|
This paper outlines the complexities involved in examining the manifestation
of crime and disorder on public transport, particularly in applying the
traditional analysis techniques commonly used within the field of environmental
criminology. It begins by highlighting the importance of examining crime
on public transport and the limitations that exist in current data, and
suggests some reasons for the paucity of studies evident. Previous research
that attempts to explain crime on public transport with reference to the
characteristics of the area a route traverses is utilised, in order to
consider whether environmental criminology rationale can be applied to
examine crime on public transport.
The analytical techniques used within the field of environmental criminology
are then examined, and are found to focus primarily on examining crime
events that have a specific location (static events). However on the public
transport system crime events may happen on a moving vehicle, which are
difficult to locate, and hence analyse spatially. Thus the public transport
arena contains a mixture of static and non-static events. Although techniques
to examine static events are well established, this is not the case for
non-static events. Two techniques are suggested to address this problem.
Finally, some discussion is provided as to future directions for research,
including the need for empirical testing of some of the ideas presented
in this paper, and the development of techniques to analyse crimes that
do not have a single location.
Crime and Disorder on Public Transport
There has been limited attention afforded to incidents of crime and disorder
on public transport, particularly in comparison to the occurrence of such
incidents outside of the public transport arena. In the United Kingdom
(UK), perhaps the single most conclusive evidence of this is provided
by the fact that the police do not record incidents of crime on public
transport as a category in its own right, and there is no national or
local policing unit dedicated to bus services, although the British Transport
Police are responsible for policing rail services. In addition to this,
difficulties in obtaining accurate data on the location of public transport
crime (described later in this paper) have further restricted analysis
into the manifestation and distributions of such crime.
On public transport fear of crime and concerns for personal security,
a generic problem in many aspects of society today, are clearly a limiting
factor to patronage and levels of usage (Levine and Wachs 1986b; Benjamin
et al. 1994; and Ingalls, Hartgen and Owens 1994). The then Department
of the Environment, Transport and the Regions White Paper (DETR 1998),
suggests that patronage on public transport could be increased by 3 percent
at peak and 10 percent at off peak times if fear of crime were reduced.
A vicious circle exists here, as fuller trains and buses make people feel
safer when traveling. Furthermore, a recent survey of both users and non-users
of public transport in Merseyside (UK) indicates that, after service and
reliability, over 50 percent of respondents felt personal security should
be the priority area for improvement (Baker and Bewick 2001). In addition
to the safety of passengers, it is also important to consider the security
of staff. Budd (1999) examines risk of violence at work, using data from
the 1994, 1996 and 1998 British Crime Surveys (BCS). This report states
that public transport staff have a high risk of both 'threats' (5.6% compared
to an average risk of 1.5%) and 'assaults' (2.8% compared to an average
risk of 1.2%).
This has wide ranging implications, because, as highlighted in the DETR
report cited above, public transport is important as a means of gaining
access to health, leisure and other facilities, and thus in making a contribution
to minimise social exclusion. Furthermore, the environmental benefits
highlighted by promoting public transport as a means of sustainable mobility
are undermined by real fears of personal security on public transport.
Therefore, the reduction of crime and fear of crime on public transport
should be viewed as an area of paramount importance.
In order to reduce fear of crime, it is necessary to reassure the public
about the safety of using public transport. To achieve this requires reliable
information about the nature and extent of crime and disorder on public
transport, the "who, what, when, where, and why" questions of
crime. A better understanding of the prevalence of crime should then enable
appropriate measures to be implemented to prevent and reduce such crime.
Indeed, "the ultimate goal should be to make riders feel safe by
ensuring that they are safe" (Nelson 1997:7). Any measures introduced
to address problems of crime and disorder on public transport should be
based upon clear and appropriate analysis. This should include reliable
information on the levels of crime and disorder on public transport, and
on what measures applied where and when are likely to prove successful.
There are a number of problems that exist when attempting to gauge the
level and extent of crime and disorder on public transport, that make
it difficult to determine whether a gap exists between the 'perceived'
and 'actual' levels of crime. This is in part due to the lack of data
collected on actual levels of crime on public transport. The amount of
under-reporting of public transport crime is also unknown and may, as
an underestimate, be 25 to 30 times below the actual level of public transport
crime (Levine and Wachs 1986a). Major obstacles to collecting accurate
data on public transport crime, particularly in the UK, are that with
the exception of data collected by British Transport Police on rail crimes,
there is no dedicated unit responsible for policing buses or trams, and
there is no requirement to collect data on levels of public transport
crime. Although the 1988 Crime and Disorder Act in the UK placed a statutory
requirement on local authorities and police constabularies to produce
local crime and disorder audits and strategies on a three-year basis (Crime
Concern 2004), this does not require the incorporation of information
on levels of crime and disorder on public transport.
Indeed, in a review of the last round of the Crime and Disorder Audits
in the UK of 2002, only a limited number of these contain reference to
public transport (Crime Concern 2004). Only a quarter of these audits
made reference to crime on or near public transport facilities, less than
one tenth used data from transit companies (rail or bus operators) and
even fewer used data from bus companies. Some of the reasons why this
data was not utilised, or perhaps not available, are; that commercial
services only provide information on a voluntary basis; staff often will
not report incidents (for example, verbal abuse to ticket inspectors)
as it is perceived as part of their job, not worth reporting, and or treated
as not serious; and that incident reporting forms are not simple to fill
in (Crime Concern 2000).
From this it is evident that the prevalence of public transport crime
is perhaps an unknown, and that it is important for public transport organisations
to address the deficiencies that exist in their crime data, before attempting
to implement preventive measures. Without an accurate evidence-base upon
which to target resources, it is not only difficult to know where and
when to target scarce resources most effectively, but also perhaps impossible
to evaluate whether a preventative scheme has been successful or not.
THE ENVIRONS OF PUBLIC TRANSPORT CRIME
A number of studies have examined the manifestation of crime and disorder
on public transport. Whilst there is not scope here to provide a thorough
review of these, some of the key ideas that stem from previous research
will be drawn upon, in order to demonstrate the complexity involved in
examining crime and disorder on public transport, and the environment
within which it occurs. For a more general overview of studies the reader
is referred to Easteal and Wilson (1991); Clarke (1996); Felson et al.,
1996; Eck (1997); Loukaitou-Sideris (1999); Smith and Clarke (2000); DTLR
(2002); Newton (2004a); and Home Office (2004).
One of the early studies that attempted to explain the prevalence of crime
on public transport journeys with reference to the environment a vehicle
passes through was work by Pearlstein and Wachs (1982). They examined
crime on buses in California, and, for a one-year period, found that only
88 out of 233 routes experienced any serious incidents of crime, and that
crime mostly occurred on routes that traversed areas with high crime rates
in general. Their research also found that most crime occurred when passenger
volumes were highest, that crime was disproportionately high during the
late evenings when violent crimes were prominent, and that theft and robberies
were most prolific during the rush hour periods. From this they argue
that crime on buses is concentrated both in time and space. A subsequent
paper by Levine, Wachs and Shirazi (1986) using survey and observational
data at bus stops in Los Angeles, provides further support for the hypothesis
that incidents of bus crime tend to be highest on bus routes that travel
through high crime areas.
The Pearlstein and Wachs study emphasises that a bus route passes through
a mix of complex and heterogeneous environments, and that consequently,
in order to meet the problems posed by these differing environments, a
range of strategies may be necessary to tackle problems of crime and disorder
on these routes. However, this is not unique to the particular mode of
transport (buses) that they discuss. On the rail network for example,
there have been studies that demonstrate how low crime rates in certain
systems can be explained by some aspect of the design of their environment
(Gaylord and Galliher 1991; Myhre and Rosso 1996; and La Vigne 1997).
