The question of whether or not offender profiling offers false hope points to the challenge that in theory or as an idea the possibility of inferring characteristics of an offender from a crime scene is ground breaking, but is it applicable in practice. The notion of being able to predict behaviour, apprehend a suspect and aid in a possible conviction or proof of innocence from a profile is attractive and one that has received much attention from the media and general public. More pertinently, for whom does it offer false hope? The investigation and police force? The lawyers, judges and jury? The victim? Or the offender? Or even false hope for forensic psychology as a discipline? Furthermore, false hope for what? Investigations? Offender apprehension? Probative guilt or innocence? A critical assessment of the scope, assumptions, professional issues, ethical issues and legal issues of offender profiling might be able to offer some answers to these questions. In the first instance, definitions of offender profiling and a historical overview will be provided before addressing the specific issues surrounding this relatively new and unknown discipline.Historical Overview
The concept of profile analysis can be traced back to the late Nineteenth Century in the case of Jack-The-Ripper, pseudonym given to the individual responsible for five murders in the Whitechapel area of London in 1888. Police surgeon and medical doctor, Thomas Bond offered a description of the characteristics likely to be associated with the offender responsible. His description included details on appearance, age, clothing and personality traits which were derived from an assessment of the victims and crime scenes. To date however, the case remains unsolved (Howitt, 2006). In 1943, Dr Walter C. Langer developed a profile of Adolf Hitler based on Hitler's speeches, his book Mein Kampf, interviews with people who knew Hitler and published reports. In his book The Mind Of Adolf Hitler (1972), Dr Langer referred to possible behavioural and psychological traits, reactions to losing World War II and he theorised Hitler would be likely to die from suicide or euthanasia, most of which proved to be correct.
The first successful application of a profile analysis dates back to 1956 in the case of New York City's "Mad Bomber", George Metsky, by psychiatrist Dr James A. Brussel . From an examination of the crime scene and the bomber's letters to the police, Brussel predicted the Bomber would be of large build, middle aged, living with siblings and would be wearing a double breasted suit neatly buttoned up when he would be found (Jackson and Bekerian, 1997). Although George Metsky was not apprehended as a result of this profile, he fitted the description perfectly.
More recently the field of profiling has been dominated by two different approaches. Holmes and Holmes (2002) identify an inductive and deductive approach the profiling. The inductive approach is based on the assumption that if different crimes committed by different offenders share similar features, then the offenders must also share common personality traits. It is basically as assessment of similar crime scenes. In contrast the deductive approach focuses on an in-depth analysis of the crime scene and evidence left behind, from which a picture of the unknown offender can be generated in terms of psychological and behavioural characteristics. Deductive profiling can be further categorised into two different techniques.
In the 1970's the concept of offender profiling began to be looked at for investigative purposes more methodically. The FBI set up their Behaviour Analysis Unit (now called The Investigative Support Unit) in an attempt to use the behavioural sciences in the investigation of serious contact crimes. The basis for this Unit was that forensic evidence was only really useful when detectives actually had a suspect or possible offender in custody and rarely offered any insight into the sort of person the investigation should be looking at. The role of the Behavioural Analysis Unit was to provide the police with a likely psychological profile of an unknown offender which could be used to direct or narrow down the investigation. The FBI approach to profiling centres around classifying the offender and predicting characteristics deducted from crime scene analysis, the nature of the attacks, forensic evidence and victimology (Jackson and Bekerian, 1997). The FBI constructed two basic models: murderers could be classified into Organised (pre-meditated attack, control displayed at the crime scene, victim is a stranger and little evidence left at the scene) and Disorganised (chaotic crime scenes displaying erratic behaviour) or a mixture of the two. They also developed four rape typologies: power assurance, power assertive, anger retaliatory and anger excitation (Jackson and Bekerain, 1997). Eventually the Crime Classification Manual was developed in 1992 and lists the type of crimes in which the behaviour of the perpetrator plays a significant role.
The process adopted by the FBI is organised into four systematic stages: data assimilation (collection of information from as many sources that are available), crime classification (organised vs. disorganised etc), crime reconstruction (generates victim behaviour, sequence of crime and the modus operandi) and profile generation. The final profile follows a standard format and constitutes of: demographic and physical information, education and intellect, existing criminal activity or records, military background, family characteristics, habits and social interests, evidence from crime scene, age and type of vehicle, personality characteristics and suggested interview techniques (Jackson and Bekerian, 1997).
