Mapping Crime: Understanding Hot Spots - REPORT

Page created by Andrea Powers
 
CONTINUE READING
Mapping Crime: Understanding Hot Spots - REPORT
U.S. Department of Justice
          Office of Justice Programs
          National Institute of Justice
AUG. 05

                                                   Special   REPORT

          Mapping Crime: Understanding Hot Spots

                                                             www.ojp.usdoj.gov/nij
Mapping Crime: Understanding Hot Spots - REPORT
U.S. Department of Justice
Office of Justice Programs
810 Seventh Street N.W.
Washington, DC 20531

Alberto R. Gonzales
Attorney General

Regina B. Schofield
Assistant Attorney General

Sarah V. Hart
Director, National Institute of Justice

This and other publications and products of the National Institute
of Justice can be found at:

National Institute of Justice
www.ojp.usdoj.gov/nij

Office of Justice Programs
Partnerships for Safer Communities
www.ojp.usdoj.gov
Mapping Crime: Understanding Hot Spots - REPORT
AUG. 05

          Mapping Crime:
          Understanding Hot Spots

          John E. Eck, Spencer Chainey, James G. Cameron, Michael Leitner,
          and Ronald E. Wilson

          NCJ 209393
Mapping Crime: Understanding Hot Spots - REPORT
Sarah V. Hart
Director

This document is not intended to create, does not create, and may not be relied upon to
create any rights, substantive or procedural, enforceable by law by any party in any matter
civil or criminal.
Findings and conclusions of the research reported here are those of the authors and do
not necessarily reflect the official position or policies of the U.S. Department of Justice. The
products, manufacturers, and organizations discussed in this document are presented for
informational purposes only and do not constitute product approval or endorsement by the
U.S. Department of Justice.
The National Institute of Justice is a component of the Office of Justice Programs, which also
includes the Bureau of Justice Assistance, the Bureau of Justice Statistics, the Office of
Juvenile Justice and Delinquency Prevention, and the Office for Victims of Crime.

Cover Credits
(From left to right)
Map by Jeff Stith and the Wilson, North Carolina, Police Department
Map by Wilpen Gore, Heinz School of Public Policy and Management, Carnegie Mellon
University, Pittsburgh, Pennsylvania
Background map by Jeff Stith and the Wilson, North Carolina, Police Department
Mapping Crime: Understanding Hot Spots - REPORT
About This Report
Much of crime mapping is devoted to
detecting high-crime-density areas known
                                               What did the researchers
as hot spots. Hot spot analysis helps          find?
police identify high-crime areas, types of     ■   Identifying hot spots requires multiple
crime being committed, and the best way            techniques; no single method is suffi­
to respond.                                        cient to analyze all types of crime.
This report discusses hot spot analysis        ■   Current mapping technologies have sig­
techniques and software and identifies             nificantly improved the ability of crime
when to use each one. The visual display           analysts and researchers to understand
of a crime pattern on a map should be              crime patterns and victimization.
consistent with the type of hot spot and
possible police action. For example, when      ■   Crime hot spot maps can most effective­
hot spots are at specific addresses, a dot         ly guide police action when production
map is more appropriate than an area               of the maps is guided by crime theories
map, which would be too imprecise.                 (place, victim, street, or neighborhood).

In this report, chapters progress in sophis­
tication. Chapter 1 is for novices to crime
mapping. Chapter 2 is more advanced,
                                               Who should read this study?
and chapter 3 is for highly experienced        Crime analysts and researchers in police
analysts. The report can be used as a com­     departments.
panion to another crime mapping report
published by the National Institute of
Justice in 1999, Crime Mapping: Principle
and Practice, by Keith Harries.

                                                                                               iii
Mapping Crime: Understanding Hot Spots - REPORT
Contents
About This Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

Chapter 1. Crime Hot Spots: What They Are,
Why We Have Them, and How to Map Them . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Chapter 2. Methods and Techniques for
Understanding Crime Hot Spots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Chapter 3. Spatial Analysis Tools for Identifying Hot Spots . . . . . . . . . . . . . . . 35

Chapter 4. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

                                                                                                                                          v
Mapping Crime: Understanding Hot Spots - REPORT
Chapter 1. Crime Hot Spots: What They Are,
Why We Have Them, and How to Map Them
John E. Eck, University of Cincinnati