Easteal and Wilson (1991) emphasise that each mode of public transport
(they discuss bus, train, taxi, and aircraft, although tram, ferry and
other forms of transport could also be included) exhibits its own set
of unique problems. They argue that each mode of transport has a distinct
set of problems due to its unique environment, and hence efforts to reduce
each type of crime on each system may require a discrete set of solutions.
What is important to emphasise here is that each mode of transport has
its own unique set of environments that are distinct from other modes
of transport.
These two ideas can be combined. Pearlstein and Wachs suggest that the
environment that a bus route passes through is a mix of complex and dissimilar
environments (this almost certainly applies to other forms of transport).
Easteal and Wilson advocate that each mode of transport (bus, train, tram)
will traverse its own set of unique environments that are distinct from
those of other forms of transport. Merging these two ideas suggests that,
within each separate set of unique (dissimilar) environments each particular
mode of transport will traverse, there will also be a unique set of environments
associated with the particular route traveled by individual vehicles.
This is depicted in Figure 1, whereby area A refers to the bus environment
and area B represents the rail environment. The external environments
(the physical and social characteristics and crime levels) that influence
these distinctive modes of transport will will be different. For example,
the environmental influences on passengers 1 to 4 (bus journey) will be
different to riders 5 to 8 (rail system). In addition to this, there may
be areas with similar environmental characteristics where both rail and
bus systems coincide, and this is depicted in area C. Further to this,
only two modes of public transport are considered here, but, other forms
of transport such as trams would add further dimensions to Figure 1.
Figure 1. The Environs of Public Transport Crime.
 
In addition to the different external environments traversed by different
modes of transport (bus versus train) like modes of transport (for example
two buses or two trains) may also pass through different environments.
This is illustrated in Figure 1, whereby the solid lines represent how
a bus may move through these environments and the dotted lines represent
the train journey. At point 1, the user is traveling on a bus in a high
crime area, and at point 2 riding a bus in a low crime area. At point
7 the passenger is traveling on a train in a high crime area, and at point
8 the rider is on board a train in a low crime area. All four situations
may have a unique environment, and the experience of the passenger may
also differ dependent upon whether the bus or train stops or does not
stop within these low and high crime areas. This may influence the amount
of crime experienced on the route, by transferring offenders and potential
targets between these low and high crime rate areas and environments.
In addition to these external influences on a particular mode of transport
along the duration of its route, the internal environment of each mode
of transport is likely to have a bearing upon the levels of crime experienced
by public transport users. A bus or train will have its own unique internal
environment when a person is inside a bus or train carriage. The importance
of this internal environment was suggested by Mayhew et al. (1976) who
examined the effect of supervision on damage to buses in Manchester (in
the North West of England). They found that damage was greatest on buses
without a conductor, and more prevalent on upper decks, especially the
rear seats. They also discovered that on buses with a rear staircase,
graffiti and vandalism was more prominent upstairs at the front of the
bus. After adjusting the figures to account for where people are likely
to sit, they concluded that lack of supervision was an important factor
in the occurrence of vandalism and graffiti on buses.
In addition to the influence of the changing external environment a public
transport vehicle will traverse, and the internal environment of that
vehicle itself, the actual infrastructure of the public transport system
is also likely to relate to the prevalence of crime. The interaction between
these internal and external environments occurs at stops, stations, underground
stations, and interchanges, and these have an important role in that they
provide the only inputs and outputs on the public transport system.
This is also depicted in Figure 1. At point 5, the user is waiting for
a train at an underground station in a high crime area, and at point 6
waiting for a train (above ground) in a high crime area. At point 3 the
user is waiting at a bus stop in a low crime area, and at point 4 the
passenger is waiting for a bus in a high crime area. The importance of
this is that the environments at all these points are very different,
and their impact upon crime rates is also likely to vary considerably.
For example, the high crime rates and environmental characteristics above
ground (point 5 above the underground station in Figure 1) are likely
to have less of an influence on the passenger in the subway station, than
the characteristics at point 6 (waiting at a rail station at street level
in a high crime area), and perhaps also at point 4 (waiting at a bus stop
in a high crime area). These ideas are now discussed further with reference
to previous research.
There have been some studies into crime and disorder near such public
transport facilities. Block and Davis (1996) examined the geographical
distribution of street crime in four districts of Chicago, to ascertain
whether the area adjacent to rapid transit stations is a focus for street
crime or not (as opposed to looking at crime within stations). They found
that in the low crime rate areas street robbery was concentrated near
(but not immediately outside) rapid transit stations. In the high crime
rate areas, although robbery was most prominent on main streets, over
the two-year period at least one robbery occurred on every block. They
also found that robbery varied temporally, concentrated late at night
(11.00 to 12.00 pm, with a peak time of 2.00 am).
Loukaitou-Sideris (1999) looks at the connection between criminal activity
at bus stops and environmental factors based on empirical observations,
mapping and survey research. Ten high crime bus stops were analysed along
with four low crime 'control' stops in Los Angeles. Across the whole system
incidents were rare (there were fewer than 5 crimes per 100,000 passengers).
They found that the ten high crime bus stops that they examined accounted
for 18 percent of the total crimes reported out of 19,650 stops. Although
passenger levels at these stops were high, other nearby high patronage
stops exhibited little or no crime.
By examining the physical and social context of the surveyed bus stops,
they found an abundance of 'negative' environmental factors and a general
lack of defensible space at the high crime stops, whereas the four comparative
low crime rate stops lacked negative environmental factors and offered
better surveillance opportunities. These negative factors (within 300
feet of a stop) included "liquor stores, bars, check cashing establishments,
seedy motels, pawn shops, vacant lots/buildings and adult book stores
and movie theatres" (Loukaitou-Sideris 1999: 06). The authors conclude
that their empirical research indicates that environmental attributes
and site conditions at bus stops do have an impact on crime levels. This
finding is supported by Newton (2004b) who examined criminal damage to
bus shelters in Merseyside (UK), and found damage was related in a systematic
and predictable way to known attributes of a shelter's location
Liggett et al. (2001) extended the work of Loukaitou-Sideris, to investigate
how environmental factors around a bus stop could be used as a predictive
rather than an indicative measure, to estimate the likely amount of crime
as a result of placing a bus stop at a particular location. Using a series
of regression models, they determined a number of environmental predictors
of bus stop crime, and found that the most important predictor was location.
Any examination of the public transport needs to consider the transport
journey from start point to destination point. The system should be examined
in its entirety, and, indeed the holistic approach to the public transport
journey advocated by the DETR, the 'whole journey approach' is needed
to tackle crime on public transport. A transport journey consists of a
number of discrete, inter-linked components, and passengers and staff
need to feel secure during all aspects of a journey. "The best priority
is a holistic treatment. If one link of the journey is wrong the whole
journey may be cancelled or replaced by a car trip" (DETR 1999: 109).
When examining the public transport system from this standpoint, it is
possible to distinguish (at least) three possible scenarios in which a
crime event can occur. These are:
i) Walking or transferring between stops on foot (departure point to
stop or station, between stops or stations, stop or station to destination
point).
ii) Waiting at boarding or embarking points (at a stop, station or interchange).
iii) On board a mode of public transport (bus, train, tram or other
mode).
The above discussion outlines how public transport systems contain a
number of settings, which include the mode of transport (bus, train, tram
for example) and the infrastructure (stops, stations and interchanges).