The FBI style of 'intuitive' profiling has been challenged by the more statistical/actuarial approach largely associated with the work of David Canter in Britain (Howitt, 2006). In Britain, profile analysis finds its roots in the 1980s as a result of the successful application of its techniques to a number of high profile cases such as John Duffy, "The Railway Rapist" (Canter, 1995). While the statistical approach also relies on the assumption that features of a crime scene contain evidence of behavioural traits of the unknown offender, the statistical approach established these inferences via statistical techniques. There are three main issued which are addressed in the statistical approach. The first pertains to the classification of the crime scene which refers to the information collected at the scene and whether these features merit being classified as a particular type, grouping or cluster. The second issue is whether or not these features reveal any psychological details about the likely offender and thirdly what degree of stability is there in the crime scenes of offenders, re-occurring patterns and how stable and consistent these may be over time (Howitt, 2006). The classification of crime scenes is empirically based using statistical methods such as smallest space analysis. This method aims to identify how likely features of a crime scene are to exist or co-exist with another feature at a particular crime scene (Howitt, 2006). According the Howitt (2006), most statistical profiling remains research based and has yet to be applied to police investigative work, but the potential is there for its' use.
While there are important merits to the inductive approach as defined by Holmes and Holmes (2002), for the purpose of this discussion, the focus will be placed on the deductive approach to profiling, associated with both the FBI's clinical/intuitive methods, and David Canter's statistical approach.Definitions
Jackson and Bekerian (1997) argue that no offender nor any crime are completely unique. Regardless of the crime, the number of motives and the modus operandi tend to be fairly restricted and there are always patterns to be recognised and comparisons to be made against other cases. In stranger violent contact crimes the police are often subjected to information overloads and a distinct lack of referential material. This is the basis for profile analysis, whereby the profile enables an investigation to narrow down its search. There are however various differing definitions of offender profiling. The connecting factor in all these is that from the analysis of a crime, the scene, the victim and other information, descriptive, behavioural and psychological characteristics of a likely offender can be inferred. Jackson and Bekerian (1997) define profile analysis as making inferences about a likely offender from behaviour exhibited at the crime scene or series of crime scenes. They argue profiling uses principles from a wide array of disciplines such as clinical psychology, forensic psychiatry, environmental psychology, social psychology and cognitive psychology. Homant and Kennedy (1998) go on to differentiate between three types of profile analysis: (1) Crime Scene Profiling: analysing evidence from the crime scene to build a picture of the offender. (2) Psychological profiling: the use of personality tests and interviewing to assess how the suspect fits a known personality template of a certain type of offender. (3) Offender Profiling: collecting empirical data in order to generate the characteristics of those involved in a certain type of offense.Scope of application
Profile analysis had proved to be an important and potentially useful tool in the investigative process and apprehending a likely offender. Yet the first major problem presented by offender profiling is the scope of its' application. It has been widely acknowledged that not all crimes are suitable for offender profiling. Jackson and Bekerian (1997) suggest arson and sexually motivated crimes showing evidence of sadistic torture, ritualistic behaviour, evisceration, posturing of the body, staging or acting out a fantasy are most suited to profile analysis. In contrast crimes involving vandalism, robbery and murder committed in the commission of a robbery and drug induced crimes are not suitable for profiling as the personality of the offender is rarely reflected at these crime scenes. The Home Office British Crime Survey for 2007/2008 shows the percentage of types of crimes committed in England and Wales for that period as recorded by both the British Crime Survey and Police Records. It shows the majority of crimes are property related. Vandalism accounts for 27% of crime as per the BCS and 21% from police records. Vehicle related theft accounts for 15% of all BCS crimes and offenses against vehicles for 13% of recorded crime. Burglary accounts for 7% of all BCS crime and 12% of recorded crime. Violence against the person and sexual offenses, account for 19% of the BCS crimes and 1% of crime recorded by the police. Most serious violence against the person where injury inflicted or intended is life threatening only accounts for 2% of all violence against the person (www.homeoffice.gov.uk). The percentage of crimes suitable for profile analysis therefore represent a small fragment of all crimes and more so, in a large number of these cases the crime is solved relatively quickly as the assailant is more often than not known to the victim (Jackson and Bekerian, 1997). In essence stranger violent contact crimes, which are those best suited to profiling techniques represent such a small proportion of the total number of crimes, it would seem, that it does question how much importance should be placed on profiling techniques. The media and general public have brought offender profiling to the forefront of society and naturally profile analysis offers hope as a tool for explaining and rationalising heinous crimes, and in aiding investigations around unsolved crimes, undetected crimes and/or unreported crimes but this is not a sound basis for advocating its usefulness and efficiency. The question is, is profiling a sound scientific tool or is it there to excuse and rationalise the darkest aspects of humanity? After all, the first question addressed by the public when faced with a heinous crime is What sort of person does this?Assumptions
