GEOGRAPHICAL PROFILING

Over the last two decades numerous research studies using the theory of mental maps and home range have been applied to study criminals’ spatial behavior. These studies reflect the importance of the journeys criminals habitually take around the areas close to their homes. The research also points out that such journeys provide criminals with information around which they are likely to plan their next crime; for instance, locations frequently passed by the criminal while traveling home, such as bars, shops and restaurants, are therefore proposed as defining the criminal’s range and thus tuning his or her perceptions as to which areas are likely to be safe, both geographically and psychologically.
The concept of home range is based upon the notion that there is a geographical area around our homes in which we travel and use more regularly than areas further away. This area would typically include shopping areas, homes of friends and relatives and social activity areas. In 1913, it was suggested that home range can be thought of as a kind of cognitive or mental map; that is, images strongly related to residential location. Cognitive maps are representations of what is possible and where, built up through our daily activities and social transactions. As such, the mental maps we draw of an area change over time, and reflect how much time we spend in an area and the variability of our purposes for being there.
Each of these spatial approaches has a defining role in the specific aspect of geographical behavior investigated; for example, the location of the offender’s home base; the location of the crime; and the relationship between the offender’s home base and the crime scene. It is the latter area which has the most significance in the emerging field of geographical profiling. The identification of an offender’s home base and the distances that the offender travels from home to commit crime is of value to police investigators who are investigating a series of crimes, for example, serial murder, rape or arson.


The Ecological Approach

Approaches to the study of geographical movements of criminals originated from what has become known as the ecological tradition, which developed in America between 1900 and through to the early 1970s. The ecological tradition is closely linked to a particular theoretical concept originally developed at the Chicago School of Sociology. Research in the Chicago School of Sociology in 1929, 1931 and 1969 exemplified the ecological principles. Using the ecological principles in 1942, Shaw and McKay confirmed the geographical coincidence of a number of social problems with high rates of delinquency and criminality. For example, they found that crime rates in American cities tended to be highest in the wealthier areas that were juxtaposed to poorer areas where there were identifiable concentrations of criminal residences. Shaw and McKay demonstrated the persistence of areas with a high rate of criminality over time and through changes in the ethnic make-up of these cities. Their research findings had a major impact on police procedures, and consequently have become accepted as facts. Shaw and McKay’s studies have influenced all subsequent sociological theory dealing with the geography of crime and delinquency.
However, the confusion in Shaw and McKay’s ecological principles is that they assumed that the characteristics of an area with high proportion of criminal residents also identified and described the social characteristics of individuals who were likely to commit crimes. The attempt to use general data as an explanation for individual behavior has been termed the ecological fallacy by criminologists critical of the Chicago School of Sociology. As Brantingham and Brantingham pointed out in 1981, the problem comes when ecological data on the location of criminal residences is expected to answer questions about criminal motivations. The weakness of the ecological approach is that its findings are based on patterns of association between crime and its potential causes at an aggregate level and do not consider individualistic offending data. However, the ecological studies of the Chicago School of Sociology did draw attention to the potential of studying the spatial distribution of various urban, social and criminal indices.

Environmental Criminology

Deriving out of the ecological tradition, Brantingham and Brantingham in the late 1970s termed the phrase environmental criminology. Environmental criminology is concerned with criminal mobility and the relationship between offenders’ home bases and their target areas. Environmental criminology attempts to predict the geographical area that an offender will victimize, based not on demographic features but on the individual’s own mental image of the area. The Brantinghams proposed a theoretical spatial model for looking at journeys to crime as they occur in urban space. The Brantinghams’ model uses concepts of opportunity and offender motivation together with the concepts of mobility and perception to predict the next likely target area and offender’s resident. Table 1 lists six major components that make-up the Bran-tinghams’ hypothesized spatial model.
One element not shown in Table 1 is the spatial patterns of hunting and target selection. However, in an attempt to fill this void, the Brantinghams expanded their model using theoretical cases. The simplest case scenario and subsequently the foundation for current geographical profiling systems, is the basic search area for an individual offender. An example of the hypothesized search area can be seen in Fig. 1. The initial conditions in Fig. 1 are single offender, uniform distribution of potential targets and the offender based in a single home location.
The Brantinghams refer to the phenomenon of distance decay in describing the location of an offender’s crimes. Briefly, distance decay refers to the reduction of activity or interaction as distance from the home increases. In the hypothesized case shown in Fig. 1, the expected range would be circular and most offenses would occur close to home, with the likelihood of an offense taking place in a particular location decreasing with distance from home.