There are perhaps two distinct influential environments, the differing
external environments the public transport vehicle traverses, and the
internal environment of the vehicle itself. These external environments
will vary by crime levels, socio-economic characteristics, land use, demographics,
and the physical infrastructure (the layout of roads and buildings or
the amount of open space). The internal environment will vary by the design
of the vehicle itself. The link between these two environments is provided
at the stops, stations and interchanges, which provide the gateway between
the internal and external environments, or the entry onto and exit from
the system. These entrance and exit points will also vary by the way they
are designed, be it the layout of a large station, a single stop, or the
entrance to the vehicle itself. These exit and entry points provide the
inputs and outputs to the system. There are a number of potential victims
of crime on the system, including passengers, staff, and facilities. There
are also a number of entry and exit points onto the system for potential
offenders, and capable guardians. Thus, examining the nature of crime
and disorder on public transport becomes a highly complex process.
APPLYING ENVIRONMENTAL CRIMINOLOGY RATIONALE TO PUBLIC TRANSPORT
Environmental criminology theories (Bottoms and Wiles 1997; and Clarke
and Eck 2003) examine how the convergence of a number of factors, is more
likely to result in the occurrence of crime. These features include location,
environment, the potential opportunity to commit a crime, the absence
of capable guardianship, the presence of offenders and targets, and the
juxtaposition of all these elements in time and space. Three of the most
influential theories of environmental criminology are Routine Activities
Theory (Cohen and Felson 1979), the Rational Choice Perspective (Cornish
and Clarke 1986), and Crime Pattern Theory (Brantingham and Brantingham
1993).
Routine Activities Theory states that for a criminal event to occur there
must be a convergence in time and space of three factors. These are (a)
the presence of a motivated offender, (b) the absence of a capable guardian,
and (c) the presence of a suitable target (person or object). Whether
or not these elements converge or coincide is a product of the routine
activities (day-to-day movements) of potential targets and offenders.
Public transport journeys may encompass part of the routine activities
of offenders, suitable targets (staff, passengers and facilities), and
capable guardians (for example, police officers, security staff, CCTV
cameras, or members of the public). This is particularly true when considering
the whole journey approach to public transport, from destination point
to end point (door to door). It is possible that the availability or lack
of public transport may actually influence a person's routine activities.
The use of public transport may also be shaped by obligatory (that an
individual must undertake) and discretionary (that a person chooses to
undertake) routine activities (LeBeau 2002).
'Rational Choice Perspective' suggests that offenders will choose their
targets and achieve their goals in a manner that can be explained (Cornish
and Clarke 1986). This seeks to explain the way in which crimes are distributed
by weighing up the potential cost of a crime (chance of apprehension,
cost of journey) against its possible benefits (potential reward, ease
to commit). Crime is assumed to be purposive to the offender's needs,
and constrained by limits such as time and the availability of information
(Felson and Clarke 1998). The offender rationally chooses the situation
with the highest net outcome. There is no reason to suggest that an offender
would not make a rational choice about committing a crime because they
are within the public transport domain, and, indeed, the decisions and
choices they make are likely to be influenced by the system itself.
Crime Pattern Theory argues that 'crime is an event that occurs when an
individual with some criminal readiness level encounters a suitable target
in a situation sufficient to activate that readiness potential' (Brantingham
and Brantingham 1993: 266). This approach to understanding crime contends
that crimes are patterned, but these patterns are only discernible when
crimes are viewed as aetiologically complex, occurring within and as a
result of a multifaceted environment (Eck and Weisburd 1995). Crime is
best viewed as an action that occurs within a situation at a site on a
changeable backcloth. This environmental backcloth includes social, cultural,
legal, temporal, spatial, and physical infrastructure characteristics
(Brantingham and Brantingham 1993). When broken down the model described
above is complex because the backcloth, site, situation, an individual's
criminal readiness and the distribution of targets are all required to
be examined coincidentally with each other in order to explain individual
crime events.
The three principal components of Crime Pattern Theory are nodes, paths
and edges (Brantingham and Brantingham 1981) and these appear to be particularly
transferable to the public transport arena (Felson et al. 1996). The idea
of personal activity nodes closely resembles routine activities, and refers
to a number of behaviour settings (slices of time and place where certain
activity occurs) that alter with time. These nodes are linked by paths,
which represent journeys between different activity nodes. Edges define
the boundaries around nodes and paths. Certain crimes may occur at these
edges, where people who are not familiar with each other meet (for example,
racist attacks and robberies). Public transport journeys may represent
such paths, and facilitate the movement of persons between some of the
activity nodes. These paths on the public transport system are separated
by edges, defining by the outer extents of the system, and regulated by
the various inputs and outputs to the system (stops, stations and interchanges).
A final concept that has been previously applied to public transport systems
is the idea of crime generators and crime attractors (Brantingham and
Brantingham 1995). The authors suggest that public transport stations
may be crime generators, crime attractors or fear generators. Transport
stations contain a number of people congregated together and this may
produce the opportunity for a crime to occur (a crime generator). At certain
times of the day these crowds and the characteristics they exhibit (for
example, commuters during rush hour) may produce suitable conditions for
a particular type of crime (for example, attract offenders who believe
there is opportunity to pick-pocket), and hence stations may act as a
crime attractor. Fear of crime can be generated in number of ways, especially
if the environment appears unclean, uncared for, not well lit or poorly
supervised (the Broken Windows Theory, after Wilson and Kelling 1982).
This discussion suggests that all these theories could be used to explain
crime on public transport. From this a number of potential directions
for future research can be identified. One avenue for exploration is to
examine whether public transport systems act as crime generators or crime
attractors, or both (which Brantingham and Brantingham (1995), begin to
explore). A second possible study is to investigate if certain paths on
the public transport network facilitate crime. Belanger (1997) starts
to investigate this, by examining how far offenders traveled from their
place of residence to the place where they committed subway crime. There
appears to be scope to utilise environmental criminology theories as an
explanatory focus for crime events on public transport, and the following
sections examine the methodological approaches necessary for this, to
ascertain the validity of such an approach.
STATIC AND NON-STATIC CRIME EVENTS
Earlier in this paper three situations were identified where a crime could
occur, when using the holistic approach to public transport journeys.
These include walking to, from and between stops, waiting at a stop, and
travelling on a moving vehicle. From an environmental perspective, these
can be described respectively as the following three different (but interlinked)
situations:
i) The walking environment
ii) The waiting environment
iii) The en-route environment
When considering a crime event in simple terms, as being something that
happens (Eck and Weisburd 1995), it may be argued that the above circumstances
describe two types of crime events. These situations describe a mixture
of 'static' and 'non- static' events, in terms of where
and when the crime event actually occurred. The first two situations primarily
describe a static crime event. For the purposes of this paper static crime
events refer to a crime occurring at an exact place, that can theoretically
be pin pointed to a specific location (x,y co-ordinate). An example of
this would be assault at a bus stop or train station. The second possible
scenario, however, implies the crime to be moving and this can be described
as a non-static crime event. When a crime occurs on a moving mode of transport
(bus, tram, or train for example) it is difficult to pin point the exact
location where the crime event occurred, as the crime happened on a moving
vehicle.
Non-static crime events may have more than a single location, and have
a start point (the place the crime started) and an end point (when the
crime finished). These two locations may be different, even if the crime
event only lasted for a short duration (for example over a thirty second
time frame). An example of this may be an assault that occurs on a moving
vehicle.
It is acknowledged that in certain circumstances the distinction between
static and non-static crime events is less clear. It could be contended
that the walking environment implies movement and therefore should be
viewed as non-static. When a crime event occurs in the walking environment,
however, it is likely that the target is stationary at the time of the
crime event, or movement is over a very short distance, perhaps a few
feet, and this location can be recorded as static (x,y location). The
speed of travel here is an important factor, as over the same time period
that the pedestrian moves a few feet, a bus or train may move several
hundred metres.
