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

© 2000 by CRC Press LLC

Geography Of Crime

“Murder Alley” was the name given to an unpaved lane near downtown
Kenosha, Wisconsin, where seven bizarre homicides took place between 1967
and 1981 (Newton, 1990a). The frequency and unusual nature of the crimes
— including a corpse found in a hearse, another buried beneath a rose
garden, and a triple homicide — prompted investigators to publicly comment
that something strange was happening in this “Bermuda Triangle of murder.”

During its early years, Los Angeles was a violent frontier town. In 1850,

the year California joined the Union, “Los Diablos” experienced a murder a
day; with a population of 4000, this translates into a homicide rate of 1
murder per 11 residents (Blanche & Schreiber, 1998). Most of these murders
took place close to the central plaza in a narrow alley slum named Calle de
los Negros (“Nigger Alley”), a collection of shabby buildings, housing
saloons, brothels, and gambling dens. Shootings, riots, lootings, and lynch-
ings were common activities.

“What surfaces in terrible places... is a rapport between persons and

places such that the techniques of the habitat and forms of personation
become indistinguishable” (Seltzer, 1998, p. 233). Routine activity and ecol-
ogy of place theories provide an explanation for the phenomena of dangerous
places or “dreadful enclosures.” The related concepts of “fishing holes” and
“trap lines” — used by serial killers and rapists in the hunt for victims —
also help explain the clustering of predatory murders and rapes. The geog-
raphy of crime encompasses the study of the spatial and temporal distribution
of criminal offences.

The methodological and theoretical approaches developed in a variety

of disciplines are often adopted to the study of crime and criminals, and the
fields of geography and urban analysis provide several analytical tools for
criminologists. Perspectives that make particular use of geographic tech-
niques include the social ecology of crime, environmental criminology, geog-
raphy of crime, routine activity approach, situational crime prevention, and

7

© 2000 by CRC Press LLC

problem-oriented policing (Brantingham & Brantingham, 1981b; Branting-
ham & Jeffery, 1981; R. V. Clarke, 1992, 1997; Felson, 1986; Goldstein, 1990;
Lowman, 1986; Smith, 1986). Work in such areas has provided the conceptual
foundations for geographic profiling.

7.1 Geography and Crime Studies

Geographical patterns in crime have been noted since the mid-19

th

 century

pioneering work of Andre-Michel Guerry and Lambert-Adolphe Quetelet
who mapped, on a national basis, violent and property offences and exam-
ined their spatial relationship to poverty (Brantingham & Brantingham,
1981c; Vold & Bernard, 1986). The most famous spatial crime studies were
conducted in the early 20

th

 century, when the city of Chicago served as an

inspiration source and a field of experimentation for University of Chicago
sociologists (Warren, 1972; Williams & McShane, 1988). The geographical
focus of criminology had shifted from regional areas to city neighbourhoods.

The Chicago School’s human ecology and theories of urban growth,

developed by Robert Park and Ernest Burgess, served as guides for former
probation officer Clifford R. Shaw and his colleague Henry D. McKay (Brant-
ingham & Brantingham, 1984; Vold & Bernard, 1986). While working for
the Illinois Institute for Juvenile Research, Shaw and McKay examined a
plethora of urban social ills, including the politically important issue of crime
(Williams & McShane, 1988; Wilson, 1984). Much of this study was done
along the margins of society; researchers observed, talked to, and examined
criminals, gang members, hoboes, immigrants, and slum dwellers. Most
importantly, they did this within the “natural” urban contexts associated with
the people they were studying. This attempt to triangulate the subjective
human perspective with more objective demographic statistics exemplified
the work of the Chicago School (Smith, 1986).

While there is a long tradition of spatial crime analysis, the ultimate units

of concern are individual human beings and their day-to-day social interac-
tions, not census statistics or demographic data. This point seems to have
been forgotten during the unfruitful positivistic geographic research of the
1960s, when complicated but theoretically weak factor analytic methods were
used to correlate a spectrum of spatially distributed variables (Smith, 1986).

