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

© 2000 by CRC Press LLC

investigation since the development of the Henry System for fingerprinting
(Bigbee, Tanton, & Ferrara, 1989; Burke & Rowe, 1989; Eisner, 1989; Gaud-
ette, 1990; Kelly, Rankin, & Wink, 1987; Lowrie & Wells, 1991; Wambaugh,
1989). As both semen and blood contain DNA, the potential for ascertaining
and verifying links in cases of violent and sexual crimes is significant. One
of the necessary steps to realize the potential of DNA analysis is the estab-
lishment of centralized indices to facilitate computerized searches and com-
parisons (Adams, 1989; Miller, 1991).

In 1990, the FBI began development of a national DNA identification

index, completing the combined federal, state, and local police agency pilot
project in 1993 (Brown, 1994; see also Weedn & Hicks, 1998). Designed for
the compatible storage and comparison of DNA records, the Combined DNA
Index System (CODIS) consists of two investigative indices — the forensic
index for unsolved crimes, and the convicted offender index for known felons.
A missing persons and unidentified bodies index will also be part of the system.

One of the important functions of such a system is the establishment of

series crimes (Brown, 1994). For example, DNA pattern matching linked 18
unknown suspect serial cases together in Minnesota, eventually helping to solve
them by matching specimens of two offenders to the crime scenes. Such results
are more common in Britain where a DNA database has been in existence for
some time (Davies & Dale, 1995b; “DNA profiling,” 1995). As of May 1997, it
has been responsible for over 5000 hits — defined as a match between crime
scenes or between crime scene and offender — in the U.K. Canada has also
enacted legislation for the establishment of a national DNA data bank.

Automated fingerprint identification systems (AFIS) — the method that

identified Richard Ramirez as the Night Stalker — are becoming increasingly
common in police agencies, allowing comparisons and matches that sheer
volume would have previously precluded (Sparrow, 1994). Programs such as
the 

Integrated Ballistics Identification System

 (

IBIS

), networked into the joint

ATF/FBI National Integrated Ballistic Information Network (NIBIN), link
crimes committed with firearms through computerized comparisons of
microscopic bullet striation patterns and shell casing marks (Dees, 1994;
Strandberg, 1994; see also Davis, 1958; Di Maio, 1985). And the British
Shoeprint Image Capture and Retrieval System (SICAR) is a national data-
base that connects crime scenes through the geometrical shapes associated
with stored images of footprint evidence.

4.2.1.2

Offender Description

Descriptions of offenders have provided a common and long used method
of linking crimes. Mug shot books, while still in existence, are being replaced
by computerized photographic databases that allow for certain physical
description parameters to be used to narrow the search.

© 2000 by CRC Press LLC

This is all predicated upon the assumption that there was a witness to

the crime. There may not be one in the case of murder or arson. Even in a
sexual assault, the victim must see, remember, and accurately recall (acquire,
retain, and retrieve) the offender’s description. The ability for investigators
to obtain an accurate physical description depends upon such factors as
lighting conditions; whether the offender was masked; if the attack was from
the rear; and the level of victim trauma, stress, fright, alcohol use, forgetful-
ness, and cooperation.

Individual physical appearance is also subject to modification. Weight

changes slowly, hair more quickly, and clothing daily. Even more stable
descriptors such as age, race, height, and build will be viewed subjectively by
different victims. Ted Bundy, for example, was able to tremendously vary his
description as can be seen from photographs and police “wanted” posters
(Michaud & Aynesworth, 1983; Winn & Merrill, 1979).

The prevalence of video cameras has enhanced the ability of police to

use offender descriptions. Most banks and financial institutions, and many
businesses, transportation centres, and building lobby areas have been out-
fitted with video cameras. CCTV (closed circuit television) coverage of city
centres, parking lots, and other public places is now common in Britain.
Police agencies make routine use of the photographs and videotape from
these cameras to assist in their investigations.

Automated recognition systems for digital facial images now exist and

are used for both comparison and identification purposes (Mardia, Coombes,
Kirkbride, Linney, & Bowie, 1996). Distance and angle measurements are
taken between certain facial points (e.g., from lower point of the ear to nose
tip, the tangent to chin tip and nose tip, etc.) for both anterior (front) and
profile (side) views. Horizontal and vertical landmarks are calculated to
produce a mathematical “profile” of the face. This landmark-based data then
provides the means of determining correlations between facial images.