Crime is not spread evenly across maps.         scarce. Community policing is particularly
It clumps in some areas and is absent in        attentive to high-crime neighborhoods,
others. People use this knowledge in their      where residents have great difficulty exert­
daily activities. They avoid some places        ing social controls. Problem-oriented polic­
and seek out others. Their choices of           ing pushes police officials to identify
neighborhoods, schools, stores, streets,        concentrations of crime or criminal activity,
and recreation are governed partially by        determine what causes these concentra­
the understanding that their chances of         tions, and then implement responses to
being a victim are greater in some of           reduce these concentrations. Much of
these places than in others. In some            what is called crime analysis is dedicated
places people lock their cars and secure        to locating concentrations of crime—hot
belongings. In other places they do not.        spots—and much of crime mapping is
Along some streets people walk swiftly          devoted to their detection.
and view approaching strangers with sus­
picion. Along other streets they casually       This chapter discusses how different inter­
stroll and welcome the next interesting         pretations of hot spots require different
person they might meet, and notice others       types of crime maps. The principal theme
making the same choices in the same             is that crime hot spot maps can most
areas.                                          effectively guide police action when pro­
                                                duction of these maps is guided by theory.
Some might argue that this behavior             With the appropriate crime theory, crime
merely shows that people are unreason­          maps can communicate vital information
ably fearful of some areas but not of oth­      to police officials and community mem­
ers. This may often be true, but the fact       bers efficiently and effectively.
that people are not equally fearful of all
places suggests that they understand that       Many useful crime theories provide guid­
crime is not evenly distributed. People         ance for selecting mapping symbols.
might be mistaken about the risks of            Which theory is most useful depends on
some places, but they are not mistaken          the type of problem being mapped. Maps
that their risk of being a victim of crime is   that are not based on theory will provide
not geographically constant.                    officers with inadequate and even mislead­
                                                ing information.
Police use this understanding every day.
Decisions about how to allocate scarce          The term hot spot has a number of mean­
resources are based partially on where the      ings. This chapter begins with a discussion
demands for police are highest and where        of what the term means and how the
they are lowest. Officers are told to be        meanings relate to the concept of levels of
particularly attentive to some behavior in      spatial analysis of crime. Different theories
some areas, but are given no guidance           of crime explain crime at different levels,
about other areas where this behavior is        so this chapter briefly describes various

                                                                                                1
Mapping Crime: Understanding Hot Spots - REPORT
SPECIAL REPORT / AUG. 05

    levels of crime theories and explains how       crime or disorder. It also suggests that
    they can be depicted on maps. This chap­        some hot spots may be hotter than others;
    ter examines four types of crime theories       that is, they vary in how far above average
    in greater detail: place (point) theories;      they are.
    street (line) theories; area (polygon) theo­
    ries; and repeat victim theories, which can
    operate on point, line, or polygon level.       Levels of hot spot analysis
    These theories describe the levels of hot
    spots and how these levels can be depict­       If hot spots are merely areas with an
    ed on maps. This chapter examines why           above average amount of crime or disor­
    crime theory, crime mapping, and police         der, why do practitioners and researchers
    actions need to be consistent. The end of       use the term in such a variety of ways? In
    the chapter examines how the map sym­           fact, with recent developments in crime
    bols implied by each theory communicate         mapping, one can find hot spots of any
    to users of crime maps.                         size—from hot spot places to hot regions.
                                                    Although all of these perspectives on hot
                                                    spots have something in common—con-
                                                    centrations of crime or disorder separated
    What is a hot spot?                             by areas with far less crime or disorder—
    Areas of concentrated crime are often           they differ in the area covered by the hot
    referred to as hot spots. Researchers and       spots. More importantly, the factors that
    police use the term in many different ways.     give rise to hot spot places are different
    Some refer to hot spot addresses (Eck and       from the factors that give rise to hot spot
    Weisburd, 1995; Sherman, Gartin, and            streets, hot spot neighborhoods, or hot
    Buerger, 1989), others refer to hot spot        spot cities. Further, the actions one takes
    blocks (Taylor, Gottfredson, and Brower,        to deal with a hot spot place will be differ­
    1984; Weisburd and Green, 1994), and oth­       ent from the actions needed to address a
    ers examine clusters of blocks (Block and       hot spot street, hot spot neighborhood, or
    Block, 1995). Like researchers, crime ana­      hot spot city.
    lysts look for concentrations of individual
    events that might indicate a series of relat­   These approaches differ on the level of
    ed crimes. They also look at small areas        analysis, or the size of the geographic area
    that have a great deal of crime or disorder,    of crime about which one is concerned.2
    even though there may be no common              The level at which one examines crime or
    offender. Analysts also observe neighbor­       disorder is dictated by the question one
    hoods and neighborhood clusters with high       asks, which will determine the usefulness
    crime and disorder levels and try to link       of the results. Consider two related, but
    these to underlying social conditions.          very distinct, questions: Where are drugs
                                                    being sold? What is the market for drugs?
    Though no common definition of the term
    hot spot of crime1 exists, the common           The precise answer to the first question
    understanding is that a hot spot is an area     requires identifying specific drug-dealing
    that has a greater than average number of       locations or street segments (very small
    criminal or disorder events, or an area         areas) where drug dealers and customers
    where people have a higher than average         routinely meet. To answer the second
    risk of victimization. This suggests the        question, the analyst needs to find out
    existence of cool spots—places or areas         where the customers are coming from,
    with less than the average amount of            just as he would if he asked the question,

