Biases
and pitfalls in profiling
There are a number of
misconceptions about profiling, usually based on its fictional use and
psychodynamic portraits of politicians. Rarely does profiling provide the specific
identity of the offender, and this is not its purpose. The aim is to narrow
the field of the investigation and suggest the type of person who committed the
crime (Douglas et aI, 1986).
The profile report will try to establish the gender, approximate age,
marital status, educational level and details of possible occupation of the
offender. There may be suggestions of whether this person has a previous
police record and if another offence is likely.
Evaluation of offender profiling
Whether profiling
is effective or not is a key question, and historically there are famous
successes and failures. One of the best-known failures in America was the case of
Albert DeSalvo (known as the ‘Boston Strangler’). A profile suggested the
offender was a male homosexual schoolteacher living alone. When arrested,
DeSalvo was found to be a heterosexual construction worker living with his
family. In the UK, the Rachel Nickell case is seen as a failure of offender
profiling.
Offender profiling
captures the imagination. But with the Rachel Nickell case, its use has been
brought into question.
‘Rachel Nickell was a young woman who was brutally murdered in
mid-morning while walking on Wimbledon Common in south London. As part of the
investigation into the killing, a profile was commissioned from a psychologist.
A suspect was eventually identified and it was noted that he seemed to fit the
profile well. An elaborate operation, drawing partly but not only on the
profile, was put together in which a police woman befriended the socially
isolated and inadequate suspect, offering the promise of an intimate relationship
in exchange for descriptions of his sexual fantasies and a confession that he
murdered the woman on the Common. The confession was not forthcoming, but he
was still arrested. The case fell apart.’
When Pinizzotto and Finkel (1990) attempted to
discover whether professional profilers would be more accurate than informed
lay-persons, they asked groups of profilers, detectives, clinical psychologists
and students to examine two closed police cases (a sex offence and a homicide)
and to draw up profiles. What they found was that the profilers did indeed
produce richer and more detailed profiles and in relation to the sex offence,
they were more accurate than non-profilers, but the detectives were more
accurate on the homicide case. Pinizzotto and Finkel concluded that the success
of the profilers was the result of both confidence and experience rather than
the use of an exclusive technique. The implications would therefore be that
both training and practical experience are vital in developing profiling
expertise and that productive liaison between the police and psychologists is
the way forward in order to achieve both investigative and clinical objectives.
Pinizzotto and
Finkel (1990) argue that profiling is most effective in serial sexual offences
because of the extensive research base, and least effective for fraud,
burglary, robbery, theft and drug-induced crimes. Holmes (1989) feels that it
is most useful when there is a psychopathology involved, such as sadistic
assault.
Holmes (1989) cites FBI data, which reveal that
in 192 cases of profile generation in 1981, arrests were made in 88, but in only
17% of these did the profile contribute to the arrest. Others (e.g. Oleson,
1996) point out that the seminal work of the FBI in establishing offender
profiling may be methodologically flawed since no control groups were used to
compare the evidence obtained from interviews with offenders and there is no
mention of the statistical techniques used to analyse their data. Moreover,
much of the evidence used by the FBI was simply information obtained in
interviews with offenders and was accepted at face value.
More recent research has made greater claims for the
usefulness of offender profiling (for example, approximately 80% of cases
solved were helped by offender profiling according to Canter and Heritage,
1990).
In the UK, a survey of detectives in 48 police forces, who had worked
with offender profiling concluded that identification of the offender came in
2.7% of cases and general help in 16% (Copson, 1995). What the survey did find was variety in the individuals who did
the profiling. Those involved included clinical psychologists, forensic
psychiatrists, academic psychologists, clinical psychiatrists, forensic
psychologists and consultant therapists. The skill of the individual
profiler determined whether the police officers were satisfied with profiling
generally. ‘Indeed the research suggests that, at this stage of the development
of profiling in Britain, approaches to profiling are idiosyncratic’ (Copson,
1995, p. 25).
Did the advice
given by the profile: |
Yes |
No |
Assist in solving the case? |
14.1% |
78.3% |
Open new lines of enquiry? |
16.3% |
82.1% |
Add anything to information supplied? |
53.8% |
38.6% |
Prove operationally useful? |
82.6% |
17.4% |
How was the
advice operationally useful? |
% |
Led to identification of offender |
2.7 |
Furthered understanding of offender |
60.9 |
Offered structure for interviewing |
5.4 |
Not useful |
17.4 |
Source:
Copson, 1995.
Bartol (1999) admits that ‘profiling is probably at least 90% an art and
speculation and only 10% science’ (p. 239).
The limited amount of information available in Britain to date suggests that many senior detectives have somewhat negative views as to the usefulness of information provided by profilers (Copson, 1995; Jackson et al, 1997; Smith, 1998). Britton (1997) also admits that a large number of cases continue to be solved, not by profiling, but by routine police work, or the use of forensic evidence. Such apparently poor results should however be viewed in the light of comments made earlier in respect of police officers’ possible reluctance to admit that ‘outsiders’ did help in an investigation. Profiling does at least allow the police to better focus their investigations. This can be important for as Canter (1994: 21) has noted the alternative is that the police will simply throw more and more resources at a crime in the hope that ‘something will turn up’. By using a more scientific approach, profiling, thus, has advantages in terms of the resources needed to solve any particular crime.