These systems can be connected to video cameras placed within airport

terminals and used to scan custom and immigration lines. A match with an
image in the facial database alerts officials to the possible presence of a person
of interest, such as a criminal, missing person, fugitive, abducted child, or
terrorist. Police investigators have also used computerized facial recognition
systems to assess the probability that two robbers captured on bank cameras
are, in fact, the same person. These methods increase the value of offender
image data, and echo the early work of Alphonse Bertillon in criminal iden-
tification through body measurement.

21

21 

Ear patterns are also unique and have been used successfully to identify both criminals

and victims.

© 2000 by CRC Press LLC

4.2.1.3

Crime Scene Behaviour

The behavioural analysis of crime scenes provides a third, and rapidly devel-
oping, methodology for offence comparison.  Linking crimes behaviourally
requires comparing similarities versus differences for both related and unre-
lated crimes (see Robertson & Vignaux, 1995).  Like crimes should show
more similarities than differences, and unlike crimes, more differences than
similarities.  These comparisons are usually assessed in terms of proximity
in time and space between offences, comparable modus operandi (method
of operation), and the presence of signature.

Crimes that take place close together in space and time are obviously

more likely to be connected than those separated by significant distance and
occurring years apart. This is not to say that some offenders do not travel
great distances, change residences, or interrupt their criminal activity for
personal, employment, or institutional reasons. It only means that geo-
graphic and temporal factors affect the probability of offence linkage, with
those more proximate more likely to be connected.

Modus operandi (M.O.) involves the mechanics of the crime, and can

be broken down into three chronologically-ordered stages comprising the
methods used by an offender to: (1) hunt (find and attack the victim); (2)
protect identity; and (3) escape from the scene. Modus operandi has also
been defined as those actions used by the offender to: (1) commit the crime;
(2) commit the crime; and (3) escape from the scene. It has been our expe-
rience, however, that significant offender behaviour exists in the search and
hunt for a victim, and therefore M.O. should include these activities. Pro-
tection of identity can also occur at any and all stages of the offence.

The matrix in 

Table 4.1

 helps categorize modus operandi behaviour and

can serve as a framework for comparative case analysis. It is based on the
above three stages divided into aggressive (what is done to accomplish the
crime) and defensive (what is done to protect the offender) actions:

Modus operandi is not constant, but rather varies and changes for a

variety of reasons. Like all human behaviour, it is subject to individual devi-
ation and random fluctuations. M.O. is responsive to environmental influ-
ences, such as victim reaction, the physical conditions of a crime setting, and
ongoing police activities and media attention. Displacement may result from
police activities, and can take the form of changes in the spatial, temporal,

Table 4.1    Modus Operandi Matrix

Action/Stage

Hunt

Crime

Escape

Defensive
Offensive

© 2000 by CRC Press LLC

target, tactical, or functional characteristics of the crimes (Gabor, 1978;
Reppetto, 1976). Over time, an offender’s method of operation often evolves
as the result of education, maturity and experience; fantasy progression and
development can also occur.

Unlike M.O., “signature” is constant, though certain aspects of its expression

may evolve and improve over time (Keppel, 1995; Keppel & Birnes, 1997).
Signature is defined as behaviour that goes beyond the actions needed to commit
the crime; it is a fantasy-based ritual or combinations of rituals that represent
a unique and personal expression of the offender (Douglas & Munn, 1992).
When present, it provides a useful method for establishing links between crimes
and can indicate certain underlying needs of the offender (Homant, 1998).

Staging is another consideration when examining behaviour (Douglas &

Munn, 1992). Staging occurs when the crime scene is purposely altered. This
is usually done by the offender in an attempt to mislead the police and
typically involves a criminal who knew the victim and hopes to create a
convincing alternative scenario. Crime scenes have also been changed by
family members to protect the victim from embarrassment (in cases of auto-
erotic fatalities, for example).

One of the functions of criminal profiling is the assessment of the like-

lihood that one or more crimes were committed by the same offender. In
Projects Eclipse and Kayo, a team of profilers assessed several unsolved mur-
ders of prostitutes and young women in Vancouver and Edmonton, respec-
tively, in the effort to establish possible connections between the crimes
(MacKay, 1994). Geographical analysis can also help disentangle different
series of crimes (see the discussion on parsing and crime sets in Chapter 9).

The reality, however, is that the ability of the police to link serial crimes is

limited. Establishing offence connections is often a difficult exercise, especially
so in busy urban environments characterized by high levels of crime. The
backdrop of other, similar offences can interfere with the process, and such
“background noise” can make it difficult to know who is responsible for what.
Consistency and constancy are not characteristics of many criminal offenders,
and when an analysis does indicate that certain crimes are connected, the links
are usually referred to in terms of probabilities rather than certainties. This is
made all the more difficult when multiple offenders are involved.

A further complication is that the links might be between different inci-

dent types. For example, in the case of a serial rapist who breaks into apart-
ments to attack women, police must review not just other rapes and attempts,
but also sexual assaults, residential burglaries, and prowlings. Correctly estab-
lishing a serial criminal’s pattern can involve analyzing hundreds or even
thousands of crimes.

The major response to these problems has been the establishment of

computer systems containing centralized investigative information networks.

© 2000 by CRC Press LLC

The idea of a national computer database to link murders originated with
Pierce Brooks, a retired captain of detectives in the Los Angeles Police Depart-
ment (LAPD) homicide unit (Brooks, 1984; Howlett, Hanfland, & Ressler,
1986). During his investigation of California serial killer Harvey Glatman in
the late 1950s, Brooks found he had to resort to combing various newspaper
files in order to locate murders that might fit the pattern from outside LAPD’s
jurisdiction (Newton, 1998; Ressler & Shachtman, 1992). With input from
the National Serial Murder Advisory Group, the Violent Criminal Apprehen-
sion Program (VICAP) Program was formed in the early 1980s (Keppel &
Weis, 1994).

Under the auspices of the National Center for the Analysis of Violent

Crime (NCAVC), VICAP went operational in 1985. It uses a standardized
form containing a series of behaviourally-oriented questions concerning the
crime, victim, and offender that allows for computerized matching of similar
cases from a national database. As part of its protocol, VICAP identifies
critical time and location factors, including the different types of scenes
associated with a murder: (1) victim’s last known location; (2) site of
offender’s initial contact with victim; (3) murder or major assault site; and
(4) body recovery site (Ressler et al., 1988). This allows for the calculation
of time and distance relationships, the importance of which was stressed by
the National Serial Murder Advisory Group (Keppel & Weis, 1994).

Britain and Europe have had modus operandi-based crime notification

systems for series detection — the matching of crime with crime, and offence
with offender — since 1907 (Brooks et al., 1987, 1988; Green & Whitmore,
1993; Rebscher & Rohrer, 1991). The British police developed a Crime Pattern
Analysis (CPA) computerized database within the National Criminal Intelli-
gence Service (NCIS) following the Yorkshire Ripper case. Currently, the
Serious Crime Analysis Section (SCAS), National Crime Faculty, is mandated
to conduct comparative case analysis for murders, rapes, and abductions.

Several states have also developed their own computerized crime linkage

systems, some of which feed into VICAP. Examples of these programs include
the Washington Homicide Investigation Tracking System (HITS) (also used in
Oregon and Idaho); the New York Homicide Assessment and Lead Tracking
(HALT) system; the Florida Violent Crime Investigation System (ViCIS); the
Indiana Criminal Apprehension Assistance Program (ICAAP); the Iowa Sex
Crimes Analysis System; the Michigan Sex Motivated Crime File; the Minnesota
Sex Crimes Analysis Program (MN/SCAP); the New Jersey Homicide Evalua-
tion and Assessment Tracking (HEAT) System; the North Carolina State Homi-
cide and Assault Reporting (SHARE) system; the Oklahoma Sex Crime Analysis
and Reporting System (SCARS); and the Pennsylvania ATAC Program (Cryan,
1988; Geberth, 1994; Keppel & Weis, 1993a, 1993b; Collins, Johnson, Choy,
Davidson, & MacKay, 1998). Also, six Regional Information Sharing Systems