Table 1 Major elements of the Brantinghams’ hypothesized spatial model

• Individuals exist who are motivated to commit specific offenses
• The environment emits many signals about its physical, spatial, cultural, legal and psychological characteristics
• The commission of an offense is the result of a multistaged decision process that seeks out and identifies a target or victim positioned in time and space
• As experiential knowledge grows, an individual learns which individual cues, clusters of cues and sequences of cues are associated with good victims or targets.These can be considered a template, which is used in victim selection
• An individual who is motivated to commit an offense uses cues from the environment to locate and identify targets
• This template becomes relatively fixed and influences future behavior
The Brantinghams suggest offenders will have more cognizable mental maps about potential crime opportunities close to their home bases. They also proposed that offenders are more likely to be noticed and identified close to their home bases by other individuals who live in the same vicinity. Consequently, the Brantinghams argue that there would be a ‘confront zone’ directly around the offender’s home base where little criminal activity would occur; this is referred to as a ‘buffer zone.’ For example, in 1980 it was suggested that it takes effort, time and money to overcome distance; these things are what have been termed the ‘friction’ of distance.
The Brantinghams' hypothetical offense area for a single offender.
Figure 1 The Brantinghams’ hypothetical offense area for a single offender.
In later research, the Brantinghams refined their hypothesized spatial model and proposed a complex search area for offenders. An example of this hypothesized model is shown in Fig. 2. The initial conditions allow for the fact that criminals, like noncriminals, are not tied just to geographical locations near their home base. Rather, offenders, like nonoffenders, go to work, shop and relax and the pathways between them all combine to form, what the Brantinghams termed, the individual’s awareness space. Arguably, the Brantinghams’ hypothesized spatial theories suggest that, given equally distributed opportunities, offenders will tend to offend within a minimum and maximum range of distance from their home, independent of direction and other physical or psychological constraints. Research that tends to echo this hypothesized spatial model was carried out in 1974; for example, it was found that in Cleveland, Ohio, property offenders tended to travel further from home than personal offenders did. Conversely, another study in 1982 found that black robbers and not burglars traveled longer distances to their crime sites. However, there are exceptions to these findings. For example, a 1969 study did not find any significant differences based on offense type and the distances between the offender’s residence and crime location for 1960 Philadelphia data. Also, a study as far back as 1955 showed that 40% of all Houston homicides, including domestic homicide, between the years 1945 and 1949 occurred within one city block of the offender’s residence.