When trying to evaluate the usefulness of a profile it may also be important to consider whether the profile is seen in isolation or merely as one part of more general guidance which a psychologist might provide to investigators. Jackson et al (1997) conclude that when profiles are considered as a separate entity, they seldom, if ever, offer enough foundation or impetus to steer or guide an investigation in a new direction’ (Jackson et al, 1997: 131).
For this reason, these authors report that in the Dutch profiling unit, profiles are never treated as a separate entity, but rather as a management instrument. As such any profile should be accompanied by practical advice on how best to proceed with a particular investigation. Copson (1995) appears to support this viewpoint. He suggests that some of the negative comments made by senior detectives about profiling may stem from a misunderstanding of what ‘profiling’ can actually achieve. Psychologists can offer different forms of assistance to an investigation, only one of which might be the drawing up of a profile of the most likely offender characteristics.
We should also be aware of the danger of the self-fulfilling prophecy with respect to profiling (Ainsworth, 1995a: chs 3 and 10). Detectives must bear in mind that a psychological profile may well fit a number of people and may not be totally accurate. The fact that the suspect currently being interviewed happens to fit the profile does not ‘prove’ that they actually committed the offence. There may be several people who share the suspect’s make-up and the police should be cautious before making a presumption of guilt. As was noted earlier, the psychologist will tend to work on probabilities. By contrast, the police may be likely to operate in more absolute terms of guilt or innocence. The danger is that once a person has been labelled as a suspect and brought in for questioning, the police will make a presumption of guilt and see their role as merely to elicit a confession (Ainsworth, 1995b).
One ethical dilemma, which profilers will need to address, is the possibility that information, which they provide, might result ultimately in the wrongful conviction of an innocent person. While much of the blame for such a miscarriage of justice might well rest with the police (Gudjonsson, 1992) the profiler must acknowledge that their initial contribution may have been significant.
As we have noted earlier, there is little good scientific research to which one can turn in trying to answer the question of how useful psychological profiling is. Success or failure are not so easily measured when one is dealing with the sort of material used in psychological profiling. If a profiler’s information proves to be 50 per cent accurate and 50 per cent inaccurate should this be counted as a success or a failure? As was noted earlier, we also need to be sure whether the information, which a profiler provides, is of a type that others (e.g. detectives) could not reasonably have deduced for themselves.
Commentary
Canter is critical of media and fictional portrayals of what he does. He emphasizes his scientific basis. Boon and Davies (1992) calls this a ‘bottom-up’ approach. This is working with the details and building up specific associations between the offences and the offender’s characteristics. The USA approach is classed as ‘top-down’ based on interviews with the known criminals. Canter is also critical of the American approach based on interviews with ‘serial killers’, who are known to be manipulative.
EVALUATION OF PROFILING
Campbell (1976) suggested
ominously that ‘there is no clear evidence that psychologists are any better
than bartenders at the remote diagnosis of killers’ (p.119) and 22 years later
Canter and Alison (1999) suggested that in many accounts of profiling
‘systematic observations or procedures are discouraged in favour of instinct
and intuition’ (p.7). Their view is that profiling as it is popularly described
is insufficiently grounded in psychological principles of empirical investigation,
and does not always adhere to the ethical framework required by the British
Psychological Society.
O’Keefe and Alison (2001)
suggest that profiling may suffer from a possible Barnum effect whereby people,
in this case police officers, are presented with ambiguous, generalised
statements and then see what they want to see in the content, disregarding what
does not fit in with their overall perspective. When there is considerable
pressure on police officers to make an arrest, they may seize upon statements
made by profilers as a result of the pressure rather than the convincing nature
of the profile. Within this context profilers are seen by some to operate in
much the same way as astrologers and psychic detectives who have previously assisted the police. O’Keefe and Alison found in
their study that psychic detectives were in fact no more accurate in their
predictions than controls, and relied on a variety of rhetorical devices
associated with the cold reading strategies used by psychics in telling people’s
fortunes.
Pinizzotto and Finkel (1990)
attempted to discover whether professional profilers would be more accurate
than informed laypersons in drawing up profiles from two closed police cases,
and so they compared groups of profilers, detectives, psychologists, and
students. What they found was that the profilers did indeed produce richer and
more detailed profiles, and that they correctly predicted the characteristics
of the convicted offenders. Pinizzotto and Finkel concluded that this was the
result of their confidence and experience rather than an exclusive technique.
When Kocsis et al. (2000) repeated
this paradigm, however, including a group of self-declared psychics, the
professional profilers demonstrated a superior set of skills for the task,
while the psychics seemed to rely on little more than social stereotypes.
Significantly the performance of the group of psychologists was better than
that of the other non-profiler groups, suggesting that some psychological
knowledge is of benefit to the process of profiling.
Ainsworth Peter B., 2000, Psychology and Crime, Longman, ch 6
Brewer Kevin, 2000, Psychology and Crime, Heinemann, pp 37-42
Dwyer Diana, 2001, Angles on Criminal Psychology, Nelson Thornes, pp 47-52
Harrower Julie, 1998, Applying Psychology to Crime, Hodder & Stoughton, pp56-62.
Harrower Julie, 2001, Psychology in Practice, Crime, Hodder & Stoughton, pp 51-64
Psychologists used to 'solve' crimes by measuring your
ears. Now they simply size up your neuroses.
Nick Cohen
Sunday July 8, 2001
The Observer
Criminal
Profiling:
Fact, Fiction, Fantasy and Fallacy
by Wayne Petherick