Additionally, when a moving vehicle is stationary (perhaps at lights or
at a stop), it could be argued that this is static. Whether a crime here
is recorded as static or non-static would depend on a number of factors.
These include; where the crime happened (did the crime happen only when
the vehicle was stationary, or include some movement of the vehicle before
and or after the stop); the duration of the crime; the speed of movement;
the distance travelled; and whether the event can be recorded at an exact
location (x,y co-ordinate) or between two points and times.
Finally, a missile (any item that could be thrown at a vehicle including
rocks, stones, bricks, and eggs for example) projected at a vehicle implies
the object has been thrown from a static location, onto a vehicle that
is moving (static to non-static). This situation here is unique as it
represents one of the few interactions between the internal (inside a
vehicle) and external (outside of a vehicle) environments of the transport
system that does not necessarily occur at a station, stop or interchange.
At the point of impact, the crime event could be pinpointed as static.
It can be argued that any crime event could take place over a time period
and moving space, for example a person gets knocked down, dragged into
a car and driven away, or a shop is 'ram-raided', (when a car, usually
stolen, is driven through a shop front) property is stolen from the shop,
and driven away (usually in another car). However in these situations
the crime events can be split into three separate acts, each with three
separate locations, whereas an assault that occurs on a moving bus or
train is a single continuous act with a moving space and time. The difficulty
faced is that no single precise location or time can be provided for the
crime event (the assault).
These static and non-static ideas may apply not only to the crime event,
but also to its environmental backcloth. The movement of this backcloth
may influence the situational factors that converge in time and space,
and result in crime events (both static and non-static). How is the convergence
of these situational factors influenced by non-static situations? Here
the fundamental question arises: can the existing theories of environmental
criminology that are largely focussed on static events be applied or adapted
to explain crime and disorder on public transport? When examining this
further, questions that arise include; are the existing theories limited
to the extent that they can't be applied to public transport; can they
be adapted; or do new theories need to be developed to explain crime and
disorder on public transport?
THE ANALYSIS OF CRIME EVENTS
Environmental criminology studies have primarily considered 'static' crime
events. These events have two key attributes, a space or place, and a
time. Numerous examples exist of the analysis of crime events outside
the public transport arena, including studies into domestic and commercial
burglary, assault, theft, robbery, car crime, domestic violence, racial
harassment, criminal damage, arson, and juvenile disturbances (Clarke
1997; Goldsmith et al. 2000; Hirschfield and Bowers 2001; and Ratcliffe
2002). The common feature of all this research is that the crime event
can be located at an exact place, by a geographical co-ordinate (x,y),
at a point in time (t).
As an extension to this, research by Ratcliffe (2002) developed the idea
of aoristic crime analysis. This considers that a crime may occur at a
single place, but it is difficult to define the exact time of this crime
event. In this analysis burglary incidents are examined, which, by their
nature, happen without the presence of a person to report the time of
the incident. They can be captured between the time a person left a property,
and the time someone has discovered the incident. Thus, the crime occurs
at a single location (x,y), but occurs between two time points (t1 and
t2). These characteristics could also apply to any crime event that occurs
at a single point and have a start and end time, between times t1 and
t2, that differ significantly.
Furthermore, there has been research that examines an offenders journey
to crime (Wiles and Costello, 2000), from the point an offender traveled
(x1,y1) at time t1 to the point the offender committed the crime (x2,y2)
at time t2. When the actual crime is committed, the crime event itself
is at (x2,y2) at time t2. An alternative to this it to examine the relationship
between where a crime occurs and property is recovered, for example theft
of a vehicle. In this example the theft of the vehicle would be at point
and time (x1,y1,t1) and the recovery of the vehicle at point and time
(x2,y2,t2), but the actual crime event itself would be (x1,y1,t1).
Crimes do not occur randomly or uniformly over time or space, and the
purpose of examining the patterns and distributions of crime events and
the environment where they occur, is to seek to explain the patterned
non-uniformity or non-randomness that real crime events exhibit. The technological
developments in Geographical Information Systems (GIS) and the growth
of crime mapping and crime analysis (Getis et al., 2000; Hirschfield and
Bowers 2001) have led to the development of a number of tools for the
spatial, temporal, and spatio-temporal analysis of crime patterns. Some
of these techniques are highlighted in later sections of this paper.
Spatial analysis techniques require information on the location of the
crime event. The term spatial analysis covers a wide area, but can be
defined as "the assemblage of analytical techniques and models in
which a clear association is maintained between quantitative data and
the spatial co-ordinates which locate them" (Chorley 1972, after
Wise and Haining 1991, 3.24.3). A variety of clustering algorithms have
been used to examine the spatial distribution of crimes (Anselin et al.,
2000), including neighbourhood hierarchical ellipses, kernal density interpolation,
LISA (Local Indicators of Spatial Association), K-Means clustering, STAC
(spatial and temporal analysis of crime), Voronoi analysis, GAM (Geographical
Analysis Machine), CrimeStat, and SCAS (Spatial Crime Analysis System).
In addition to examining the location of crimes, it is useful to examine
the environmental characteristics of the area where crimes occur, to add
further explanations to the occurrence of crime. Such features include
land use, the physical infrastructure of the area (the physical layout
of buildings), socio-economic, and demographic information. Hillier and
Shu (2000) discuss how the layout of urban space may influence crime levels.
In order to explore this further, micro level data is required, at a fine
scale (individual level) on both the exact location of crime, and its
environmental characteristics.
It is important to include data not only on the spatial location of a
crime, but also non-spatial information that can provide valuable insights
into the occurrence of a crime. An example of this is the concentrations
of crime evident on public transport, such as Pealstein and Wachs' findings
(1982) that only 18 out of 233 bus routes had a serious crime incident,
or the results of Loukaitou-Sideris research (1999) that found 10 high
bus stops (out of almost 20,000 stops) accounted for 18% of crime incidents
at bus stops. Combining this with the temporal concentrations of crime
evident in these studies and the findings of Levine et al. (1986) should
enable highly effective targeting of resources. Furthermore, this generates
questions such as why is crime clustered at these routes and stops, and
why do other routes and stops experience lower levels of crime?
It is important to include criminological theory when performing any crime
analysis or crime mapping, as the spatial element of a crime on its own
has a limited usefulness for future crime prevention. Pease (2001) likens
this to knowledge of a footballers position on a pitch, its meaning and
usefulness is informative only when we have knowledge of the laws and
tactics of the game. It is essential to incorporate environmental criminology
theory within any spatial, temporal or other quantitative analysis of
crime patterns. The following section explores whether the traditional
analysis methods embodied within current environmental criminology theories
can be applied or adapted to analyse crime events on public transport.
Methods for Analysing Crime Events on Public Transport
Crime events on public transport, as described earlier, may occur within
the waiting, walking, and en-route environments of the whole journey.
Thus, these 'static' and 'non-static' crime events can be translated into
three types of situations using the various crime analysis techniques
described previously. Crimes may occur in the following situations:
i. At x1,y1,t1 (for example an assault at a bus stop)
ii. Between x1,y1,t1 and x1,y1,t2 (for example criminal damage to a
bus shelter)
iii. Between x1,y1,t1 and x2,y2,t2 (for example assault on a moving
vehicle)
The first two situations above describe a 'static' crime event, and the
latter a 'non-static' crime event. In situation (i) for example, the place
where the crime event occurred (x1,y1) and the time of the event (t1)
are both known. In situation (ii) the location of the incident is also
known (x1,y1) but the precise time it occurred is not known, only that
it happened sometime between time t1 and t2. As the crime event happens
at a unique location (x1,y1) it can be termed a static crime event.
In situation (iii) the crime event has a starting point and time, and
end point and time that are different. There are a number of considerations
here that could be used to capture information about this non-static crime,
and these are depicted in Figure 2.
- The departure and termination points of the vehicle (points A and
B).
- The environments through which the vehicle has traversed before
the crime event occurred.
- The points where the vehicle stops along the route.
- The point the offender boards/alights the vehicle (points C and
D).
- The point the victim (if a person) boards/alights the vehicle (points
E and F).
- The point where the crime begins (trigger point or start point G)
and the point the crime ends (point H).
Figure 2. Non-Static Crime Events.
For a crime that occurs between two points, a question arises as to
whether to capture the two points that demarcate the location of the crime
event, or further detail such as the start and end points of the journey,
and where the offender and victim boarded the vehicle. During its journey,
the external environment that a vehicle traverses will vary. The characteristics
of the areas surrounding stops will influence who is boarding and alighting
the vehicle, and this influences the on board environment in terms of
who is on the vehicle (although the actual design of the on board environment
does not change). The demarcation of the crime event could be between
two points where the crime occurred (G and H on Figure 2 respectively),
or between two stops (C and D on Figure 2 respectively). The crime event
may also span several stops if a moving vehicle does not stop at certain
stops whilst a crime is happening. In the example of assault on a moving
vehicle, the start point would be where the vehicle was when the assault
commenced (point G), but this does not distinguish where the offender
(point C) or the victim (point E) actually entered the vehicle, nor the
last point they could have boarded the vehicle (point E), nor why they
first committed the crime where they did (trigger point G on Figure 2).
For some situations the crime might occur when the vehicle is stationary.
Here it might be possible to consider the crime as a static crime event
(point I in Figure 3), for example if the crime event is a single incident
without a start and end point, or if the vehicle is stationary at the
time of the incident.
Figure 3. Static Crime Events on Moving Vehicles.
Another potential scenario is that an object is thrown or missile is projected
(from a static location) onto a travelling vehicle (a moving entity).
Here, it is suggested it is more important to capture information about
where the missile is thrown from (point J on Figure 3) and the position
of the vehicle upon impact (point K on Figure 3), as the route the vehicle
has traversed previously is unlikely to have any influence on the position
the missile was thrown from. In this situation the trigger point would
be where the missile was thrown from, but if the vehicle did not pass
this point, or no missile were available, the crime would not have occurred.
This is unique in that it represents one of the few situations where the
internal and external environments converge outside a stop or station.
There are added difficulties in locating public transport crime. For some
crimes (for example graffiti or damage to a vehicle) the incident may
not be discovered until the end of the journey, or when the driver returns
the vehicle to the depot, and thus the crime could have happened any time
between when the vehicle was last checked and the time the damage was
discovered, along the route (or routes) it has traveled through between
these times.
This mixture of static and non-static crime events in the public transport
environment, presents a situation that is perhaps unique for the analysis
of crime events. The question posed here is how to apply the traditional
methods of crime analysis to both static and non-static crime events.
Analysing Static Crime Events
Earlier in this paper, a number of clustering algorithms and spatial analysis
techniques used to examine patterns of crime were described, and these
have been applied in a number of studies outside the public transport
arena. These methods can also be readily applied to examine the spatial
patterns of static crime events on public transport. This is because these
static crimes on public transport have a spatial location, or geographic
coordinate (x,y), of where the crime event occurred. An example of the
spatial analysis of static public transport crime events is research into
bus shelter damage (Newton 2004b).
One of the more common approaches used to examine static events is hot
spot analysis. A hot spot can be described as a "geographical area
of higher than average crime and or disorder. It is an area of crime or
disorder concentration relative to the distribution of crime and disorder
across the whole region of interest (e.g. ward, district, or county).
Hotspots are areas of clusters of crime or disorder that can exist at
different scales" (Chainey 2002). Whilst it is acknowledged that
hot spots in an area can vary by the time of day, for example hot spots
in an area at 12.00 midday may be very different to the hot spots in the
same area at 9.00 pm, these incidents here can still be considered 'static'
crime events, examined at two different times of the day.
The spatial analysis of crime data (Anselin et al., 2000) uses either
information on the unique location of individual crimes (disaggregate
data with an x,y point ), or aggregated data (that combines individual
point data into larger areal units). These two techniques have been combined
and displayed on a single map in Figure 4, which examines incidents of
criminal damage to bus shelters in Merseyside (UK) over a one-year period.
The analysis of individual (disaggregate) point data is demonstrated through
the use of the red circles in Figure 4. These circles represent the top
10% of individually damaged shelters in Merseyside over the one-year period.
These individual shelters could also have been analysed statistically
using the various clustering algorithms described earlier such as kernel
density interpolation or Nearest Neighbour Hierarchical (NNH) analysis,
to identify the hot spots of shelter damage. The advantage of this type
of analysis is that crimes are not aggregated into pre-defined areas,
thus patterns identified may be more tangible to the real world, since
offenders committing crimes are not constrained by administrative or other
boundaries (Hirschfield and Bowers 2001).
In Figure 4 the wards from the 1991 Census of Population are also portrayed.
An example of using aggregated data to analyse the patterns of crime is
the use of the light and dark shaded wards. In each ward the number of
times an individual bus shelter is damaged can be counted (using disaggregate
data). This information can be merged and aggregated for each ward. The
wards with the highest (dark shading) and lowest (light shading) 10 percent
of incidents of bus shelter damage in Merseyside are highlighted in Figure
4. It is noticeable that a preventive measure aimed at reducing crime
at the top 10 percent of individual shelters that were damaged would focus
on different shelters to a reduction measure aimed at the tackling the
10 percent of wards that experienced the most shelter damage.
Figure 4. Bus Shelter Damage on Merseyside, North West
England in 2002.
These individual incidents of shelter damage could have been aggregated
into a number of other areal units, such as census areas, police beats,
social service areas, or other administrative boundaries. The user may
also create these areas, around housing estates or to map socially perceived
neighbourhoods or communities for example. A number of spatial autocorrelation
techniques exist for the statistical analysis of aggregated data (see
Anselin et al. 2000). The advantages of using aggregated information is
that data sets with coterminous boundaries can be cross-referenced, for
example comparing crime levels aggregated to census wards with the socio-economic
characteristics of those wards. The disadvantages of this are that within
these areas there may be localised pockets of high or low crime areas
that are not apparent at the aggregated level (Hirschfield and Bowers
2001). Furthermore, such analyses are prone to errors that arise due to
the Modifiable Areal Unit Problem (Openshaw and Taylor 1981) and the Ecological
Fallacy (Brown 1991).
Analysing Non-Static Crime Events
The traditional spatial analysis techniques described above cannot readily
be applied to 'non-static' crime events, due to the difficulties in locating
a moving crime event. It is possible that this requires alternative techniques
to be developed. However, it is contended here that non-static crime events
do contain information on the location of the crime incidents. It is possible
to demarcate a crime between two points (x1,y1 and x2,y2) and two times
(t1,t2) as a single linear event at a single snapshot of time. Thus, instead
of applying analysis techniques to points or areas (as with the traditional
approaches), spatial analysis methods can be performed on these linear
routes or lines. Thus, for the purposes of analysis these non-static crime
events can also be treated as static.
Two possible methods to analyse these linear crime events are depicted
in Figures 5 and 6. It is likely that alternative and perhaps better spatial
analysis techniques can be developed around the idea of a linear crime
event, but as a starting point this discussion focuses on how the public
transport route can be delineated into smaller components or sections,
to examine the differences between sections of a route with low, medium
and high levels of crime.
In Figure 5 the route has been subdivided into a number of sections, or
areas. These could be segmented by administrative boundaries such as census
output areas, or be created by the user. For example stops and stations
along the route could be used to delineate route segments. These stops
would demarcate the last point an offender or victim could have boarded
a vehicle before the crime occurred, although the actual boarding points
may have been many stops earlier. However, this would enable crime events
on public transport to be examined for a number of route segments (between
stops and stations), and, for each segment, to be cross-referenced with
the social and physical characteristics of the surrounding areas or environments.
It would also enable a profile to be developed of the area a route has
traversed, before a crime event occurs.
Figure 5. Analysis of Non-Static Crimes by Route Section.
An alternative technique is to perform buffer analysis, around either
linear crime events (Figure 6), or to create buffers around segments of
routes. This can be achieved by creating concentric buffers around route
segments or linear crime events, at equal distances. For example a series
of concentric 50m buffer zones could be created around linear crime events,
the first 0-50m, the second 50-100m, the third 100-150m and so forth.
Within each of these buffer areas, characteristics of these areas could
be compared with the amount of crime occurring in that linear event.
From this the relationships between characteristics of surrounding environment
and levels of public transport crime could subsequently be examined. If
particular characteristics are found to influence crime levels on the
route, perhaps a relationship exists between areas of high crime in general
and transport route segments with high levels of crime for example, then
the influence of this by proximity to the route could also be examined.
Theoretically those buffer zones nearest to the route would have a greater
influence on the level of crime experienced than the zones further away
from the route. In addition to the influence of surrounding crime levels,
this technique could also be utilised to examine the relationship between
crime on the public transport route and other features of the physical
and social environment. It is acknowledged that there may be difficulties
in obtaining micro level data on the environmental characteristics of
these individual buffers zones, but the development of urban mapping systems
(such as OS MasterMap in the UK) with this type of disaggregate information,
should better facilitate such an approach.
Figure 6. Analysis on Non-Static Crimes by Buffer Analysis.
These two analysis techniques are illustrated here to demonstrate
the difficulties faced in analysing crime on public transport. It may
be necessary that alternative and more appropriate techniques need to
be developed to analyse linear patterns of crime. The difficulties in
collecting data on the location of a moving crime event, in addition to
the limited information collected about crime on public transport in general
(as described earlier in this paper), are perhaps the two primary reasons
for the paucity of studies evident in this area.
Accurately Reporting a Crimes Location on Public Transport
The analysis techniques described previously require the location of a
crime event on public transport to be demarcated at a precise location,
which may be at an individual location and time (x1,y1,t1) or between
two points and times (x1,y1,t1 and x2,y2,t2). This will allow a public
transport route to be divided into smaller subsections, to examine where
and when crime occurs along a route and the potential reasons for this.
This is depicted in Figure 7. However, careful attention should be afforded
to the methods used to report the location of crime events on public transport.
Figure 7. Capturing Static and Non-Static Crime Events.
The method by which a crime event is captured is important,
as it will not only heavily influence what subsequent analysis can be
performed, but can actually determine which analytical techniques can
be applied. A simple example of this is when a crime is located by the
general area where it occurs, for example a police beat or census area,
and not by its exact geographical position (x,y co-ordinate). This would
enable spatial analysis of crime patterns by area (for example by police
beat), but spatial analysis of individual crime points would not be possible.
The geographical position of crime can be reported in a number of ways,
and in the UK the Ordnance Survey's National Grid is commonly used to
produce precise x and y co-ordinates. However, inaccuracies in the location
of crime events evident in current crime data systems (Hirschfield and
Bowers 2001), focussed primarily on capturing static crime events, are
likely to be magnified when locating non-static crimes.
When a crime occurs on a moving vehicle, it is very difficult for a driver
or ticket officer to accurately report its location, and then for this
to be accurately transferred into a computer database. The use of a road
name, which may be several miles long, does not demarcate precisely where
crimes occur. On rail tracks it is often difficult to find a point of
reference to locate where a vehicle is. Road intersections and nearby
landmarks may help to locate crime incidents more precisely, but what
is desirable is a geographical co-ordinate, the vehicles route (by name
or number), and the direction of travel.
One potential tool to aid the accurate location of crime is the use of
Global Positioning Systems (GPS). Bus operators in the UK are developing
GPS tools to automate their revenue collection, and to develop electronic
timetable displays at stops that use real time. This time is based on
where the bus actually is in relation to the stop, and not when it is
likely to arrive based on timetable information. GPS devices may be handheld
or attached to a vehicle, but would need to be manually activated to indicate
when a crime starts, and perhaps also finishes. The police, fire and ambulance
service are at the forefront of developing more sophisticated methods
of accurately reporting where their incidents occur, particularly to ensure
a rapid response to emergency calls.
Newton and Hirschfield (2004) highlight the inconsistencies in current
methods used to capture crime data on public transport in the UK in their
examination of crime on buses in three case study areas. They found that
each study area recorded the location of crime in a different fashion.
In one area only the route number was used, in another area the crime
was located by the nearest bus stop (x,y position) although the actual
bus route was not recorded. In the final area, both the geographical position
of the actual crime (x,y) and bus route number were recorded. The analysis
that could be performed was limited by the way the data was recorded,
and for each of the case study areas it was necessary to employ a different
methodology. Interestingly however, these three different approaches all
yielded the same finding, that bus crime was positively correlated with
general crime levels in the surrounding area.
The limitation of recording only the bus route with no precise geographical
location of where the crime occurred is that, although high crime bus
routes were shown to pass through high crime areas, it was not possible
to test whether the risk of crimes on buses was greatest when the bus
was in high crime areas. The limitation of recording bus crime by nearest
stop and not by route number is twofold. Firstly the crimes location is
slightly distorted, and secondly although the location of high bus crime
incidents corresponded with high crime areas, there was no information
on where the bus had previously traversed. This is important as the characteristics
of the areas surrounding these previous stops, could influence who boards
and alights the bus (potential offenders and targets).
There are a number of ways by which crime events on public transport can
be located, but it is suggested that the following information is essential.
For static crimes:
(i) the point at which the crime occurs (for example x1,y1)
(ii) the route name or number (if applicable)
(iii) the direction of the vehicle (start and destination point, if
applicable).
For non-static crimes:
(i) the points and times the crime started (x1,y1,t1) and finished
(x2,y2,t2) and or
(i) the last stop before the crime started and the first stop after
it finished
(ii) the route name or number
(iii) the direction of travel (start and destination point)
This information will demarcate the actual crime offence in terms of
its location, but will not define where the offender or victim boarded
a vehicle. Perhaps the most appropriate mechanism for reporting the location
of non-static incidents warrants further research, particularly as this
is likely to impact upon the analysis that can be performed.
CONCLUSION
This paper has highlighted the complexity involved in examining crime
and disorder on public transport and the difficulties this poses when
attempting to analyse such crime events. The public transport environment
itself has been shown to be a multifaceted arena, with a number of settings,
and a range of potential offenders, victims and guardians. These settings
include different modes of vehicle (including buses, trains, trams and
other forms of transport), and facilities such as stations, stops and
interchanges. Moreover, the external and internal characteristics of these
environments need to be carefully considered. The vehicles will traverse
through a range of different external environments, and in addition to
this, will have their own internal environment unique to that vehicle.
This may vary between two different designs of buses, and between a bus
and a train for example. The design of the stops, stations, and interchanges
themselves may also influence crime on the system. These stops, stations,
and interchanges act as the gateway between the internal and external
environments, and control the input and output of potential victims, offenders,
and guardians onto the system.
When the holistic approach to public transport is considered, which is
necessary due to the interlinked nature of public transport, three scenarios
exist whereby a crime event may occur, the waiting, the walking, and the
en-route environments. These can be considered as static and non-static
crime events, as the crime may occur; at time and place x1,y1,t1; at time
and place x1,y1 and between times t1 and t2; or between places and times
x1,y1,t1 and x2,y2,t2. Current theories of environmental criminology focus
on the first two situations (static crime events), but do not consider
the latter of these, the non-static crime events.
As a result of this, the techniques developed to analyse crimes have centred
upon analysing points (disaggregate data) and areas (aggregate data).
These traditional crime analysis methods can be applied to static crime
events on public transport. What has not been considered is how to analyse
the non-static crime events. It is contended, for the purposes of crime
analysis, it is possible to capture non-static crime events as static.
By representing the crime event in a linear format, the route a vehicle
travels between x1,y1 and x2,y2 and times t1 and t2, depicts a static
crime event (a line) at a single snapshot in time. Thus, it could be argued
that the terms static and non-static crime events are arbitrary definitions,
and, for the purposes of analysis, all public transport crime can be captured
as static. The location of these events might be a single point, a single
place or area, or a line between two points. It is suggested that the
traditional theories of environmental criminology are very applicable
to public transport systems, and the difficulties that are faced are more
analytical than theoretical.
The techniques available to analyse linear patterns of crime are perhaps
underdeveloped, and it is suggested efforts are needed to address this.
There are a number of potential benefits here, to investigate not only
crime on public transport routes, but also crime across corridors in general.
This may have particular relevance to crime pattern theory, and to the
idea of nodes and paths. The developments in the field of GIS, particularly
in network analysis, alongside the collection of more data on the location
of crime on public transport, would enable this to be explored further.
This may be important in furthering our understanding of how the public
transport system may act as a crime generator or crime attractor, and
the implications this has for crime prevention. On public transport, the
influences of the external environment a vehicle traverses, and the internal
environment of that route, need to be further explored to understand the
complexities of the public transport system. In addition to this, the
links between these two environments, the internal and external environments,
provided by stops, interchanges, and stations, is perhaps a key area for
future research.
This paper has highlighted that the external environment a public transport
vehicle traverses can influence the level of crime experienced. This was
shown to be influential on both the level of crime on buses, and at bus
stops. This may have implications for situational crime prevention measures,
as altering some aspect of this environment could potentially reduce crime
on the system. However, a greater understanding of the factors that influence
levels of crime on public transport is required, in order to select appropriate
reduction measures.
This paper has highlighted some of the difficulties faced when persuading
passengers and staff to report public transport crime, and in demonstrating
to operators the need to collect such information, which has contributed
to the limited availability of public transport crime data. It is suggested
that in addition to this, the difficulties in precisely locating the location
of crime on a moving vehicle, and the limited knowledge of how to analyse
such information, are some of the primary reasons behind the paucity of
research in this area.
The importance of accurately locating a crime event was highlighted, because
it has a direct influence on the techniques and methods that can be applied
to examine the prevalence and distribution of this crime. Indeed, as the
accuracy of data on the location of a crime event decreases, increased
limitations are placed in the choice of available analysis techniques,
and the potential for error in analysis also increases. The growth of
available and accurate data on crime on public transport currently underway
in the UK, combined with the increased awareness of some of the issues
discussed above, will favour more empirical testing of these ideas, and
the development of improved analysis techniques to examine crime on public
transport.
ENDNOTES
1. The author thanks Professor Alex Hirschfield of the Applied Criminology
Group, University of Huddersfield (UK), for his advice and guidance, and
Dr Kate Bowers of the Jill Dando Institute of Crime Science, University
College London (UK), for her ideas and observations. Also thanks to Professors
Pat and Paul Brantigham of the Simon Fraser University (Canada) for their
inputs and suggestions. The author would also like to express his gratitude
to the anonymous reviewers for their comments and recommendations.
REFERENCES
Anselin, L., J. Cohen, D. Cook, W. Gorr, and G. Tita, 2000, "Spatial
Analysis of Crime." In Criminal Justice 2000. Measurement and
Analysis of Crime and Justice, edited by J. Reno, D. Marcus, M. Leary,
and J. Samuels [Online]. Available:
http//www.ojp.usdoj.gov/nij/criminal_justice2000/vol4.html. Accessed 12/02/2002.
Baker, J, and I. Bewick, 2001. Personal Security on Public Transport.
Final Report Prepared for Merseytravel. Bristol, UK: Transport and
Travel Research Ltd.
Belanger, M., 1997. "Crime Trips in the NYC Subway." Transit
Policing: 7(2) 15-22.
Benjamin, J., D. Hartgen, T. Owens, and M. Hardiman. 1994. "Perception
and Incidence of Crime on Public Transit in Small Systems in the Southeast."
Transportation Research Record 1433:195-200.
Block, R. and S. Davis. 1996. "The Environs of a Rapid Transit Station:
A Focus for Street Crime or Just Another Risky Place". Pp. 237-257
in Preventing Mass Transit Crime. Crime Prevention Studies Vol. 6,
edited by R. Clarke. Monsey, NY: Willow Tree Press.
Bottoms, A. and P. Wiles. 1997. "Environmental Criminology."
Pp. 305-349 in The Oxford Handbook of Criminology, edited by M.
Maguire, R. Morgan, and R. Reiner. Oxford, UK: Clarendon Press.
Brantigham, P. and P. Brantingham. 1981. Environmental Criminology.
Beverly Hills, Ca: Sage Publications.
Brantigham, P. and P. Brantingham. 1993. "Environment, Routine and
Situation. Toward a Pattern Theory of Crime." Pp. 259-294 in Routine
Activity and Rational Choice: Advances in Criminological Theory Volume
5, edited by R. Clarke and M. Felson. New Jersey: Transaction Publishers.
Brantigham, P. and P. Brantingham. 1995. "Criminality of Place:
Crime Generators and Crime Attractors." European Journal on Criminal
Policy and Research: 3(3) 5-26.
Brown, P. 1991. "Exploring Geodemographics." Pp 221-258 in
Handling Geographic Information: Methodology and Potential Applications,
edited by I. Masser and M. Blakemore. London: Longman.
Budd, T. 1999. "Violence at Work: Findings from the British Crime
Survey." Home Office Occasional Paper. London: Home Office.
Chainey, S. 2002. Focus Areas and Hot Spots. [Online]. Available:
http://www.crimereduction.co.uk/toolkits/fa0102.htm. Accessed 21/09/2002.
Clarke, R. 1996. Preventing Mass Transit Crime. Crime Prevention Studies
Vol. 6. Monsey, NY: Willow Tree Press.
Clarke, R. 1997. Situational Crime Prevention: Successful Case Studies.
Second Edition. New York: Harrow and Heston.
Clarke, R. and J. Eck. 2003. Become a Problem-Solving Crime Analyst
in 55 Small Steps. Devon, UK: Willan Publishing.
Cohen, L. and M. Felson. 1979. Social Change and Crime Rate Trends: A
Routine Activity Approach. American Sociological Review 44: 588-608.
Cornish, D. and R. Clarke. 1986. The Reasoning Criminal. New York:
Springer-Verlag.
Crime Concern. 2000. Reporting and Recording Incidents: Bus Related
Incidents of Crime and Disorder. Unpublished Report. Bristol, UK:
Crime Concern.
Crime Concern. 2004. Understanding the Links Between Bus-Related Crime
and Other Crime. Analysis of Crime and Disorder Audits Strategies, Round
2 - 2002. London, UK: Department for Transport (DfT). Forthcoming.