Computerized Geographical Profiling

Following on from the work of Brantingham and Brantingham, Rossmo, calling his technique criminal geographical targeting (CGT), has combined concepts from environmental criminology with a mathematical model, based on a distance decay function, derived from the locations in which killers leave their victims’ bodies, to indicate the area in which an offender may be living. The psychological principles on which Rossmo’s work is based are not articulated in any detail but appear to derive from the postulate propounded by Brantingham and Brantingham’s research suggesting that victims are probably spatially biased toward the offender’s home base. This theory was illustrated in a study by the Brantinghams in 1981 in Washington, DC, where they found that offenders generally victimized areas they know best, concentrating on targets within their immediate environments and surrounding areas. This spatial bias is the proposed cause of a decay function, such that the further an offender is from home the less likely he or she is to commit an offense.
The Branginghams' hypothetical offense area for an individual offender, taking into account his or her action space. Arrows indicate movement paths.
Figure 2 The Branginghams’ hypothetical offense area for an individual offender, taking into account his or her action space. Arrows indicate movement paths.
The reasons for the proposed decay proposed in Rossmo’s criminal geographical targeting model are not exactly clear but appear to be based on the least-effort principle. As defined by one author in 1950, least-effort principle postulates that when multiple destinations of equal desirability are available, all else being equal, the closest one will be chosen. However, the notion that criminals choose targets in their confront zones is usually modified by two further considerations. One is most readily supported by the 1982 study referred to above, which found that specialist selectivity is reflected in the effort an offender puts into planning a crime. This choice would lead to more selective and more carefully planned crimes being committed further away from home. This theory has been supported by comparisons across different types of crime. For example, a series of studies in 1975 and 1976 found that the apparently more impulsive crimes of rape are committed nearer to home than robbery, and that armed robbers travel further on average than those who are not armed, and tend to net larger sums of money.
Rossmo incorporates another principle into his geographical profiling system, which has been put forward as a basis for crime locations: that there will be a tendency for serial killers to avoid committing crimes close to where they live, which was referred to earlier as a buffer zone. The proposed reason for this so-called ‘confront zone’ is so that criminals will avoid leaving incriminating evidence near to where they live. However, the evidence for this is sparse. For example, a 1997 study of 54 American serial killers found that offenders, on average, tended to make initial contact with their victims closer to home than the locations in which they eventually placed the bodies, which could suggest that a buffer zone is highly probable for body disposal sites but not for the abduction sites. A study in 1995, for example, found no evidence for a buffer zone in a limited study of single rapists. Furthermore, the drawback to relying on the distance decay theory is that the actual distances proposed as buffer zones are often larger than would be consistent with leaving local clues.
A variety of geographical profiling processes have thus been outlined by those with an ecological and environmental criminology bias and these processes could be seen to be logically in conflict. One is a tendency to minimize effort close to home, which would predict that crimes are in a closely circumscribed area. A second is the tendency to keep a minimum distance away from home. These two processes combined would lead to the prediction of an optimum distance from home to all cases of a particular type of offense. However, the general finding is one of an aggregate decay of the frequency of crimes as their distances increase from home. These processes are derived from a consideration of instrumental crimes often with a tangible material benefit, such as theft or robbery. So, although they doubtless have relevance to geographical profiling, there are questions about how important emotional issues are ignored by such rational models. For example, two further complexities raise questions about the relevance of these rational processes to geographical profiling criminals. The present author has pointed out that there is typically more than one location involved in serial killers’ activities. Besides the location of the site where the victim’s body is dumped, there is also usually at least one other important site, the point at which the victim is first encountered. For example, in a series of murders, all three of the processes indicated above, least effort, buffer zone and decay function, would likely predict that the abduction and body dump sites would be close together. However, research on 54 serial killers by this author found that the locations where victims go missing were on average 1.5 miles (2.4 km) from the offenders’ home bases compared with 5 miles (8 km) for the body dump locations. Furthermore, none of the three processes would lead to any predictions in the changes of the distance of the crimes from home over time.

Environmental Psychology

Following on from the environmental criminological theories proposed by Brantingham and Brantingham is the emerging field of environmental psychology. Environmental psychologists see the journey to crime as an expression of a complex interaction between the offender, his or her background characteristics, predispositions, knowledge and perceptions, and the location and type of target, in terms of perceived risks, rewards, opportunities and attractions. For example, in 1989 environmental psychologist, Canter, hypothesized that the actual nature of the location selected may be indicative of the purpose and experiences of the offender. Canter pointed out that there may be patterns of space use typical of different criminals, relating to where they are living at the time of their crimes.
Using an environmental psychology approach, Canter and his colleague in 1994 proposed research into the relationship that may exist between the criminal range and the location of the home base of serial sexual offenders. The proposed model of individual sexual offenders’ spatial activity was based upon 45 British male sexual assaulters who had committed at least two assaults on strangers. The study hypothesized two general models to characterize the relationship between the home base and criminal area of offenders. The first assumption that was suggested regarded the geometry of a criminal’s domain: that it would be an area defined by a circle around the offender’s home, which Canter and his colleague defined as the commuter hypothesis. According to this theory, the area around the home and the area in which the crimes are committed are represented as circles. An example of the commuter hypothesis is shown in Fig. 3.
In describing the commuter process, Canter suggests that an offender travels from his or her home base into an area to offend. Central to this hypothesis is that there will be no clear relationship between the size or location of the criminal’s domain and the distance it is from an offender’s home. As such, the commuter hypothesis model proposes little or no overlap between these two areas, suggesting that the offender moves an appreciable distance to a district outside his home range to offend.
The second model proposed by Canter and his colleague is based upon the marauder hypothesis. An example of this hypothesized model is shown in Fig. 4. This theory argues that the offender’s home base acts as a focal point for his or her crimes; it is suggested that the offender moves out and returns to the home base each time he or she commits a crime.
muter hypothesis.Open circle, home base; solid circle, offense.
Figure 3 Commuter hypothesis.Open circle, home base; solid circle, offense.
Marauder hypothesis.Open circle, home base; solid circle, offense.
Figure 4 Marauder hypothesis.Open circle, home base; solid circle, offense.
Arguably, this hypothesis would predict that, on average, the further the distance between crimes, the further the offender must be traveling from home to offend. However, research on 54 serial murderers found that the marauding process may be viable for victims’ body dump locations but not for their abduction sites.
Canter’s study found that 91% of the sample of offenders had all their crimes located within the circular region, which was termed the circle hypothesis. It was also found that 87% of the offenders had a base within the circle hypothesis prediction area. These results provided strong support for the general marauder hypothesis as being most applicable to this set of offenders; that is, that the majority of the sample had operated from a home base from which they traveled to commit their crimes.