2
Mapping Crime: Understanding Hot Spots - REPORT
MAPPING CRIME: UNDERSTANDING HOT SPOTS

“What is the market for new cars?” The            Crime hot spot theories
answer to the first question—specific
locations or street segments—is not par­
ticularly useful for answering the second
                                                  Place theories
question. Rather, the analyst would be            Place theories explain why crime events
interested in larger areas with high con­         occur at specific locations. They deal with
centrations of drug users. These areas            crimes that occur at the lowest level of
might surround the locations and street           analysis—specific places. They involve
segments identified when answering the            looking at specific incidents and asking
first question, or they may be physically         such questions as, “At what places are
separated from the dealing sites (as would        burglaries occurring and at what places are
occur when suburban high school and col­          they not occurring?” Crime phenomena at
lege students drive into cities to find           this level occur as points, so the appropri­
drugs). The types of police actions that          ate units of analysis are addresses, street
might remove drug-dealing locations are           corners, and other very small places,
likely to be different from the actions           which are typically represented on maps
needed to dry up the market. So identify­         as dots. Police action, such as warrants,
ing the appropriate level of analysis is criti­   which specify exact addresses (not blocks
cal to understanding the problem and              or neighborhoods), is very precise at this
determining what action to take.                  level. Similarly, nuisance abatement focus­
                                                  es on specific locations.
Crime theories are critical for useful crime
mapping because they aid in the interpre­
tation of data (Eck, 1998) and provide            Street theories
guidance as to what actions are most              Street theories deal with crimes that occur
appropriate. Therefore, understanding             at a slightly higher level than specific
how crime theories account for hot spots          places; that is, over small, stretched areas
is critical. Several theories of crime and        such as streets or blocks. A prostitution
disorder concentration (hot spots) exist.         stroll is an example. At this level of analy­
Some theories disagree, but often the             sis analysts ask such questions as, “On
theories do not contradict each other.            which streets are prostitutes found and on
Rather, they explain different types of           which streets are they not found?” The
crime phenomena that occur at different           appropriate units of analysis can be street
geographic levels.                                segments, paths, and sections of high­
                                                  ways, which would be represented on
Each level has basic units of analysis—the        maps as straight, bent, or curved lines.
things being examined. One can think of           Police action is still relatively precise,
units as corresponding to the geographic          although not as precise as at the place
areas being depicted on maps: points,             level. Concentrated patrolling occurs at
lines, or polygons (Harries, 1999). Some          this level, for example, as well as efforts to
theories help explain point concentrations        change traffic and street patterns.
of crime. Other theories help explain linear
concentrations of crime or hot spot crime
polygons. However, theories of crime are          Neighborhood theories
useful for helping to guide crime and disor­      Some theories of crime attempt to explain
der mapping only if one selects a theory          neighborhood differences.3 At a higher
appropriate for the level of analysis and         level than place or street, neighborhood
action.