Recent developments in the spatial study of crime have been more prom-

ising. The integration of geographical perspectives, urban planning tactics,
environmental criminology, and ecological approaches based on econometric
studies has led to a series of interesting studies (Brantingham & Brantingham,
1981a; Lowman, 1986; Smith, 1986). While the field has been primarily
instrumentalist in nature, writers such as Lowman and Smith have com-

© 2000 by CRC Press LLC

menced the development of a critical geography of crime, and left realists
are now concerned with the situational and spatial nature of crime (Kinsey,
Lea, & Young, 1986; Lea & Young, 1984).

Geographic concepts and terms are important in understanding criminal

target patterns and constructing crime site selection models. They also pro-
vide the elements necessary to develop methods of offender residence pre-
diction. Some of these ideas are presented below.

7.1.1 Journey-to-Crime Research

The journey-to-crime literature includes numerous studies of crime trips for
a variety of offence types in several North American and European cities.
This research has often explored the influence of such offender and offence
characteristics as sex, race, age, prior criminal experience, nature of home
area, crime type, target area attributes, and perceived level of gain (Baldwin
& Bottoms, 1976; Capone & Nichols, 1975; LeBeau, 1987a; Nichols, 1980;
Reppetto, 1974; Rhodes & Conly, 1981).

These studies have produced several common findings, including the

following observations:

•

Crimes often occur in relatively close proximity to the home of the
offender. “While criminals are mobile, they don’t seem to go very far
in committing a crime. A majority of crimes appear to take place
within a mile of the criminal’s residence” (McIver, 1981, p. 22). Time
is a commodity and almost all people act in a manner to conserve its
use (Brantingham & Brantingham, 1984; LeBeau, 1992; Rebscher &
Rohrer, 1991; Rengert & Wasilchick, 1985; Rhodes & Conly, 1981).
This pattern is consistent with the nearness and least-effort principles.

Table 7.1

 presents a summary of the findings from journey-to-crime

research.

•

Crime trips follow a distance-decay function with the number of crime
occurrences decreasing with distance from the offender’s residence
(Brantingham & Brantingham, 1984; Capone & Nichols, 1975; Rhodes
& Conly, 1981). This pattern is similar to that exhibited in other forms
of human movement (Jakle, Brunn, & Roseman, 1976).

•

Juvenile offenders are most likely to commit their crimes within their
home area, and are less mobile than adult offenders (Baldwin & Bot-
toms, 1976; Gabor & Gottheil, 1984; Warren et al., 1995).

•

Differences in crime trip distances between offence types have been
consistently found. Violent crimes, for example, usually occur closer
to the offender’s residence than do property offences (Baldwin &
Bottoms, 1976; Gabor & Gottheil, 1984; LeBeau, 1987a; Rhodes &
Conly, 1981).

© 2000 by CRC Press LLC

•

Most cities contain high crime rate neighbourhoods, the arrangement
and location of which influence the patterning of crime trips (Gabor
& Gottheil, 1984; Rhodes & Conly, 1981).

A crime trip, or the journey to crime, is usually defined by some direct

or surrogate measure of distance between offender residence and location of
offence. This may or may not have been the actual journey taken by the
offender who could have originated his or her travel from work, the house
of a friend, or the local bar. For example, in Pettiway’s (1995) study of
Philadelphia crack cocaine users, 25.7% originated their drug buying trip
from home, with street corners and outside hangouts (22.9%), friends and
relatives (10.7%), shopping and business (10.8%), recreation (7.4%), and
other places (22.4%) comprising the remaining origins. Of significance, how-
ever, is the finding that, regardless of origin, most destinations were in the
offender’s neighbourhood (61%). Almost all trips (93%) began and ended
within the city limits of Philadelphia. These results “suggest that even when
individuals engage in activities away from their places of residence, they tend
to return to locations near those places” (p. 515).