Home Range and Directional Travel of Criminals

There is a further complexity about environmental criminology, Rossmo’s geographical profiling model and the circle hypothesis processes to consider; that is, the hypothesized spatial models fail to take into account directional bias. For example, research in 1985 demonstrated that crime trips of suburban burglars are more likely to be skewed in the direction of nodal centers, such as workplaces and city central districts, and that crime trips were more likely to be wedge-shaped. Considering the plethora of research literature on the sectoral mental maps (directionality) of intraurban travel patterns and migration of people, the consideration of directionality in geographical profiling has been mute. One author argued that intraurban migration takes place in accordance with sectoral and directional biases in the urban spatial cognition. The findings suggested that intraurban migration patterns are sectorally biased towards the center business districts, in that they are wedge-shaped.
The theoretical analysis of people’s bonds with the tangible surroundings of the home environs is found in several disciplines; for example, migration and shopping behavior studies. It has been pointed out that, through daily travel, the home environment becomes a unique place of familiar, known and predictable activities, people and physical elements, a focal point of one’s experiential space. Thus, through habitual, focused and satisfying involvement in a residential locale, the tangible home area becomes an enduring symbol of self, of the continuity of one’s experiences, and of that which is significant and valued by the inhabitant. The landscape around the home base may thus be hypothesized to provide criminals with those enduring symbolic experiences. If their crimes, as hypothesized, do indeed develop as an elaboration of their daily activities, rather than as some distinct work-life activity, then it would be predicted that the home would be geographically as well as symbolically central to their criminal activities.