                                                                                                   3
Mapping Crime: Understanding Hot Spots - REPORT
SPECIAL REPORT / AUG. 05

    theories deal with large areas. Here ana­        Exhibit 1 organizes and summarizes the
    lysts are interested in such questions as,       discussion of hot spot analysis so far and
    “What areas are claimed by gangs and             introduces what is to come. The first col­
    what areas are not?” The appropriate units       umn describes the geographic concentra­
    of analysis are quite varied and can include     tion at various levels of interest. The
    square blocks, communities, and census           second column describes the basic pat­
    tracts, to name a few. Two-dimensional           tern formed by hot spots at each level. The
    shapes such as ellipses, rectangles, and         third column lists the geometric dimen­
    other polygons are used on maps to repre­        sion to be used on a crime map to depict
    sent crime phenomena at this level. At this      each type of hot spot. Place theories sug­
    level police action is far less precise          gest maps with dots, street theories sug­
    because the areas are typically too large        gest maps that emphasize lines, and area
    for effective concentrated patrolling            theories suggest the use of polygons on
    (Sherman, 1997). Nevertheless, depending         maps. Repeat victimization theories do not
    on neighborhood characteristics, relevant        directly correspond to a single dimension
    action might include efforts to engage resi­     or level. They can be depicted on maps by
    dents in collective action against crime         dots, lines, or polygons. The last three
    and disorder. If offenders are mobile            columns highlight points discussed next.
    throughout an area, rather than concen­          Examined are four types of hot spots—
    trated at a few places, then efforts to          places, victims, streets, and areas.
    deter them should occur at this level.

    Other large area theories                        Types of hot spots
    Still other theories attempt to explain dif­     Repeat places hot spots
    ferences in crime patterns at much higher
    levels of aggregation. For example, theo­        The most basic form of a hot spot is a
    ries of crime differ among cities and            place that has many crimes. A place can
    among regions. On the city level, suggest­       be an address, street corner, store, house,
    ed actions may include citywide changes          or any other small location, most of which
    in economic, transportation, education,          can be seen by a person standing at its
    welfare, and recreation policies, to name a      center (Sherman et al., 1989). Places typi­
    few. On the multijurisdictional or multi-        cally have a single owner and a specific
    state regional levels, suggested actions         function—residence, retail sales, recre­
    against concentrations of crime could            ation, school (Eck and Weisburd, 1995).
    include even broader scale policies or           Crime often is concentrated at a few
    social change. Although these are interest­      places, even in high-crime areas. Although
    ing theories, they are far less useful for       hot places often are concentrated within
    local police agencies. Thus, they are not        areas, they often are separated by other
    examined here.                                   places with few or no crimes. Because
                                                     such hot spots are best depicted by dots,
                                                     they have a dimension of zero.
    Repeat victimization theories
    Finally, repeat victimization theories pertain   Underlying causes. Routine activity theo­
    to questions of why the same victims are         ry helps to explain why crime often is con­
    targeted repeatedly. They can operate at         centrated at specific places. In particular,
    any of the three levels discussed: points,       routine activity points to how behavior is
    lines, or polygons. However, not all repeat      regulated at the location by place man-
    victimization can be shown on maps.              agers—owners of places or people acting
                                                     on an owner’s behalf. Behavior regulation

4
MAPPING CRIME: UNDERSTANDING HOT SPOTS

falls under place management theory, a                               place (Brantingham and Brantingham,
part of routine activity theory. For exam­                           1995). Repeat places tend to be stable
ple, the difference between a bar that has                           over time (Spelman, 1995a), and this is
few or no incidents or assaults and a bar                            consistent with the routine activity theory
with frequent assaults is likely to be that in                       that an absence of effective place man­
the first instance the bar employees regu­                           agement is at the heart of the problem.
late the behavior of patrons to minimize
the chances of an assault, and in the sec­                           Maps for repeat places. Maps for repeat
ond instance, they do not. Such regulation                           places include—
has three effects. It directly prevents crimi­
                                                                     ■   Graduated symbols. When looking for
nal activity through early intervention (e.g.,
controlling the number of drinks a patron                                hot places, dot maps are superior to
can consume), it attracts place users who                                other forms of mapping. The goal is to
desire a well-regulated location over a                                  identify isolated high-crime locations,
weakly regulated place (such people are                                  which can be done in a number of ways.
less likely to create problems), and it                                  One can use graduated dots, so that dot
repels place users who desire a weakly                                   size is proportional to the number of
regulated location over a well-regulated                                 crimes at the location. This method

Exhibit 1. Hot spot concentrations, evidence, theory, and causes

                           Map                    Geometric
Concentration             pattern                 dimension                  Theories         Likely causes         Examples

