Table 10.1 The Newcastle murder series
Crime 1 Crime 2 Crime 3 Crime 4
Murder and rape Murder and rape Murder Murder
Occurred at night Occurred at night Occurred at night Occurred at night
Female rape victim threatened with knife Female rape victim threatened with knife
Male victim killed with rock (from scene) Male victim killed with rock (from scene) Male victim killed with rock (from scene) Male victim killed with rock (from scene)
Blitz (sudden) attack Blitz (sudden) attack Blitz (sudden) attack Blitz (sudden) attack
Rape of female Rape of female Male victim found undressed (potential female rape victim?)
No property stolen Property stolen (telephones, purse) No property stolen
CASE STUDY 10.2 STEVEN FORTIN (HAZELWOOD & WARREN, 2003)
Steven Fortin was tried for the murder of Melissa Padilla on several occasions (Meyer, 2007). He was finally convicted and sentenced to life imprisonment without parole for her murder in June 2010. In his first trial retired FBI profiler, Roy Hazelwood, presented evidence of behavioural similarity between the actions of the perpetrator who murdered Melissa Padilla and the behaviour displayed during a sexual offence in the state of Maine for which Steven Fortin had been convicted. Fifteen similarities in modus operandi (behaviour necessary to commit the offence) were identified, which included characteristics of the victims, the environment in which they were attacked, the use of force and the injuries they sustained as a result. In addition, five ritual behaviours, which reflect motive, were highlighted as being common to both crimes: both victims were brutally beaten about the face, manually strangled from the front, raped anally, and bitten on their chin and breast. Roy Hazelwood concluded that the 15 shared modus operandi behaviours alongside the five shared ritual behaviours formed a unique signature across the two crimes.
Hazelwood also considered behavioural differences between the crimes that might have suggested the offences had been committed by different people. In this case, Hazelwood identified 11 differences, which included the locations of the offences, the time/weekday on which they occurred, and physical characteristics of the victims (e.g. their height and weight). He explains in his article with Janet Warren, however, that these differences can be attributed to the opportunistic nature of the crimes and the attack in Maine being disturbed by a third party, and therefore they “pose no threat to the signature opinion” (Hazelwood & Warren, 2003, p. 316). Hazelwood, therefore, concluded that the two crimes were committed by the same person.
While the case studies show how crime linkage has been used in court proceedings, it is more commonly used in an investigative context. Within this context, there are several different scenarios in which crime linkage might be conducted (see Figure 10.1). An analyst might be provided with a specific index crime (that is either solved or unsolved) and then asked to find other crimes that may be committed by the same person from within police crime databases (scenario 1 in Figure 10.1, where the orange dots represent crimes that are linked to the index crime). Alternatively, an analyst might be presented with a pre-defined group of offences (e.g. 10 rapes) and asked to decide whether the crimes are committed by the same person or not (scenario 3). These two scenarios are described as reactive crime linkage because the analyst is responding to a specific request to link crimes (Woodhams, Bull, & Hollin, 2007). In scenario 2, however, the analyst proactively searches for potentially linked offences amongst a larger database (with the orange dots in Figure 10.1 representing one crime series committed by a particular offender and the purple dots a second series committed by a different offender).
FIGURE 10.1 Three different crime linkage scenarios
PHOTO 10.1 Information about how offenders behave at crime scenes, whether that comes from witness/victim accounts or CCTV footage, is used by crime analysts to conduct crime linkage.
Source: © ERproductions Ltd/Getty Images
Where the crime linkage practitioner suspects they have identified two or more crimes that might have been committed by the same offender they conduct a detailed analysis of the behaviour displayed at each crime scene. They identify behaviours that are common to the crimes and they consider how the behaviour displayed might differ. Where repeated behaviours are identified, the practitioner must also consider how often these behaviours are displayed by all offenders. For example, if within a series of stranger rapes the offender always tied up the victim, the practitioner would need to consider how often this behaviour occurs in all stranger rapes. This relates to the assumption of crime linkage that offenders must be relatively distinctive in their offending styles. If a behaviour is relatively rare and it occurs across all offences in the potential series the practitioner can be more confident that these crimes were committed by the same offender. The practitioner would, however, normally take into account multiple behaviours when arriving at such a decision.