DETR (Department of the Environment, Transport and the Regions). 1998.
A New Deal for Transport. Better for Everyone. London, UK: DETR.
DETR (Department of the Environment, Transport and the Regions). 1999.
Personal Security Issues in Pedestrian Journeys. London, UK: DETR.
DTLR (Department for Transport, Local Government and the Regions). 2002.
Get On Board. An Agenda for Improving Personal Security in Bus Travel.
London, UK: DTLR.
Easteal, P. and P. Wilson. 1991. Preventing Crime on Transport.
Canberra: Australian Institute of Criminology.
Eck, J. 1997. "Preventing Crime at Places". In Preventing
Crime: What Works, What Doesn't, What's Promising, edited by L. Sherman,
D. Gottfredson, D. MacKenzie, J. Eck, P. Reuter and S. Bushway. [Online].
Available: http://www.ncjrs.org/works. Accessed 12/07/2001.
Eck, J. and D. Weisburd. 1995. "Crime Places in Crime Theory".
Pp 1-33 in Crime and Place. Crime Prevention Studies Vol IV edited
by J. Eck and D. Weisburd. New York: Criminal Justice Press.
Felson, M. and R. Clarke. 1998. "Opportunity Makes the Thief."
Crime Detection and Prevention Series, Paper 98. London, UK: Home Office.
Felson, M., M. Belanger, G. Bichler, C. Bruzinski, G. Campbell, C. Fried,
K. Grofik, I. Mazur, A. O'Regan, P. Sweeney, A. Ullma and L. Willaims,
L. 1996. "Redesigning Hell: Preventing Crime and Disorder at the
Port Authority Bus Terminal." Pp 5-93 in Preventing Mass Transit
Crime: Crime Prevention Studies Vol. 6, edited by R. Clarke. Monsey,
NY: Willow Tree Press.
Gaylord, M.and J. Galliher, 1991. "Riding the Underground Dragon:
Crime Control and Public Order on Hong Kong's Mass Transit Railway."
British Journal of Criminology. 31 15-36.
Getis, A., P. Drummy, J. Gartin, W.Gorr, K. Harries, P. Rogerson, D.
Stoe and R. Wright, 2000. Geographical Information Science and Crime
Analysis. [Online]. Available: http://www.urisa.org/Journal/protect/Vol12No2/getis/abstract.htm.
Accessed 15/09/2003.
Goldsmith, V., P. McGuire, T. Mollenkopf, and T. Ross, 2000. Analysing
Crime Patterns: Frontiers of Practice. Thousand Oaks, CA: Sage Publications.
Hillier, B. and S. Schu, 2000. "Crime and The Urban Layout. The
Need for Evidence." Pp. 224-250 in Secure Foundations. Key Issues
in Crime Prevention, Crime Reduction and Community Safety edited by
S. Ballintyne, K. Pease and V. McLaren. London: Institute for Public Policy
Research.
Hirschfield, A. and K. Bowers, 2001. Mapping and Analysing Crime Data.
Lessons from Research and Practice. London: Taylor and Francis
Home Office, 2004. Crime Reduction Toolkits: Public Transport.
[Online].
Available: http://www.crimereduction.co.uk/toolkits/pt00.htm. Accessed
15/01/2004.
Ingalls, G., T. Hartgen and T. Owens, 1994. "Public Fear of Crime
and Its Role in Bus Transit Use." Transportation Research Record
1433: 201-211.
LaVigne, N., 1997. Visibility and Vigilance: Metro's Situational Approach
to Preventing Subway Crime. Washington DC: National Institute of Justice.
LeBeau, J., 2002. "The Impact of a Hurricane on Routine Activities
and on Calls for Police Service: Charlotte, North Carolina and Hurricane
Hugo." Crime Prevention and Safety, an International Journal.
4 (1): 53-64
Levine, N. and M. Wachs, 1986a. "Bus Crime In Los Angeles I - Measuring
the Impact." Transportation Research A, 20 (4): 273-284.
Levine, N. and M. Wachs, 1986b. "Bus Crime In Los Angeles II - Victims
and Public Impact." Transportation Research A, 20 (4): 285-293.
Levine, N., M. Wachs, and E. Shirazi, 1986. "Crime at Bus Stops.
A Study of Environmental Factors." Journal of Architectural and
Planning Research 3(4): 339-361.
Liggett, R., A. Loukaitou-Sideres, and H. Iseki, 2001. "Bus Stop-Environment
Connection. Do Characteristics of the Built Environment Correlate with
Bus Stop Crime?" Transportation Research Record 1760: 20-27.
Loukaitou-Sideris, A., 1999. "Hot Spots of Bus Stop Crime. The Importance
of Environmental Attributes." Journal of the American Planning
Association 65 (4): 395-411.
Mayhew, P., R. Clarke, A. Sturman and M. Hough, 1976. "Crime as
Opportunity." Home Office Research Studies Number 34. London,
UK: HMSO.
Myre, M., and F. Rosso, 1996. "Designing for Security in Meteor:
A Projected New Metro Line in Paris". Pp 199-216 in Preventing
Mass Transit Crime. Crime Prevention Studies Vol. 6, edited by R.
Clarke. Monsey, NY: Willow Tree Press.
Nelson, K., 1997. "Policing Mass Transit: Serving a Unique Community."
FBI Law Enforcement Bulletin, January 1997. Washington DC: FBI
Publications.
Newton, A., 2004a. Crime and Disorder on Buses: Towards an Evidence
Base for Effective Crime Prevention. Unpublished PhD thesis. University
of Liverpool, UK.
Newton, A. 2004b. "Routing out the Hot Spots: Towards using GIS
and Crime-Place Principles to Examine Criminal Damage to Bus Shelters.
In Spatial Data for Evidence Based Policy Making. New Data, New Methods.
Innovations in GIS 10, edited by S. Wise, and M. Kraglia. London,
UK: Taylor and Francis. Forthcoming
Newton, A. and A. Hirschfield, 2004. "Understanding the Link Between
Bus-Related Crime and Other Crimes. Report 3B." Crime Analysis
and Mapping; A Report on the Four Case Study Areas. London, UK: Department
for Transport (DfT). Forthcoming.
Openshaw, S. and P. Taylor, 1981. "The Modifiable Unit Areal Problem."
Pp 60-69 in Quantitative Geography: A British View, edited by N.
Wrigley and R. Bennett. London: Routledge.
Pearlstein, A., and M. Wachs, 1982. "Crime in Public Transit Systems:
An Environmental Design Perspective." Transportation, 11:
277-297.
Pease, K., 2001. "What to do about it? Lets turn off our minds and
GIS". Pp 225-236 in Mapping and Analysing Crime Data. Lessons
from Research and Practice, edited by A. Hirschfield and K. Bowers.
London: Taylor and Francis.
Ratcliffe, J., 2002. "Aoristic Signatures and the Temporal Analysis
of High Crime Patterns." Journal of Quantitative Criminology.
18 (1): 23-43.
Smith, M. and R. Clarke, 2000. "Crime and Public Transport".
Pp 169-233 in Crime and Justice. A Review of the Research. Vol 6,
edited by M. Tonry and N. Morris. Chicago: University of Chicago Press.
Wiles, P. and A. Costello, 2000. "The Road To Nowhere: The Evidence
for Travelling Criminals." Home Office Research Study No 207.
London, UK: Home Office.
Wilson, J and G. Kelling, 1982. "Broken Windows." The Atlantic
Monthly 249 (3): 29-38.
Wise, S. and R. Haining, 1991. "The Role of Spatial Analysis in
Geographical Information Systems." AGI Proceedings 1991. London,
UK: AGI.
|