Place—at            Point concentration;       Zero; concentration         Routine activity   Management of         Bar fights,
specific            a few places with          at points                   theory; place      behavior at places    convenience
addresses,          many crimes and many                                   management                               store
corners, or other   places with few                                                                                 robberies,
places              or no crimes. Repeat                                                                            ATM patron
                    crime places are often                                                                          robberies,
                    concentrated.                                                                                   drug dealing
                                                                                                                    locations
Among               Often confused with        Zero, one, or two;          Routine activity   Victim routines       Domestic
victims             repeat crime places        concentration at            theory;            and lifestyle         violence
                    (above). Only visible      points, lines, and          lifestyles         choices
                    on maps if victims are     areas
                    concentrated at places,
                    on streets, or in areas.
Street—along        Linear concentration       One; concentration          Offender search    Offender movement     Outside
a street or block   along major thorough-      along lines                 theory             patterns and target   street
face                fares; a few blocks                                                       concentrations        prostitution,
                    with much crime and                                                                             street drug
                    many blocks with little                                                                         dealing,
                    crime                                                                                           robberies of
                                                                                                                    pedestrians
Area—neighbor-      Concentration covering     Two; concentration          Disorganization    Low collective        Residential
hood areas          multiblock areas           in areas                    theory and re-     efficacy, social      burglary,
                                                                           lated ecologic     fragmentation,        gang
                                                                           theories of        concentrations        violence
                                                                           crime;             of youth,
                                                                           opportunity        economic disinvest-
                                                                           theories           ments; concentra-
                                                                                              tions of crime
                                                                                              targets

                                                                                                                                    5
SPECIAL REPORT / AUG. 05

        allows the depiction of repeat and non-      Repeat victimization hot spots
        repeat places on the same map and per­
        mits comparison among repeat places          Repeat victimization refers to the multiple
        about the number of crimes. Graduated        attacks on the same individual, regardless
        dots also allow one to find concentra­       of location. It often is confused with
        tions of hot places (e.g., an area that      repeat crime places. A repeat place might
        contains several repeat assault bars).       have a number of different victims.
        Because graduated dots can obscure           Clearly one can have both repeat victim­
        nearby features (e.g., a large dot may       ization and repeat crime places (Eck,
        overlap nearby smaller dots), this tech­     2000). For example, a person could fre­
        nique is best used on large-scale maps.      quent a bar where he is assaulted on a
                                                     number of different occasions. But if
    ■   Color gradient dots. Two other               repeat victimization is distributed over
        approaches are useful on small-scale         many locations (as would occur if repeat
        maps. One is to use a color gradient—        victims are assaulted at different bars,
        yellow through red, for example—to           but never the same bar twice), it will not
        depict the number of crimes at each          show up on a map as a hot spot place
        location. A yellow dot may be used to        (zero dimension). Repeat victimization
        represent places with a single crime, a      could show up as lines (one dimension) if
        light orange dot may represent locations     the victims are repeatedly attacked along
        with two crimes, and a deeper orange         the same thoroughfares, or as a polygon
        dot might represent places with three        (two dimensions) if victims are repeatedly
        crimes. This approach has the advantage      attacked in the same neighborhoods.
        of the use of graduated symbols but
        overcomes the overlap problem.               Mapping repeat victimization is more likely
                                                     to reveal patterns with vulnerable popula-
    ■   Repeat addresses. Another method is          tions—potential victims who engage in
        to select the most serious hot spot          similar activities. Consider taxicab rob­
        addresses. For example, one might want       beries and homicides. These crimes are
        to find the worst 10 percent of the          unlikely to be concentrated in places. One
        addresses. This is called repeat address     might find attacks on this victim group
        mapping (RAM). The addresses would           occurring along specific streets where the
        be the 10 percent of repeat addresses        drivers are particularly vulnerable or where
        that have the most crimes. They would        offenders have a better chance of escape.
        be plotted on a map using dots to repre­     More likely, however, taxicab robberies
        sent hot spots. This method has two dis­     and homicides will be spread over a neigh­
        tinct advantages. First, the map is          borhood or in a multineighborhood area
        clearer because it has less clutter.         within a city.
        Second, such maps are useful for clearly
        specifying police targets. The deficiency    Underlying causes. Repeat crime places
        of RAM is that it leaves out information     with different victims and repeat victimiza­
        about the other locations. This deficiency   tion with different places have different
        can be overcome by producing supple­         causes. Repeat crime places (with differ­
        mentary maps that show all locations or      ent victims) can be attributable to the
        by combining RAM with the use of a           behavior of place managers, but if the vic­
        color gradient so that the targeted hot      timizations occur at different places, place
        spots have a distinct color (Eck, Gersh,     managers have less of a role. In those
        and Taylor, 2000).                           cases, one should look at the occupations,

6
MAPPING CRIME: UNDERSTANDING HOT SPOTS

commuting patterns, or lifestyles of the        (including offenders), targets situated
potential victims (Farrell and Pease, 1993;     along thoroughfares face higher crime
Spelman, 1995b; Stedman and Weisel,             risks than targets on side streets far from
1999). The most obvious example comes           thoroughfares. Further, some types of tar­
from the increasing evidence that the peo­      gets concentrate along major streets.
ple most likely to become victims of            Convenience stores, fast food stores, gas
assault are those people most likely to be      stations, and other retail places are sited
involved in deviant and criminal activity       along major thoroughfares because that is
(e.g., drug dealing, drug use, heavy alcohol    where their customers concentrate. So for
consumption, prostitution) (Menard,             both reasons of offender movement pat­
2000). Some occupations increase the            terns and target placement patterns, many
chances of victimization, which can             crime hot spots are actually hot lines.
increase repeat victimization. Police offi­
cers, for example, have a greater rate of       Some offenses may be concentrated at
victimization than many other occupations       points or along lines. Street drug dealing is
(Block, Felson, and Block, 1985). However,      one example. Many street drug dealers
the things that make a person a target for      simultaneously work along streets but
crime are sometimes difficult for that per­     anchor their activities to a specific
son to change.                                  address. In such circumstances, one
                                                might find a concentration of drug dealing
                                                along a few street segments and concen­
Repeat streets hot spots                        trations of drug locations at anchor sites.
Repeat streets are those thoroughfares or       Weisburd and Green (1995) used street
streets with a high degree of victimization.    segments to identify drug hot spots in
Repeat places and some repeat victimiza­        Jersey City because of offender move­
tion hot spots show up as dots on crime         ment patterns. Eck (1994), however, identi­
maps. If one increases the dimension of         fied drug-dealing places because they
the hot spot from zero to one, hot spots        seemed to be the anchor points of the
that form lines appear. Linear hot spots        drug trade in the San Diego neighborhood
are likely to be the results of the interac­    he was studying.
tion of targets and offenders along thor­
oughfares. Brantingham and Brantingham          Distinguishing hot places from hot streets
(1981) describe the search behavior of          can be difficult. In fact, one can some­
offenders. Their offender search theory         times find both. Imagine robbers attacking
points to the importance of street patterns     pedestrians on a street leading from
for how offenders look for targets.             restaurants and bars to a parking area. The
                                                attack sites may form a line along this
Underlying causes. Offenders find tar­          street. But even along this hot street, hot
gets while going about their normal legiti­     places where multiple attacks have
mate business—going to and from work,           occurred may exist. However, one should
recreation, shopping, school, and other         always be suspicious of such findings. It
nodes of activity. Potential targets that are   might be that the hot places are not actual
not along the routes or near nodes used         robbery occurrence sites. Instead, they
by offenders will unlikely be victimized, but   may be locations to which victims run for
those close to offenders’ routes and nodes      help, or they may be addresses that offi­
have elevated risks of victimization. Since     cers put in their reports when they cannot
major thoroughfares concentrate people          easily find the correct robbery address.