Geographical Profiling the Angles Between Crimes Using Predator

Predator is a geographical profiling system developed from research carried out by this author. The Predator system employs two location-allocation theories. Firstly, it employs a method that looks at the location of central facilities and the allocation of flows to them. Such a method projects, theoretically, the probable trips of the offender to the center. Secondly, Predator employs an analysis of a dispersing offender and reverses the directional assumption of the location-allocation model, while also keeping the assumptions of a monotonous plain and of optimization. Rather than assuming that the victim travels to the offender’s home base, it assumes that the offender travels outward from his or her home, so the ‘allocation of flows’ is ‘from’ rather than ‘to’ the center. This process allows the criminal’s spatial movements to be modeled from all directions; no matter if they are moving in or out.
Mapping the crime data involves recording each killer’s crime locations and home base and their respective latitude and longitude geographical coordinates. Once this task is completed, all the geographical coordinates of latitude and longitude are converted into The Universal Transverse Mercator (UTM) grid system. Converting the latitude and longitude coordinates into the UTM coordinate system allows the data to be entered into the Predator geographical profiling system. Each distance is then entered separately into the Predator system. The program then creates on the computer screen a scale of the actual crime area in miles or kilometers. For each event, the northing and easting UTM coordinate serves to express the unique, absolute location of each crime event. Equally important, the UTM coordinate system is impartial, and the values of locations are independently verifiable. The geographical program then plots each of the crime locations on the computer screen. Briefly, each case is analysed separately.
By way of example and demonstration of how offense angles can be calculated, Fig. 5 shows the cognitive maps of four American serial killers that were calculated by the Predator system. Briefly, the data were collected from the Homicide Investigating Tracking System (HITS) based in Seattle, Washington. In instances where there was more than one home base, the mean of the angles between the crimes is reported. The finding that serial killers’ spatial behavior has a tendency to be wedge-shaped has several implications for geographical profiling and predicting the likely home base area of a serial killer. For example, in the study of British sex offenders discussed earlier, the home base was found 71% of the time between the two furthest crime sites, as Canter’s circle hypothesis suggests. However, there is still a rather substantially large search area for the police to investigate. If police were to rely on the wedge-shaped theory as a function in geographical profiling, the search area including the offender’s home base would be considerably smaller. It is therefore argued that the wedge-shaped theory has heuristic value when developing geographical profiles of serial killers. For example, research on serial rape conducted in 1993 found that crime patterns had a distinct sense of directionality that could be described as a windshield wiper pattern.
Cognitive maps of killers' movements between home bases, victims' abductions and body dump sites. (A) Offender 1: 1 inch = 1 mile (2.54cm = 1.6 km); 130°. (B) Offender 2: 1 inch = 10 miles (2.54cm = 16km); 33°. (C) Offender 3: 1 inch = 20 miles (2.54cm = 32km); 246°. (D) Offender 4: 1 inch = 10 miles (2.54cm = 16 km); 162°. O, home base; solid circle, dump site; ■, abduction/last seen site; numbers represent order of crimes.Note: Difference in circle sizes is due to map magnification.
Figure 5 Cognitive maps of killers’ movements between home bases, victims’ abductions and body dump sites. (A) Offender 1: 1 inch = 1 mile (2.54cm = 1.6 km); 130°. (B) Offender 2: 1 inch = 10 miles (2.54cm = 16km); 33°. (C) Offender 3: 1 inch = 20 miles (2.54cm = 32km); 246°. (D) Offender 4: 1 inch = 10 miles (2.54cm = 16 km); 162°. O, home base; solid circle, dump site; ■, abduction/last seen site; numbers represent order of crimes.Note: Difference in circle sizes is due to map magnification. ® 1998 Predator.
In another study on directional bias, it has been suggested that the sectoral mental map may not be anchored on downtown, based on regular daily activities, but may instead focus upon recreation areas beyond the city which are visited weekly. To be sure, there is ample basis for the argument that criminals’ spatial patterns depend on limited mental maps or mental images of the city and areas surrounding their homes. When they carry out their crimes, the loca-tional choices are affected by reliable knowledge of the city or area, which forms wedge-shaped images that are shapes in focus for places close to home or other parts of the home sector, and blurry or blank for distant places, such as the other side of town. This paper certainly supports the view that investigative efforts would be greatly enhanced if geographical profiles took into consideration the directionality of crime locations.

Conclusions

This article has discussed theories and concepts associated with the geographical behavior of criminals, with particular focus on four principal approaches: ecological school, environmental criminology, environmental psychology and the wedge theory. As previously mentioned, research into geographical profiling originated from the ecological approach and was concerned with the identification of crime areas in relation to offender demographics rather than the location of specific criminal events.
The environmental criminology approach provided a new perspective on the relationship between an offender’s home base area and the location of his or her offenses. Hence, it moved away from looking at the causes of crime and emphasized the importance of where victims are targeted and where crimes occur. Although the foundation on which environment criminology is built is entirely theoretical, it does suggests that criminals do have patterns in their spatial behavior, which could have implications for predicting future target areas. In a different vein, geographical profiling from the environmental psychological approach deals directly with the prediction of the likely behavior of the serial offender, as demonstrated in research on rape by Canter and his colleague. Environmental psychology holds more promise than previous nonpsychological-based geographical profiling methods.
The article has presented evidence suggesting that criminals have limited spatial knowledge, and that these patterns appear to be wedge-shaped. Such findings are in accord with a perspective that sees journeys to crime growing out of the daily activities and contact patterns of the offender. For example, when serial killers make environmental choices involving locational considerations, their mental maps are used as preferences as to which areas to forage for potential victims and dispose of their bodies. The fact that serial killers’ spatial behavior is sectorally biased suggests that police investigators are not dealing with economic killers, with perfect knowledge, but rather real individuals with imperfect knowledge and a predilection for satisfying behavior. Such a finding has direct implications for systems such as Rossmo’s criminal geographical targeting system, leading to the hypothesis that his procedure might be more efficient if it considered the directionality of crimes.
Admittedly, the interpretations of this article are ex post facto and one has to speculate if there are additional features of the criminal’s spatial behavior that can be predictive. Further research is needed to find out what additional psychological and environmental factors influence the geographical behavior of criminals.

Next post:

Previous post: