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

© 2000 by CRC Press LLC

(Cleveland, 1985). Hickey (1997) observed that 20% of the serial killers in
his research (U.S. cases from 1800 to 1995) were black, and that they were
more likely to be involved in recent cases. Newton (1992) found 16% of the
U.S. serial killers in his study were black.

Levin and Fox’s (1985) sample included only one female and they note

that, according to FBI statistics, fewer than 7% of all simultaneous homicides
involved female offenders. All 36 sexual murderers in the study by Ressler et
al. (1988) were male, and at least one of the authors believes there has never
been a true female serial killer (Ressler, 1993; Ressler & Shachtman, 1992;
but see also Epstein, 1992). Hickey (1997), however, found 16% of the serial
murderers who operated in the U.S. between 1800 and 1995 to be female,
proportional to the involvement of females in murder generally (Holmes &
De Burger, 1988; Levin & Fox). He also notes that females were involved in
36% of the team killer cases. Females comprised 10% of the serial murderers
in Newton’s (1992) sample.

Segrave (1992) discusses 83 female serial and mass killers from around

the world who murdered from 1580 to 1990. She describes the typical female
serial killer as having murdered 17 people over a five-year period, most often
by poison. These crimes usually take place within the home of the offender
or (less often) the victim. The murders almost never occur on the street or
in a public place. Many of the victims are from powerless groups, and most
are related or closely tied to the killer (see also Hickey, 1986; Scott, 1992).

In their study of solo female serial murderers, Keeney and Heide (1994c)

found a similar lack of mobility; 13 of 14 offenders were place-specific killers,
their victims typically in custodial care (43%) or family members (37%).
They observed differences between these females and their male counterparts
in terms of behaviour patterns, methods of victim search and attack, use of
torture, crime scene organization, motive, and psychiatric disorders (histri-
onic or manic-depressive diagnoses are more likely). They suggest that
unique as well as common aetiological factors are involved in cases of female
serial murder (see also Pearson, 1997; Scott, 1992).

Hinch and Scott (1995) observe that female serial killers who act alone

generally wish to avoid attention, employing more subtle methods of murder
than their male counterparts through concealing bodies or explaining victims’
deaths by accident or illness. Perhaps this is an example of how sex roles
influence even the most extreme forms of criminality. Or maybe these women
just wish to avoid suspicion as they typically kill relatives or acquaintances.

2.1.2 Incidence, Population, and Growth

Though statistically rare,

4

 serial murder has a disproportionate impact on

communities that extends far beyond the range of the immediate victims and
their families. Changes in single indices of crime seriousness (weighted

© 2000 by CRC Press LLC

measures designed to assess the relative seriousness of criminal offences) can
be almost completely accounted for by shifts in homicide rates, despite the
fact that over 90% of all major crime is property related (Epperlein & Nien-
stedt, 1989). Also, risk of victimization and fear of crime are not spread
uniformly throughout society, but are influenced by such factors as sex, age,
race, income level, and geography (Coburn, 1988; Fattah, 1987; Kposowa,
1999; Langan & Innes, 1985; Lea & Young, 1984; Rose & McClain, 1990).
Certain groups seem to be particularly vulnerable to serial murder — pros-
titutes, young women alone, hitchhikers, children, homosexuals, skid road
derelicts, elderly women, and hospital patients (Hickey, 1990, 1997; Levin &
Fox, 1985).

Statistical analyses of the incidence and increase of serial murder must

deal with definitional difficulties and linkage problems, in addition to vari-
ations in social and historical context, police response, and official record-
keeping practices. By definition, a serial murder must be committed by a
serial murderer, but determining just exactly who is a serial murderer is
problematic. Estimates of the incidence of serial murder and the size of the
serial killer population are often inflated, inaccurate, and varied, based on
inadequate data sources and unarticulated or haphazard methodologies (see
Jenkins, 1994). By its very nature serial murder is a major news item, and
therefore often used as a political rallying point.

5

 Both Kiger (1990) and

Jenkins (1992a) insightfully discuss the role of hysteria, inflated research
claims, and media impact in the construction of social problems surrounding
serial murder.

