Файл: [D._Kim_Rossmo]_Geographic_Profiling(BookFi.org).pdf

© 2000 by CRC Press LLC

During the hearings, FBI personnel pointed out in rebuttal that the 

U.S.S.

Iowa

 equivocal death analysis was focused on answering the single question

as to whether the explosion was suicide, homicide, or suicide/homicide;
accidental death was not an option. And while psychologists on the APA
panel criticized the FBI’s methodology, not all disagreed with their conclu-
sions (Michaud & Hazelwood, 1998).

Homant and Kennedy (1998) help explain the police perspective on

profiling validity. “From a law enforcement point of view, there is no need
to wait for assurances that profiling in general is a valid process, as long as
there are not any more promising alternatives and as long as the process is
used cautiously” (p. 323). While suspects and lines of inquiry should not be
excluded just because they fail to fit the profile, investigators sometimes must
make hard decisions quickly; under such conditions, profiling information
— even of the “informed guess” type — may be of value.

Still, the APA review contains important warnings regarding the use,

limitations, and dangers of psychological profiling. “Profiles should be viewed
as supportive and not substantive” (Jackson et al., 1993a, p. 31). The FBI
itself cautions that profiling is still more of an art than a science, and should
never be considered a replacement for traditional investigative methodologies
(Ault & Reese, 1980; Hazelwood & Douglas, 1980; Porter, 1983, 52; Ressler
& Shachtman, 1992). Discussion has taken place in regards to developing
automated crime profiling through the use of artificial intelligence and com-
puter-assisted expert systems, but to date this goal has not been realized
(Icove, 1986; Reboussin & Cameron, 1989; see also Benfer, Brent, & Furbee,
1991; Bunge, 1991). Dietz (1985) suggests that “it may be more accurate to
regard profiling as a process of logical reasoning that draws on experience,
insight, and judgment at each step of the process. In this sense profiling may
resemble the process of clinical reasoning in medicine” (pp. 217–218).

5.5 Evaluation Studies

These concerns, critiques, and suggestions have led to a variety of evaluation
studies designed to examine the investigative merits of criminal profiling.
Evaluative research has occurred in the U.S., Britain, The Netherlands, Can-
ada, and Australia. These studies often approach the assessment task by
addressing three key issues: (1) how accurate is profiling; (2) how reliable is
the process; and (3) how useful are the results? For a profile to be useful, it
must assist in the investigative decision making process. Suggestions that are
vague, general, unworkable, or of low probability are not likely to produce
helpful leads. These questions and requirements provide an important frame-

© 2000 by CRC Press LLC

work for the development of any new investigative methodology, psycholog-
ical and geographic profiling included.

One of the first criminal profiling evaluation projects was an FBI in-

house survey of 192 users of BSU-prepared profiles, comprising 209 cases
(65% involving homicide, 35% rape, and 27% other offences)

26

 (Institutional

Research and Development Unit, 1981). The study determined that only 46%
of these crimes had been subsequently solved. Of these 88 investigations, the
profile was found to have helped: (1) focus the investigation (72% of the
cases); (2) locate possible suspects (20%); (3) directly identify the suspect
(17%); (4) assist in the prosecution of the suspect (6%); or (5) was of no
assistance (17%).

27

 Of the 104 unsolved investigations, the profile was still

seen as helpful in terms of generating leads, suggesting motives, and con-
firming other findings. The BSU typically receives only the most difficult of
cases, those that have defied traditional investigative efforts, and these find-
ings need to be placed within that context (Fox & Levin, 1992; see also Jackson
et al., 1994).

Pinizzotto and Finkel (1990; see also Jackson et al., 1994; Pinizzotto,

1984) examined criminal profiling outcomes and processes through the
expert/novice approach, a technique used in cognitive psychology to explore
expertise in specific domains. Between-group differences were analyzed in a
study that included profilers, police detectives, psychologists, and students.
The researchers found that the group of profilers wrote more detailed and
valid reports than did the other groups, but they observed mixed results
concerning ability to correctly predict offender characteristics. No between-
group qualitative differences in information processing were noted.

Kocsis, Irwin, Hayes, and Nunn (forthcoming) conducted a similar study

to test the influence of investigative experience, knowledge of criminal psy-
chology, objective and logical analysis, and intuition on the psychological
profiling process. Predictions for a murder case were compared between
groups of profilers, police officers, psychologists, students, and psychics.
Profilers were slightly more accurate than nonprofilers in the identification
of offender cognitive processes, physical characteristics, social history and
habits, and offence behaviours. Psychologists did better in some areas such
as determining the offender’s personality characteristics. Psychics appeared
to rely upon nothing more than social stereotypes of murders.

The Netherlands Institute for the Study of Criminality and Law Enforce-

ment (NISCALE) conducted several evaluation studies of specific profile
analysis and investigative advice as provided by the Scientific Advisory Unit
of the Dutch National Criminal Intelligence Division (CRI) (Jackson et al.,
1994; Jackson, van Koppen, & Herbrink, 1993a, 1993b). They first explored

26 

Percentages do not add to 100 as a single case may involve more than one type of offence.

27 

Percentages do not add to 100 as multiple responses were allowed.

© 2000 by CRC Press LLC

differences between the operational mental schemata of experienced police
investigators and offender profilers (Jackson et al., 1994). Models for struc-
turing knowledge consist of domain, inference, task, and strategic layers
(Adhami & Browne, 1996), and expert knowledge is derived from a series of
“if-then” rules (some areas of expertise contain heuristics numbering in the
tens of thousands). To be useful, an expert’s knowledge must be applicable,
have the potential to be generalized to other situations, contain structure and
organization, and possess depth.

The cognitive representations and strategies employed by detectives were

found to be qualitatively different from those used by profilers. Specifically,
the if-then production rules of the police detectives were either too global
(general) or too specific (idiosyncratic) to be of much investigative value.
Another component of the study found process differences between an expe-
rienced sexual offence detective and an FBI-trained Dutch profiler. The
former was observed to work in a bottom-up fashion, concerned with the
“what,” while the latter had a top-down approach, and focused on the “who.”

The research revealed interesting differences in how profilers and detec-

tives assimilated data (Jackson, 1994). On average, profilers devoted substan-
tially more time per case on police tip reports than did detectives (12 hours,
29 minutes vs. 2 hours, 33 minutes), and about the same amount of time on
the autopsy reports (14 minutes vs. 16 minutes). Profilers spent less time per
case on the main police reports (4 hours, 54 minutes vs. 6 hours, 28 minutes),
the geographical locations (18 minutes vs. 31 minutes), and the crime scene
photographs (26 minutes vs. 1 hour, 9 minutes). Overall, profilers spent
considerably more time than did detectives on data assimilation (18 hours,
21 minutes vs. 10 hours, 57 minutes). The observation that, on average,
detectives spent more of their time on the geographic locations than did
profilers (4.72% of total time vs. 1.6%, a factor of 1.7 in terms of absolute
time), is interesting in light of the findings of Pinizzotto and Finkel (1990):
“It was the detective group that scored higher in the homicide questions
dealing with the offender’s employment and the offender’s residence in rela-
tion to the crime scene” (p. 224).

The NISCALE study also looked at “consumer satisfaction” and found

the great majority of detective-customers were satisfied with the criminal
investigative analysis service (Jackson et al., 1994). While the profiles them-
selves received mixed reviews, it was observed that several other, perhaps
more important, functions were also provided (Jackson et al., 1993a). Inves-
tigative suggestions, personality assessments, profile writing, crime assess-
ments, interviewing techniques, and threat assessments were all part of the
profiler’s repertoire. Detectives noted additional benefits resulting from the
profilers’ expertise with bizarre crimes and the fresh perspective they brought
to an investigation. “By taking an independent stance, not bogged down with

© 2000 by CRC Press LLC

the inconsequential details that a detective actively working on a case has to
contend with, the professional profiler can offer directions and advice that
can result in the team achieving success in apprehending the culprit” (Jackson
et al., 1993a, p. 32).

The NISCALE evaluation discovered that “no criminal was actually

caught as a direct result of the profiles made” (Jackson et al., 1993a, p. 24),
but this appears to have resulted from problems with customers as well as
profilers. While the reports of the profilers were sometimes criticized as being
too vague and general, inflexibility on the part of detectives stemming from
lack of acceptance, differences in opinion, financial restraints, time delays,
and organizational considerations, all contributed to instances where the
profiles were not acted upon (Jackson et al., 1993a). The CRI now first
confirms that sufficient investigative resources exist to follow up any profiling
suggestions that might result (Jackson et al., 1994).