Profile analysis lacks a consistent foundation as it relies heavily on several assumptions. Indeed Holmes and Holmes (2002) outline four main assumptions. Firstly the crime scene reflects the personality of the offender. The assessment of the type of personality involved in a stranger violent contact crime can be derived from the physical and non physical evidence at the scene and victimology. Secondly it is assumed the modus operandi will remain the same from crime to crime. Thirdly the offender's signature will remain unchanged - this refers to the unique manner in which an offender will commit a crime. Examples range from the method of killing to the acquisition of trophies. Finally offender profiling assumes the offender will not change their personality. The assumption is that the offender will not change aspects of their personality radically over a short period and will display an inability to change. These assumptions have largely been challenged at varying different levels. Jackson and Berkerian (1997) refer to theories of memory used to explain an offender's behaviour. Such theories suggest individuals form schemas of re-occurring events in their lives and it has been suggested offenders have schemas for committing particular types of crime. The extent to which an offender's schema for crime is stable dictates the profiler's ability to identify discriminating characteristics. However, as the offender gains knowledge and experience in crime, their behaviour is therefore likely to change and the schema in accordance. While very little agreement about how offender's learn, any instability in behaviour is likely to hinder the use of profiling techniques. Jackson and Bekerian (1997) suggest profiling might be at its optimum when it is believed the offender has reached a "learning plateau" (p 219) and is therefore stable in their behaviour. The challenges already presented upon the basic assumptions of offender profiling cast a doubt as to the basis for profiling in general. If the basic assumptions are not met, then it becomes difficult to give profile analysis any reasonable consideration.
Reliability, validity and accurateness of offender profiling - a review of studiesCritical issued in profiling
As it stands currently the field of offender profiling suffers from a lack of unification (Jackson and Bekerian, 1997). The main issue here is a lack of agreement regarding methodological frameworks used for analysing behaviour. This is evidently displayed in the opposing techniques of the FBI style of profiling and the statistical/actuarial approach. The statistical approach supports an empirical methodology and attempts to give the area of profiling a solid scientific basis. An example of this would be Canter's study of criminal maps in the area of geographical profiling (Canter, 1995). The alternative approach relies on concepts from clinical and forensic psychology. Deductions between personality and behaviour are taken from multiple observations of single cases, as opposed to population statistics that generalise across multiple cases. It would seem that neither methodology is better than the other; both have their own advantages and disadvantages. Although the FBI style of profiling focuses on important individual characteristics, it runs the risk of missing constant relationships between variables and therefore might have less predictive power when applied to the general population. On the same note, averaging across individual cases as in the statistical approach, means important individual characteristics may me neglected but the empirical methodology does ensure more reliable and valid results. Furthermore the statistical approach helps avoid the possible biases associates with relying on experience to compare similar cases (Canter and Alison, 1999). Particularly so is the bias in recall of information whereby an individual makes judgements on a probability by thinking of examples that are easily accessible to memory while less relevant counter examples tend to be forgotten. Jackson and Bekerian (1997) argue the field would benefit from "hybrid profiling techniques"(Ch 12; p 210). Indeed when different approaches are combined the quality and quantity of information offered is likely to be much better.
Another professional issue in profile analysis is the differences between two individual profilers (Jackson and Bekerian, 1997). It is entirely possible that two profilers working on the same case could generate entirely different profiles. There are several reasons as to why this could occur. Profilers will differ in their analyses depending on which methodological framework they are using and the degree of experience they possess. Jackson and Bekerian (1997) referred to differences in formal and tacit knowledge between two profilers and these will influence the extent to which they draw the same conclusions about a likely offender. Alison and Canter (1999) further noted that inferences made on the basis of such tacit knowledge are subject to various distortions, biases and shortcomings associated with decision making. Particularly they refer to what is known as the hindsight bias, whereby "an individual underestimates how much they have learned, claiming they knew all along the outcome of the situation" (p 29). Once an offender is apprehended it becomes very easy to claim the success of a profile, referring only to the correct elements, and forgetting those incorrect predictions.
A further issue surrounds the lack of training and development of credentials in offender profiling (Canter and Alison, 1999). While there are rigorous professional requirements and qualifications for areas such as clinical psychology or occupational psychology, there are little agreed professional criteria in existence in order to consult on a case as an offender profiler. While there are training courses available surrounding psychological contributions to police investigations, there is not yet a course designed to produce a 'profiler', like a school of medicine would produce a medical doctor. Moreover, it is to be noted that the title 'Profiler' does not actually exist within the Investigative Support Unit in the FBI. The above is likely to lead profilers to be treated with a certain degree of suspicion and even may be treated no differently to a psychic, for example, brought in to consult on a case. It is worth noting however that guidelines are currently being drawn up by the British Psychological Society for expert evidence but it is still a long way from establishing adequate assessment procedures (Jackson and Bekerian, 1997).