Assume for the sake of argument that the unapprehended Seattle-area

Green River Killer has gone into retirement, never to kill again (see Smith &
Guillen, 1991). Would he still be counted in an estimate of the serial murderer
population? And should a potential serial killer, apprehended and incarcer-
ated after his first homicide, be included in such estimates (Bishop, 1946;
Boyd, 1988)? Aleksandr Koryakov stabbed three children and their teacher
in a single act of murder before being apprehended by Latvian police, qual-
ifying as a mass but not a serial killer (Latvian kindergarten killer, 1999). But
he later confessed this was only the start of a homicide spree intended to
rival that of Andrei Chikatilo, the Rostov Ripper, who was convicted of 53
murders in Russia. It is assumed that “serial killer” is a status theoretically
capable of capture in a count done at a single point in time. This is a
problematic assumption.

4 

All forms of homicide are uncommon and most killings involve intimates, as opposed to

strangers (Boyd, 1988; Fattah, 1991; Silverman & Kennedy, 1993; Skogan & Antunes, 1979).
Reed and Gaucher (1976) note “the amount of repetitive killing in Canada is miniscule
and certainly far less than appears to be generally believed” (p. 1).

5 

This is nothing new.  See Rumbelow (1988) for the response of the lower class to the

Whitechapel murders of Jack the Ripper in 1888.

© 2000 by CRC Press LLC

Linkage problems prevent many incidents of serial homicide from being

recognized as such (Holmes & De Burger, 1988; Norris, 1988). Murders
committed in different locations and separate police jurisdictions may not
be seen to be part of a single series. Even those killings that occur in the same
city might not be connected if there is a high murder rate, an overworked
homicide squad, or if the crimes are spread out over time.

The only official records in some cases may be missing person reports.

Without a body, it is doubtful that a homicide statistic will ever be recorded.
Many serial killers, including Jeffrey Dahmer, John Wayne Gacy, Jr., and Juan
Vallejo Corona, buried or hid the bodies of their victims. Several of the groups
at high risk — prostitutes, the itinerant, the destitute — may go missing
without anyone noticing. With no complaint, there is no official missing
person report.

Despite these difficulties, several researchers have attempted to estimate

the prevalence and extent of serial homicide. Leyton (1986) states that there
may be up to 100 of these murderers operating in the U.S., responsible for
the deaths of thousands of people. He expresses some surprise at this rela-
tively low number considering the present violent state of society. Holmes
and De Burger (1988), however, estimate that there are currently 350 serial
killers in the U.S. alone, responsible for between 3500 and 5000 deaths
annually. Norris (1988) states he has researched 260 serial murderers, respon-
sible for 10,360 victims (for a comparatively high average of 40 victims per
murderer), though he does not provide a list of who these killers are.

Jenkins (1989) found several hundred serial murder cases, 49 of which

involved 10 or more victims, recorded in newspaper archives since 1971. The
U.S. Justice Department conservatively estimates there are 35 active serial
killers operating within the U.S. (Levin & Fox, 1985), though their method-
ology has never been published. Jenkins suggests that there are approximately
400 victims of serial murder per year, while Fox and Levin (1992) have
calculated that fewer than 240 people fall prey to serial killers annually in the
U.S. Hickey (1997) found, for the peak period of 1975 to 1995, only 7.7
reported new cases (49 to 70 victims) per year.

6

 The Federal Bureau of

Investigation Behavioral Science Unit (BSU) has stated they receive approx-
imately 30 requests annually for assistance in serial murder cases (Hagmaier,
1990).

Kiger (1990) has critically evaluated the various methods used to estimate

the incidence of serial murder, and points out problems with unexplained

6 

The U.S. Justice Department’s estimates are for active murderers, while Hickey’s are for

new cases. These figures do not appear so dissimilar when it is realized that the median
number of years of homicidal activity for the serial killers in Hickey’s sample was 4.3 years
(1991). Jenkins (1988b) estimates a career length of just under 4 years for serial murderers
in England.

© 2000 by CRC Press LLC

methodologies and calculation procedures. In the U.S., the current UCR
system does not allow the tracking of single offenders, and Supplemental
Homicide Report (SHR) incident-based data, which often suffer from miss-
ing information, only allow counts of mass, not serial, murder. Some
researchers have used the sharp rise in the number of stranger and unknown
motive homicide categories in the SHR as the basis for evaluating serial
murder increases. This is an unsound extrapolation; increases in felony mur-
ders (particularly those that are drug related and difficult to solve), urban-
ization levels, and demands upon police departments all affect uncleared
murder rates (see Cardarelli & Cavanagh, 1992). Some estimates have con-
flated unknown and stranger relationship homicide categories, leading to
high estimates of the latter. Advocacy numbers claim 50% of all murders are
stranger offences, but 18 to 25% is a more accurate estimate (Riedel, 1998).
Jenkins (1988a, 1994) and Kiger (1990) have criticized the manner in which
these data have been used to estimate the occurrence of serial murder.