The NISCALE study concluded that profiling is not an end in itself, but

only a management instrument for helping direct certain types of criminal
investigations (Jackson et al., 1994). The importance of ongoing criminolog-
ical and evaluative research was also stressed. For example, while an inde-
pendent study validated some of the derivation principles used in profiling
rape cases, other rules were not substantiated within the Dutch context.
Collaboration within the scientific forum is therefore regarded as an integral
component of the CRI’s ongoing policies.

The Offender Profiling Research Programme is a comprehensive evalu-

ation of profiling use within the U.K. being conducted by the PRC and the
National Crime Faculty (Copson, 1993; Davies & Dale, 1995b). Interim study
results suggest that the perceived usefulness of profiling advice is more asso-
ciated with a strategic understanding of the crime and its investigation rather
than with the profile’s inferential aspects (Oldfield, 1995). Such a finding is
consistent with the conclusions of the Dutch research (Jackson et al., 1994).

As part of this research program, Copson (1995) surveyed detectives who

had used profiling services in Britain (n = 184); the majority of cases involved
either murder or sex crimes. Profiling was defined as “any predictions, rec-
ommendations and observations based on the inference of offender charac-
teristics from behaviour exhibited in a crime or a series of crimes, and offered
to investigators as the product of statistical or clinical expertise” (p. v).
Detectives were not always clear on what they expected from a profile, but
the following information was most commonly mentioned: (1) offender
characteristics; (2) understanding of the offender and the future threat level;
(3) interview strategies; and (4) linkages between crimes. When asked if they
would seek profiling advice again, 68.5% of the respondents stated they
definitely would, and 23.9% that they probably would.

© 2000 by CRC Press LLC

A previous study by Britton in the U.K. suggested that profiles were not

particularly helpful in directly leading to the arrest of a suspect, but the
process appears viable and shows promise and potential. A review by Gold-
blatt of profiling advice in solved cases determined that 72% of predicted
offender characteristics (n = 114) were correct, 19% were incorrect, and 9%
lacked sufficient information to be classified. Copson found that only 14.1%
of the profiles helped solve a case and a slim 2.7% actually identified the
offender. But profiling advice did open up new lines of inquiry in 16.3% of
the cases; 82.6% of the respondents stated the profile provided operationally
useful information, and 53.8% that it added value to their investigation.

In addition to the accuracy of a profile, detectives were further concerned

over the timeliness and clarity of the response. Verbal profiles, in particular,
were not helpful; written reports are less likely to be misunderstood or
forgotten. But it is also incumbent upon the investigators to ask questions
and seek explanations regarding things they do not understand. Unfortu-
nately, profiling advice was not always acted upon, limiting its potential
utility. It would be interesting to compare the results of this study with
research on such established investigative techniques as crime scene finger-
printing and eyewitness interviews.

Copson suggests that profiles are most useful indirectly, assisting detec-

tives in understanding the crime and developing an “investigative philoso-
phy.” He also notes they help ensure that a complete and proper investigation
has been done. The Association of Chief Police Officers (ACPO) Crime
Committee on Behavioural Science and Investigative Support has now devel-
oped policy regarding research, operational use, and media disclosure for the
British police service in an effort to improve the quality of profiling advice
in criminal investigations.

5.6 Profiling and Probability

To many police investigators, profiling is an arcane art, its predictions origi-
nating from a black box. For offender profiling to realize its potential, two
things must happen. First, profilers must better understand the requirements
and needs of police investigations; and second, investigators must better
understand the nature and use of profiles. Profiles do not solve crimes as that
can only be done by way of a confession, witness, or physical evidence (Klock-
ars & Mastrofski, 1991). Instead, they are best used to guide strategy devel-
opment, support information management, and improve case understanding.

There is a symbiosis between the investigation and the profile; the more

complete and thorough the former, the more accurate the latter (Teten, 1989).
For example, if a careful linkage analysis has not been done, then the inves-