                                                                                                7
SPECIAL REPORT / AUG. 05

    Knowledge of offender, victim, and police            outmigration. These changes either dis­
    behavior is critical to separating the under­        rupt social networks or prevent such
    lying crime pattern from reporting and               networks from forming. Since these net­
    recording patterns.                                  works, according to disorganization the­
                                                         ory, are responsible for most social
    Maps for repeat streets. Commonly                    control in neighborhoods, their absence
    available mapping programs make it easy              leads to higher levels of deviancy. Other
    to identify hot spot places or hot spot              factors, such as poverty and racism, also
    areas, but do not make linear hot spots              have been identified as undermining
    easy to identify. Simple dot maps can be             social networks.
    used to identify hot street segments, and
    this may be the most straightforward             ■   Social efficacy. Recent evidence from
    method. Most clustering algorithms,                  Chicago points to the role of social effi­
    unfortunately, will show areas of concen­            cacy, which is “the willingness of local
    tration even when a line is the most                 residents to intervene for the common
    appropriate dimension. If high levels of             good.” It depends on “mutual trust and
    precision are not required, such area                solidarity among neighbors” (Sampson,
    maps may be adequate.                                Raudenbush, and Earls, 1997, page 919).
                                                         Neighborhoods that have a great deal of
                                                         social efficacy have less crime and disor­
    Neighborhoods and other area
                                                         der than neighborhoods that have low
    hot spots
                                                         levels. Social efficacy—like disorganiza­
    More has been written about neighbor­                tion and social networks—is not a prop­
    hood concentrations of crime (hot spots)             erty of individual people or places, but a
    than about any other form of concentra­              characteristic of groups of people.
    tion of crime. In their pathbreaking book
    Social Factors in Juvenile Delinquency           ■   Broken windows theory. The broken
    (1931), Shaw and McKay noted persistent              windows theory also is an area theory of
    concentrations of deviancy in the 1920s.             crime concentration. Wilson and Kelling
    They noted that some neighborhoods had               (1982) claim that in most well-function-
    high levels of juvenile delinquency, year in         ing neighborhoods, small transgressions
    and year out, decade after decade, regard­           of social norms (e.g., failure to keep
    less of who lived in the areas (Shaw and             one’s yard tidy) result in social pressures
    McKay, 1969). Since that time, many                  to bring the offending party into compli­
    explanations for differences in neighbor­            ance. Once a place becomes untended,
    hood crime levels have surfaced. Most of             however, it undermines the willingness
    these theories focus on the ability of local         and ability of residents to enforce social
    residents to control deviancy (Bursik and            order. Consequently, residents withdraw
    Grasmick, 1993).                                     from enforcing neighborhood norms,
                                                         which allows further deviancy to occur.
    Underlying causes. Explanations for dif­             This in turn results in additional with­
    fering neighborhood crime levels include             drawal and fear and the neighborhood
    the following:                                       begins to spiral downward. Skogan
                                                         (1990) found evidence in support of this
    ■   Social disorganization theory. This the­         basic thesis, although others suggest
        ory suggests that the natural ability of         the evidence is weak (Harcourt, 1998) or
        people to control deviancy in their neigh­       show that the theory is seriously flawed
        borhoods is impaired in some areas by            (Taylor, 2000).
        constant residential turnover and net

8
MAPPING CRIME: UNDERSTANDING HOT SPOTS

■   Crime opportunity theories. Another          levels of hot spots. By now, it should be
    explanation for neighborhood-level hot       obvious that each form of concentration—
    spots comes from routine activity theory     place, victim, street, or neighborhood—
    and related theories that point to crime     requires its own form of mapping. It
    opportunities as the principle cause of      should also be apparent that the types of
    crime. Rather than concentrations of         actions police should take correspond to
    offenders or the absence of social con­      the type of the concentration. These fac­
    trols, opportunity theories suggest that     tors have important implications for how
    analysts should look for concentrations      maps of hot spots are constructed and
    of crime targets. For example, a dense       how the hot spots are depicted.
    urban neighborhood with no off-street
    parking will have many cars parked on        Dot maps. When hot spots are at specific
    the street. Such an area may become an       addresses, corners, and other places, the
    area hot spot for thefts from vehicles. A    relevant depiction of the hot spot is a dot
    suburban subdivision inhabited by dual-      because mappers want to distinguish
    income families will have few people at      between the places with problems and
    home during weekdays. Since their            very nearby places without problems.
    property is unprotected, their neighbor­     Such distinctions are critical for delivering
    hood can become an area burglary hot         effective and efficient action. A gas station
    spot. Note that in this type of situation,   with many robberies needs to be distin­
    several layers of hot spots can exist        guished from the gas station across the
    simultaneously. Within area hot spots,       street with no robberies. In this circum­
    defined by the subdivision in this exam­     stance, a map highlighting a street or area
    ple, might be streets with even greater      is far less useful to police than a map high­
    numbers of burglaries, and some of the       lighting the gas stations that are robbery
    homes on these streets may be broken         hot spots. Dot maps of crime places can
    into multiple times.                         identify widely spread locations that are
                                                 hot spots. Such places might be over­
Maps for area hot spots. Problems                looked if lines or polygons are used to
arising from processes described by              define hot spots.
neighborhood-level theories are best
depicted on maps by shaded areas, rather         Line maps. When the hot spots are along
than dots or lines. Area hot spots on maps       streets, point maps and area maps are of
can be shown in a variety of ways: ellipses,     far less utility than line maps. Point maps
shaded areas (choropleth maps), or crime-        draw attention to the hot spot places along
frequency gradients (e.g., isoline maps that     the street and imply that the intervening
depict crime frequency or risk as graduat­       locations have low risk, when they may be
ed contours, just as feet above sea level is     future targets. Area maps include streets
depicted on topographical maps).                 that have few or no crimes. Street rob­
                                                 beries of people leaving bars and night­
                                                 clubs are good examples of this. The bars
Selecting the Appropriate                        and nightclubs are specific points, but the
                                                 robberies do not occur there. These enter­
Hot Spot Map                                     tainment spots may be concentrated in
                                                 one neighborhood, but even within this
Action level, hot spot level,                    neighborhood, many streets do not have
and mapping                                      street robberies. The robberies may occur
The discussion so far has highlighted the­       along streets leading from the entertain­
ories relevant to understanding different        ment spots to car parking locations.