In a study of robbery in northern California, Feeney (1986, 1996) found

70% of offenders robbed in their home towns, and over one third in their
own neighbourhoods. Of the 30% who offended in a different town, only
half went there to commit a robbery; the rest were there for some other
reason. Those who commute to rob almost always choose a contiguous town.
Robbery is a casually planned crime, and the study found fewer than 5% of
offenders make detailed plans such as observing the layouts of prospective
targets, establishing contingencies, or predetermining escape routes (Feeney,
1986). Victims are chosen because of convenience, appearance of money, low
risk, and the possibility of a fast getaway. Or, in the words of one robber,
“‘Just where we happened to be, I guess’” (p. 92).

Canter and Gregory (1994) analyzed crime trip patterns for 45 serial

rapists responsible for 251 offences committed during the 1980s in Greater
London and Southeast England. The mean number of rapes per offender in
this sample was 5.6 (standard deviation = 3.6), with a minimum of 2 and a
maximum of 14. It was theorized that offenders with greater access to
resources will exhibit longer journeys to crime. Therefore, rapists with more
geographic knowledge, money, and time have the capability of travelling
further distances to commit their crimes. Canter and Gregory (1994) propose
that criminals with greater resources (geographic knowledge, financial, time)
travel further to offend. They analyzed crime trip distances for a sample of
serial rapists using dichotomized parameters of race, venue, weekday, and
offender age. It was suggested an expert system for determining area of
criminal residence could be based upon this approach. The basic viability of

© 2000 by CRC Press LLC

the system remains unknown, however, as the study negated its results by
using the same data set for both learning and testing.

White rapists were found to travel further to commit their crimes than

black rapists, the latter having an 80% chance of living within a half mile of
one of their offences. Rapists who chose outside venues typically journeyed
2.7 times as far as those who chose inside or mixed venues. Weekend rapists
averaged minimum crime trip distances 2.5 times as great as weekday offend-
ers, though this difference was not statistically significant. Older rapists (over
25 years of age) appeared to travel further than younger rapists, but this
finding also was not statistically significant. The percentages for each group
found to have travelled under a half mile from their home base to their first
known offence were: blacks, 74%; whites, 18%; inside/mixed, 70%; outside,
28%; weekday, 41%; weekend, 50%; mixed, 60%; under 25 years of age, 54%;
over 25 years of age, 38%; and total, 47%.

Results from journey-to-crime studies are usually reported in one of four

manners: (1) mean crime trip distances; (2) medial circles; (3) mobility
triangles; and (4) distance-decay functions. Studies using mean crime trips
compute the arithmetic average for the distance travelled by offenders from
home to offence location. Some researchers employ the geometric mean to
avoid the distortion of extreme values (LeBeau, 1987a), while others use
additional descriptive statistics such as the mode (Rhodes & Conly, 1981),
maximum crime trip distance (Canter & Larkin, 1993), minimum crime trip
distance (Warren et al., 1995), range (Turner, 1969), and directional infor-
mation (Rengert & Wasilchick, 1985).

The major problem with the use of the mean or mode is the limited

information a univariate statistic provides. The mean is also susceptible to
the influence of spatial outliers and may say little about the typical journey.
And as an average distance is produced from an aggregation of trips involving
several offenders, it might not accurately describe any one type of offender.

Studies using medial circles define a radius containing a certain percent-

age of offences.

34

 In a study of robbery within Boston, Reppetto (1976) found

90% of the offences occurred within 1.5 miles of the offender’s residence. As
with mean crime trips, medial circles supply only a single measure of crime
journeys, says little of the typical crime trip, and expresses almost nothing
about cases that fall outside of the set radius.

Mobility triangles were first used by Burgess (1925) to describe situations

where offence location and offender residence were in different neighbour-
hoods; this was in contrast to neighbourhood triangles where the crime takes
place in the offender’s home area. Subsequent developments of the mobility
triangle concept adopted various combinations of crime scene, offender res-

34 

If the circle contains one-half of the offences, the radius is equivalent to the median

crime trip distance.