The practitioner might also consider the geographical proximity of the offences, which is thought to be a good indicator of whether crimes are committed by the same offender because offenders tend to commit their offences within well-defined areas with which they are familiar (Bennell & Jones, 2005). If the geographical areas in which offenders commit their crimes don’t overlap a great deal, crimes by the same offender should be located close together as well as being a distance away from the crimes of different offenders.
In some countries, there are large-scale police databases that hold information about the crime scene behaviour displayed by the offender, geographical location and time/date of each crime in the database. These can be used to support crime linkage analysis since the database can be searched for crimes that are close together in time and space, and similar in crime scene behaviour.
10.2.1 Evidence-based Practice
The use of crime linkage analysis to support decision-making in police investigations, as well as its use in legal proceedings, means that the implications of incorrectly linking crimes are significant (Grubin et al., 2001). Psychologists’ professional practice guidelines emphasise the importance of engaging in evidence-based practice and using the best research evidence (e.g. American Psychological Association, 2014; Health and Care Professions Council, 2010). It is therefore important that the principles underpinning crime linkage are demonstrated to be empirically supported and that research findings continue to inform how crime linkage is conducted in practice. As stated above, crime linkage rests on two assumptions; those of behavioural consistency and distinctiveness. Research studies have been accumulating since the 1970s (Green, Booth, & Biderman, 1976) to the present day on the consistency and distinctiveness of criminal behaviour and a considerably larger body of research spanning many decades exists regarding the consistency and distinctiveness of non-criminal behaviour (Pervin, 2002).
10.2.2 Empirical Evidence
In terms of the consistency and distinctiveness of criminal behaviour, a range of crime types have been investigated by researchers including murder (e.g. Bateman & Salfati, 2007), sexual offences (e.g. Bennell, et al., 2009), commercial and personal robbery (Burrell, Bull, & Bond, 2012; Woodhams & Toye, 2007), commercial and residential burglary (e.g. Bennell & Jones, 2005), car-theft (e.g. Tonkin, Grant, & Bond, 2008), and arson (e.g. Santtila, Fritzon, & Tamelander, 2004). Broadly speaking, studies of acquisitive crime have found only limited evidence for consistency and distinctiveness in crime scene behaviour (with some notable exceptions discussed below). In contrast, studies of interpersonal types of crime report being able to identify, with a good degree of accuracy, crimes from the same series using similarity in crime scene behaviour. This provides support for the two underlying assumptions of crime linkage with interpersonal types of crime. This doesn’t mean, however, that offenders who commit acquisitive forms of crime are not consistent and distinctive too; it may simply be that police records of crime scene behaviour in acquisitive crime do not contain enough accurate and reliable information to identify meaningful patterns of consis
tency and distinctiveness.
Nevertheless, while the majority of offender behaviour has not demonstrated high levels of consistency and distinctiveness in acquisitive crimes, there are notable exceptions. Research has demonstrated, for example, that the geographical distance and number of days between offences is shorter for crimes committed by the same person than crimes committed by different people, which means that the so-called inter-crime distance and temporal proximity might be used to facilitate crime linkage (e.g., Davies, Tonkin, Bull, & Bond, 2012; Goodwill & Alison, 2006; Markson, Woodhams, & Bond, 2010). The inter-crime distance is also good at predicting whether crimes belong to the same series for interpersonal forms of crime, such as commercial and personal robbery (Burrell et al., 2012; Woodhams & Toye, 2007). However, the effectiveness of geographical proximity for linking varies for different crimes. For example, Bennell and Jones (2005) found it to be more effective for residential burglaries than commercial burglaries. As reported above, geographical proximity will best identify series when the geographical areas in which criminals offend overlap the least (the distinctiveness element of crime linkage). As explained by Bennell and Jones, commercial properties tend to be clustered together, therefore the spaces in which commercial burglars commit their crimes overlap to a greater degree.