The methodological issues and lack of professional criteria also lend to the problem of cross-cultural transferability of profiling techniques (Jackson and Bekerian, 1997). The notion that offender characteristics could be culture specific casts a doubt of the use of cross cultural training and profiling techniques in general. The authors suggest there should be more research into cross-cultural differences, that profiling techniques should be tailored to individual police forces and sensitive to the cultural aspects of crime, offender behaviour and the investigation.
Profiling techniques have also received a large amount of criticism surrounding the lack of a sound theoretical basis to accompany any empirical evidence. Jackson and Bekerian (1997). In research it is important to establish that there is a theoretical explanation for any relationships established between variables. The authors argue that this one of the larger problems that confront research in the area of offender profiling. However Holmes and Holmes (2002), have argued that there is an existing theoretical framework for profile analysis and is drawn from individual theories of crime including psychology and crime, personality characteristics, psychiatry, psychoanalysis and Freudian theories, constitutional theories, heredity and atavists, physical characteristics and crime, chemical imbalances and hormonal influences. The authors also look to social and ecological theories of crime as a basis for profiling techniques focusing on crime as learned behaviour, cultural transmission theories and social bond theories. These academic theories aid the profiler to reach an understanding of the personality of the offender and make for another tool available in the investigative process.
The quality of a profile analysis is dependent on the quality of the information gathered in order to generate it. This information ranges from sound forensic evidence, autopsy reports, offender and victim details. Witness accounts can also prove to be useful but these in turn can be plagued with problems in terms of accuracy of memory which can affected by elements such as emotional arousal, interviewing techniques and witness's reasons for remembering certain information ( Jackson and Bekerian, 1997).
Alison and Canter (1999) outline other issued in offender profiling. They refer to anchoring and adjustment as the process by which individuals will base a decision on what they have already been told, even if they have been told the information is irrelevant to making an accurate assessment. For example mentioning a suspect to a profiler could bias their approach on the case, even if the suspect is cleared at a later stage. There is also the problem of hypothesis testing, whereby individuals look for information to support their theory and ignore important counter examples. Illusionary correlation, where an individual finds as association between two variables where there is none based on prior beliefs can lead to distorting the information available. The frailties of group decision making are also among the weaknesses of offender profiling and these include group over-estimations, closed mindedness, stereotypes of out groups, pressures towards uniformity, illusion of unanimity and direct pressure on dissenters. These issues are applicable to both the psychologist and police force as well as expert witnesses in the courtroom and the jury.Ethical issues
Alison and Canter (1999) outline a number of ethical issues that need to be carefully taken into consideration in a profile analysis. These include keeping detailed documents, quoting sources of inferences made, accepting challenges made by peer review. Biases can be avoided by being aware of all the facts to hand, presenting the facts to a neutral third party, ignoring any existing suspicions regarding suspects, keeping records of the ongoing investigation and respecting all parties involved in the investigation.
Although not stated in the literature as such, an ethical or moral objection with offender profiling lies in the notion of arresting and convicting a perpetrator on the basis of a profile. If such importance were to be given to offender profiling, then it begs the question of being able to apprehend a suspect before the commission on a crime. Because an individual matches the profile of an offender, there is the question of whether this person could be monitored for criminal activity and subsequently arrested on the basis that they are likely to commit a crime. Of course there are numerous arguments to be made about these suggestions, but the point here is assuming transparency of an individual's personality and behaviour, as does a profile analysis, raises questions of fundamental human rights. If the field develops into a more systematic area of research and practical application, and gains professional accreditation, then how far would authorities go with the concept of being able to predict human behaviour?