Mortality data from the National Center for Health Statistics (NCHS)

Division of Health Statistics have been used to check UCR homicide data.
This information is not particularly useful for estimates of serial murder, but
Kiger (1990) notes that coroner data on the number of unidentified bodies
found each year, while not completely congruent, could be relevant. The FBI
National Crime Information Center (NCIC) keeps a computerized record of
such bodies. It also lists missing persons as does the Canadian Police Infor-
mation Centre (CPIC) operated by the Royal Canadian Mounted Police
(RCMP). While calculations of the number of serial murder victims, partic-
ularly children, have been based on this data, there is a problem estimating
the percentage of cases involving hidden bodies.

The number of children who go missing for reasons not related to paren-

tal abductions is relatively small. In contrast to inflammatory statements
made by certain special interest organizations, Kiger (1990) estimates some-
where between 20 to 300 children in the U.S. are murdered annually by
strangers (see also Allen-Hagen, 1989; Finkelhor, Hotaling, & Sedlak, 1990).
Between 1984 and 1988, only 3% of the 16,511 cases of missing and murdered
children reported to the U.S. National Center for Missing and Exploited
Children were abducted by strangers (Hickey, 1990). Of these 495 children,
15% (75) were found dead, and 47% (235) are still missing. Presumably,
some unknown fraction of these murders were committed by serial offenders.

Cavanagh (1993) presents a method for estimating the number of serial

murder victims from official sources. For the period from 1976 to 1989, he
aggregated SHR data from census subregions into 45 records (9 subregions
by 5 time periods). The different types of homicide that might reasonably
be related to serial murder were totaled for each record, and the combination
that best predicted serial murder determined through regression analysis.

© 2000 by CRC Press LLC

The homicide categories included stranger or unknown, bodily contact, fel-
ony, and sexual murders. The model was calibrated against recorded counts
of serial murder victims collected by Newton (1990a, 1990b). Cavanagh
cautions his model is a preliminary exploration, and notes weaknesses in
Newton’s enumeration that could bias the results. Nevertheless, he has devel-
oped a feasible approach, with an articulated methodology, for estimating
rates of serial murder from SHR and census data. This is a significant
improvement over inflated estimates and impressionistic guesswork.

While serial murder is not a problem unique to the latter half of the

20

th

 century, most researchers agree the number of such killings is increas-

ing (Hickey, 1997; Holmes & De Burger, 1988; Levin & Fox, 1985; Leyton,
1986; Newton, 1992; Norris, 1988; Ressler, Burgess, & Douglas, 1988; see
also Boyd, 1988). Based on a comprehensive study of 399 serial murderers,
Hickey (1997) found a ten-fold increase in active offenders per year for the
period 1970 to 1995, compared to 1800 to 1969.

7

 He notes the number of

victims per case has dropped, likely due to increased police efficiency (see
also James, 1991). Since 1925, this figure has been in the range of 7 to 13
total murders per case. Jenkins (1988b) estimates that five to six is a likely
annual total of victims for recent American serial murderers, and four for
English serial killers. Fox and Levin (1992) suggest that while six victims
per year is typical, the figure should actually be doubled (to 12) to account
for unknown victims.

Jenkins (1988a, 1989, 1992b) has analyzed cases of “extreme” serial mur-

der (involving ten or more victims) in the U.S. for the period from 1900 to
1990, and concludes that “serial homicide was a common experience in the
early years of this century, little less frequent than in recent years” (1989,
p. 378). He does find evidence, however, of a “serial murder wave,” particu-
larly involving lust killings, in the U.S. from 1940 to 1990. He documented
11 cases of lust murder from 1950 to 1964, 11 cases from 1965 to 1969, and
12 cases in 1973. While recognizing the actual scale of the problem may have
been blown out of proportion to benefit certain ideologies or interest groups,
he still concludes that “there simply were more serial killers, more of whom
could be categorized as lust murderers. A new problem was identified because
a new problem had come into being” (1992b, p. 13). Jenkins lists 49 mur-
derers active in an 18-year period, a four-fold increase from the first part of
the century (not adjusted for population growth). There is also historical
evidence of this type of crime in both England (1988b) and Germany (1988a).

7 

These figures are not adjusted for population growth; per capita serial murder rates have

only doubled over the last few decades.  This estimate may also be temporally biased as
the more recent a case, the greater the probability it will be located by a researcher in
newspaper files or other written works.