Returning to the issue of evidence-based practice, it is too early in the research process to make firm recommendations for conducting crime linkage. However, what researchers have found so far suggests that for acquisitive forms of crime, geographical proximity (and to a lesser extent, temporal proximity) should be used for deciding if crimes form part of the same series. The research findings for interpersonal forms of crime, at least in terms of robbery and sexual offences, are more promising in terms of using crime scene behaviour to identify crime series. There are, however, important limitations to this research that should be considered.
10.2.3 Limitations of the Research on Crime Linkage
While it is not possible to give a comprehensive review of limitations (see Tonkin, 2015, for more detail), there are some important limitations of the literature that will be briefly described. One primary limitation is that much of the crime linkage literature has relied on samples of solved crime. This is problematic because the crimes may have been solved in the first place due to having been committed in a highly consistent and/or distinctive manner (e.g., Bennell & Canter, 2002; Sorochinski & Salfati, 2010). Consequently, research that relies on solved offences might overestimate the likelihood of crime linkage working successfully in practice. While research has begun to test the assumptions of crime linkage with unsolved offences (e.g. see Woodhams & Labuschagne, 2012, and www.crimelinkage.org), much more research is needed in this area.
A second limitation is that much research has focused on examining crime linkage within specific crime types (e.g. is it possible to link two residential burglaries committed by the same person?). This is despite the fact that many offenders (particularly the most prolific) do not restrict themselves to committing just one type of offence (Farrington, Snyder, & Finnegan, 1988; Piquero, Farrington, & Blumstein, 2007). Fortunately, however, recent investigations have begun to test whether offenders are consistent across crimes of different types too (Tonkin & Woodhams, 2017; Tonkin, Woodhams, Bull, & Bond, 2012; Tonkin, Woodhams, Bull, Bond, & Palmer, 2011). In the future it may therefore be possible to identify serial offenders who are versatile in their offending behaviour.
Third, more research is needed to test crime linkage in practice, including how analysts identify behavioural consistency and distinctiveness, what factors they use to link crimes and whether their decision-making can be improved using statistically-based decision-support tools. Future work of this nature is vital if the research literature on crime linkage is to have a meaningful impact on police practice.
10.2.4 Conclusions on Crime Linkage
Identifying crime series has several advantages to the police and can benefit society. However, if a crime linkage practitioner gives the police or the courts inaccurate advice the consequences are significant. It is therefore important that crime linkage is assessed and shown to be a valid practice. Studies testing the principles of crime linkage are accumulating and suggest that accurate crime linkage should be possible if appropriate indicators of series membership are attended to. What this area of forensic psychology is lacking at present, however, is a real-world assessment of crime linkage whereby predictions about series membership are made and systematically followed up to determine accuracy. Until studies of this nature are conducted we will not know the true contribution that crime linkage research can make to the investigation of prolific serial offenders.
10.3 OFFENDER PROFILING
In contrast to crime linkage, offender profiling is the topic of a considerable number of scholarly articles and book chapters as well as forms of popular media. It is a contentious area and strong opinions exist. As you will see, “offender profiling” does not refer to a uniform practice and how it is conducted has evolved over time and continues to evolve. In researching it, we often have to use data that are held and collated by police forces, which presents its own difficulties in terms of data quality (Alison, Snook, & Stein, 2001). If you are considering studying this field, the sheer volume of literature to get to grips with can be quite daunting. While it is an exciting area to research and study, it is also challenging.
Quite a lot of the debate within this field relates to whether offender profiling is, or should be, an art or a science. As psychologists, our practice should be evidence-based and therefore empirical research into the underpinnings of offender profiling is crucial. Before considering this empirical research, the history of offender profiling and some of the major ways in which it has developed will be briefly outlined.
