.68
-0.070
27.27
36.84
0.072
32.95
Note. * p < .05. ** p < .001
32
Table 2
Incidence of Behaviours in the Sex Domain
Non-
Serial
% of
Offences
(N=50)
Random
First
Last
Behaviour
Serial
% of
Offences
(N=38)
Phi
All
Rapists
% of
Offences
(N=88)
Serial
% of
Offences
(N=38)
Phi
All
Rapists
% of
Offences
(N=88)
Serial
% of
Offences
(N=38)
Phi
All
Rapists
% of
Offences
(N=88)
Rapist Kissed Victim’s Face
46.00
28.95
-0.17
38.64
18.42
-0.29*
34.09
18.42
-0.29*
34.09
Rapist Kissed Victim’s Chest
14.00
7.89
-0.095
11.36
10.53
-0.052
12.50
7.89
-0.095
11.36
Rapist Kissed Victim on Other Area
8.00
5.26
-0.054
6.82
.00
-0.19
4.55
2.63
-0.11
5.68
Rapist Fondled Victim
30.00
52.63
0.23*
39.77
44.74
0.15
36.4
42.11
0.13
35.23
Rapist Masturbated
10.00
13.16
0.049
11.36
10.53
0.009
10.23
15.79
0.087
12.50
Rapist Performed Oral Sex on Victim
8.00
10.53
0.044
9.09
2.63
-0.11
5.68
7.89
-0.002
7.95
Rapist Used Hand to Penetrate Vagina
18.00
28.95
0.13
22.73
28.95
0.13
22.73
26.32
0.10
21.59
Rapist Used Penis to Penetrate Vagina
70.00
60.53
-0.099
65.91
63.16
-0.072
67.05
39.47
-0.31*
56.82
Rapist Penetrate Vagina from Behind
20.00
23.68
0.044
21.59
23.68
0.044
21.59
13.16
-0.09
17.0
Rapist Used Hand to Penetrate Anus
2.00
7.89
0.14
4.55
5.26
0.089
3.41
2.63
0.021
2.27
Rapist Used Penis to Penetrate Anus
18.00
28.95
0.13
22.73
23.68
0.070
20.45
15.79
-0.029
17.05
Victim Kissed Rapist’s Face
8.00
5.26
-0.054
6.82
2.63
-0.11
5.68
.00
-0.19
4.55
Victim Masturbated Rapist
2.00
18.42
0.28*
9.09
13.16
0.22*
6.82
13.16
0.22*
6.82
Victim Performed Fellatio
28.00
34.21
0.067
30.68
44.73
0.17
35.23
28.95
0.01
28.41
Rapist was Naked
10.00
7.89
-0.036
9.09
5.26
-0.087
7.95
5.26
-0.087
7.95
Victim was Naked
16.00
23.68
0.096
19.32
18.42
0.032
17.05
13.16
-0.040
14.77
Victim was Partially Disrobed
40.00
36.84
-0.032
38.64
34.21
-0.059
37.50
34.21
-0.059
37.50
Victim’s Clothing was Moved to Expose
32.00
21.05
-0.12
27.27
36.84
0.051
34.09
15.79
-0.19
25.00
Rapist Disrobed Victim
72.00
63.16
-0.094
68.18
71.05
-0.010
71.59
44.74
-0.28*
60.23
Victim Disrobed Herself
20.00
31.58
0.13
25.00
28.95
0.10
23.86
21.05
0.013
20.45
Rapist Disrobed Himself
76.00
76.32
0.004
76.14
73.68
-0.026
75.00
60.53
-0.17
69.32
Clothing was Removed without Damage
6.00
18.42
0.19
11.36
21.05
0.23*
12.50
10.53
0.083
7.95
Clothing Removed was Torn Off
16.00
15.79
-0.003
15.91
15.79
-0.003
15.91
18.42
0.032
17.05
Rapist Discussed Sex Acts
30.00
55.26
0.25*
40.91
55.26
0.25*
40.91
42.11
0.13
35.23
Note. * p < .05. ** p< .01
33
Table 3
Incidence of Behaviours in the Escape Domain
Non-
Serial
% of
Offences
(N=50)
Random
First
Last
Behaviour
Serial
% of
Offences
(N=38)
Phi
All
Rapists
% of
Offences
(N=88)
Serial
% of
Offences
(N=38)
Phi
All
Rapists
% of
Offences
(N=88)
Serial
% of
Offences
(N=38)
Phi
All
Rapists
% of
Offences
(N=88)
Wore Gloves
2.00
10.53
0.18
5.68
15.79
0.25*
7.95
10.53
0.18
5.68
Covered Victim’s Eyes
6.00
10.53
0.083
7.95
10.53
0.083
7.95
13.16
0.12
9.09
Told Victim ‘Not to Look’
12.00
26.32
0.18
18.18
18.42
0.090
14.77
23.68
0.15
17.05
Used a Condom
6.00
5.26
-0.015
5.68
7.89
0.037
6.82
2.63
-0.080
4.55
Used a False Name
10.00
7.89
-0.036
9.09
7.89
-0.036
9.09
5.26
-0.087
7.95
Warned Victim Not to Report Offence
18.00
26.32
0.10
21.59
21.05
0.038
19.32
21.05
0.038
19.32
Instructions to Ensure His Safe Escape
18.00
15.79
-0.029
17.05
23.68
0.070
20.45
10.53
-0.10
14.77
Makes Reference to Justice System
6.00
18.42
0.19
11.36
18.42
0.19
11.36
7.89
0.037
6.82
Note. * p < .05
34
Table 4
Incidence of Behaviours in the Style Domain
Non-
Serial
% of
Offences
(N=50)
Random
First
Last
Behaviour
Serial
% of
Offences
(N=38)
Phi
All
Offenders
% of
Offences
(N=88)
Serial
% of
Offences
(N=38)
Phi
All
Rapists
% of
Offences
(N=88)
Serial
% of
Offences
(N=38)
Phi
All
Rapists
% of
Offences
(N=88)
Discusses Victim’s Sex Practices
6.00
7.89
0.037
6.82
.00
-0.16
3.41
10.53
0.083
7.95
Orders Victim to Participate
8.00
10.53
0.044
9.09
7.89
-0.002
7.95
7.89
-0.002
7.95
Uses Abusive Language
14.00
23.68
0.12
18.18
21.05
0.093
17.05
15.79
0.025
14.77
Expresses Curiosity About Victim
18.00
21.05
0.038
19.32
15.79
-0.029
17.05
23.68
0.070
20.45
Rapists Discloses Information about Self
38.00
44.74
0.068
40.91
39.47
0.015
38.64
31.58
-0.067
35.23
Tries to Ingratiating Himself with Victim
12.00
13.16
0.017
12.50
10.53
-0.023
11.36
18.42
0.090
14.77
Compliments the Victim
10.00
13.16
0.049
11.36
10.533
0.009
10.23
7.89
-0.036
9.09
Apologises to Victim
8.00
15.79
0.12
11.36
13.16
0.084
10.23
13.16
0.084
10.23
Attempts to Prolong Relationship
6.00
7.89
0.037
6.82
7.89
0.037
6.82
2.63
-0.080
4.55
Displays Personal Knowledge of Victim
4.00
7.89
0.083
5.68
5.26
0.030
4.55
5.26
0.030
4.55
Says Victim Feels Enjoyment in Offence
14.00
15.79
0.025
14.77
15.79
0.025
14.77
7.89
-0.095
11.36
Justifies Actions
14.00
13.16
-0.012
13.64
13.16
-0.012
13.64
7.89
-0.095
11.36
!