Legal IssuesAlison and Canter (1999)
Alison and Canter (1999) approach the critical legal issues surrounding profiling. They argue the main problems surround relevance and admissibility in court. The attractiveness of a profile analysis is that from certain clues it is possible to generate a complete picture of the offender. However as Ormerod (1999) argues while parts of a profile may be based on scientific data, the gaps are more often than not completed with intuition, largely based on the experience of the expert. The courts will not be prepared to admit evidence based on intuition or a "gut feeling". Any evidence presented in a court must be relevant to the facts in issue from a legal perspective. This refers to the tendency of a given item to have probative value to make one of the elements of the case likelier or not. Generally evidence which is not probative is inadmissible in court. Furthermore evidence likely to lead the trier of fact to be distracted will also be rejected. A profile but definition is an educated guess or an opinion of a psychologist of the likely characteristics of an offender. The profile must be demonstrated as a fact and must be true. However as the profile in its' entirety is an opinion it cannot be submitted as a fact or true. In this sense a whole profile will not be admissible in court, even if it is relevant to the case. Ormerod (1999) suggests that it is more likely to have parts of a profile admitted in court. For instance, inferences made in the profile based on systematic reliable research could be used. The author uses the example whereby research demonstrates 90% of rapes are intra-racial. Admissability is further challenged with the concept of exclusionary discretion in the courts. The court is entitled to exclude any evidence at its' general discretion, usually based on information gathered in an unfair fashion.
Even if the evidence proves to be relevant, it will remain inadmissible if it is deemed unreliable, unfair, prejudicial or privileged (Ormerod, 1999). The risk of jury prejudice is increased in cases concerning deviant crimes. As previously stated, profile analysis is best suited to this type of crime and is likely to be used in this context. Furthermore, it is argued prejudice is also increased in cases of serial crime, again a typical offense where profile analysis is called upon. Previous misconduct can equally distract and mislead the jury, in addition to lengthening the process. It would seem here that the problem presented by a profile in a difficult case is that it could positively reinforce the pre-conceptions of the jury regarding deviant and/or serial crime. However is the probative force of the evidence sufficiently outweighs the prejudicial effect on the jury, then it can remain admissible in court.
There is then the issue surrounding similar face evidence, cumulative or sequential, used to prove the identity of the offender (Ormerod, 1999). A profile which suggests that X% of offenders who commit a particular crime, possess a certain characteristics, and that the accused possess this characteristic does not have sufficient probative force. It does not prove that the accused is the perpetrator of the crime. A consequential problem here is that the profile can create stereotypes, marginalising and criminalising certain segments of the population. For example the fact that John Duffy, the railway rapist, possessed pornography and martial arts paraphernalia, was not probative in determining guilt.Expert witnesses
Regardless of legal issues surrounding the profile, the question of a psychologist as an expert witness is equally problematic. The first stumbling block relates to expertise. The psychologist must be deemed suitably qualified in order to be received as an expert. As discussed earlier on, this is automatically a problem as there is no systematic training or agreed professional criteria in order to compile a profile analysis. Secondly, in order to bring an expert into the court, the information to be used must be considered beyond the realms of the understanding and knowledge of the jury. A profile analysis uses concepts often outside the jury's competence, particularly in relation to mental health and deviant crimes, and as such could be deemed useful in court, providing it is considered relevant and admissible in the first place.
The profile offered by an expert witness must also be considered reliable and the sources of information must be deemed valid. An objective assessment of reliability is very difficult, but the more scientific the data is the better it will be received in court. This ties in with the notion that offender profiling needs to be standardised in order to be systematically accepted in the legal system.
Beyond this the expert themselves are liable to be subject to numerous biases. As they would have more than likely assisted in the police investigation, they are likely to be biased towards getting a conviction. Furthermore, there is also the possibility that the jury could place faith in the evidence presented by the expert. While the expert is prohibited from giving their opinion on the matter the tribunal will ultimately have to make a decision on, the jury may rely too much on the expert evidence and their possible biases as they simply accept the information as true because it is labelled as expert. Inherent in this is the notion that expert statements relating to the offender in a court are transposed onto the accused. A profile or part profile is an analysis of the offender and in no way means it is about the accused. Alison and Canter (1999) argue that because opinions from clinicians are based upon experience of similar cases and statistical research is based on assumed data sets neither can be directly linked to the case actually being looked at. Yet it is easy to see how a jury might make this inference.
- Howitt, D. (2006) Introduction to Forensic and Criminal Psychology. 2nd Edition. Pearson
- Jackson, J.L. and Bekerian, D.A. (1997) Offender Profiling: Theory, Research and Practice. Chichester: Wiley.
- Canter, D.V. (1995) Criminal Shadows: Inside The Mind Of A Serial Killer. HarperCollins.
- Holmes, R.M and Holmes, S.T. (2002) Profiling Violent Crimes: An Investigative Tool. 3rd Edition. SAGE Publications.
- Homant, R.J. and Kennedy, D.B. (1998). Psychological Aspects of Criminal Profiling. Criminal Justice and Behaviour, 25 (3), 319-343.
- Canter, D. and Alison, L. (1999) Profiling in Policy and Practice. Asgate