The roots of offender profiling have been traced back to Dr. Thomas Bond who provided the UK police with an opinion regarding Jack the Ripper’s physical appearance, likely dress, state of mind and motivation for a series of murders in London in 1880 (Canter, 2004). However, as Canter (2004) explains, it was through the activities of the Federal Bureau of Investigation (FBI) in the United States that offender profiling gained its fame. The use of offender profiling (and BIA) to guide police investigations has since spread to other countries including Australia, Canada, Finland, Germany, Sweden, the Netherlands and Singapore (Alison, Goodwill, Almond, van der Heuvel, & Winter, 2010; Davis & Bennett, 2006; Snook, Cullen, Bennell, Taylor, & Gendreau, 2008).
Offender profilers come from a variety of professional backgrounds. A study in 1995 in the UK of 184 instances of profiling advice reported that the advice had been sought from forensic, clinical and academic psychologists, forensic and clinical psychiatrists, police officers and police scientists (Copson, 1995). A recent international review of the admittance of expert evidence on offender profiling and crime scene analysis in the courts (Bosco, Zappala, & Santtila, 2010) produced similar results: criminologists, psychiatrists, psychologists and FBI special agents had all served as expert witnesses regarding offender profiling. The three main areas on which they had testified were motivation for the crime, whether the behaviour of an offender corresponded with particular personal characteristics, and evidence of linkage between crimes.
PHOTO 10.2 Dr. Thomas Bond provided the UK police with an opinion regarding Jack the Ripper’s physical appearance, likely dress, state of mind and motivation for a series of murders in London in 1880.
Source: © frankie’s/Shutterstock
With such a long history, we might expect offender profiling to have been the subject of a considerable body of scientific research. Unsurprisingly, then, Dowden, Bennell, and Bloomfield (2007) identified 132 articles related to offender profiling that spanned 31 years; however, most articles were “discussion pieces” with only the minority of articles reporting inferential statistics. Similarly, a narrative review by Snook, Eastwood, Gendreau, Goggin, and Cullen (2007) of 130 articles concluded that most writing on the subject used “co
mmon-sense” rationales rather than empirically-based rationales. This lack of empirical research is a cause for concern in light of psychological principles of evidence-based practice. As Dowden et al. (2007) argue, empirical research is needed for offender profiling to be seen as a “legitimate field of study within the behavioral sciences” (p. 52). The few empirical studies of the principles underpinning offender profiling will be discussed below, but first a brief history of offender profiling will be sketched.
10.3.1 Schools of Thought in Offender Profiling
When talking about the history of offender profiling, it is often categorised into three different schools of thought (Alison et al., 2010):
The criminal investigative approach (also called the FBI approach, criminal investigative analysis or crime scene analysis)
The clinical-practitioner approach (also known as diagnostic evaluation)
The statistical approach (sometimes investigative psychology).2
A very brief and simplified outline of each approach is given below. The key difference between these three approaches is the knowledge base used by the offender profiler to make inferences about an offender’s characteristics.
10.3.1.1 The criminal investigative approach
The criminal investigative approach to profiling is informed by the investigator’s extensive knowledge developed from exposure to a large number of cases and their experience of criminal investigations (Canter, 2004). In addition, the FBI conducted interviews with incarcerated rapists and murderers from which typologies of offenders were developed (Wilson, Lincoln, & Kocsis, 1997). For example, in-depth interviews with 36 convicted sexually motivated serial murderers were incorporated with the extensive investigative experience of members of the FBI Behavioral Science Unit to produce a typology of murderers (Jackson & Bekerian, 1997). This typology distinguished between murderers who were classified as organised, disorganised or mixed, based on their crime scene behaviour. An organised murderer would show evidence of having planned the crime, of having committed the crime in a controlled manner, leaving few or no clues and with a stranger as a target (Jackson & Bekerian, 1997). In developing the typology, offender characteristics were associated with each classification and thus predictions could be made about an offender’s characteristics by classifying the murder into one of three types. For example, an organised murder would be committed by someone who lived an orderly life, was of average to high intelligence, socially competent and likely be engaged in skilled employment (Canter, Alison, Alison, & Wentink, 2004).
Forensic Psychology Page 47