                                                                                                 9
SPECIAL REPORT / AUG. 05

     Knowing which streets have the robberies           Limitations of hot spot maps
     and which do not is critical for addressing
     such a concentration. So showing this              Concentrations of victimization sometimes
     form of hot spot requires lines—straight,          can be shown with maps, but often they
     jointed, curved, or intersecting.                  cannot. If victimization risk is in part geo­
                                                        graphical, then maps are useful. A city­
     Ellipse, choropleth, and isoline maps.             wide dot map of gas stations with two or
     When hot spots cover broader areas and             more robberies within the last 6 months
     coincide with neighborhoods, they need to          shows concentration at two levels. The
     be depicted in another way. Ellipse and            dots depict concentrations of robbery at
     choropleth maps imply that the areas with­         specific places. Groupings of dots depict
     in the designated hot spots share the              streets or neighborhoods with concentra­
     same risk level, so a specific street or loca­     tions of repeat robbery gas stations. Dot
     tion within the area is irrelevant. Isoline        maps for this type of victimization makes
     maps imply a continuous gradient of risk           some sense, but they do not work for all
     within a hot spot, so a particular place has       forms of victimization concentration. If vic­
     risks similar to but not the same as an            tims are mobile, street or area maps might
     adjacent place or street. A gang-related           be more useful. However, the use of maps
     robbery problem can be an example. If              is limited for some forms of victimization
     gang members commit robberies through­             analysis. If the population of potential vic­
     out specific neighborhoods (i.e., do not           tims is spread throughout an area (not
     focus on specific streets or around specif­        concentrating at places, along streets, or
     ic sites), but refuse to commit robberies          within neighborhoods), the analyst would
     outside their territories, and their territorial   be better off using an analytical technique
     boundaries are streets, then a choropleth          other than maps to convey the concentra­
     map might be useful. One could create a            tion. For example, taxicab robberies may
     map of the gang areas and shade the                be spread quite thinly across a city. The
     areas according to the robbery frequency           relevant features of the robbery victims
     within each. If the likelihood of a gang-          might be related to the cab companies,
     related robbery diminishes the farther one         the drivers’ ages, hours of operation,
     goes from the center of gang activity, then        installed security within cabs, or a host of
     an isoline map depicting gradients of rob­         factors that cannot be shown on a map.
     bery frequency does a better job of show­          Police officers trying to investigate or pre­
     ing the problem.                                   vent such robberies would find maps less
                                                        useful than bar charts showing the charac­
     Ellipses may be far less useful. They sug­         teristics of victims and nonvictims.
     gest a firm boundary between crime on
     the inside and no crime on the outside,            Exhibit 2 links the major points discussed
     but they frequently do not follow natural          thus far. The first two columns are from
     movement patterns of people. Using an              exhibit 1. The third column shows where
     ellipse to define an area hot spot is like         the police action needs to be focused. If
     saying, “Look in this general area,”               the concentration level, action level, and
     because neither its shape nor its boundary         form of hot spot depiction are not aligned,
     are likely to conform to the nature of the         then the map will be useless at best and
     underlying problem. Consequently,                  suggest inappropriate action at worst. A
     ellipses provide police officers with far          map depicting hot streets or areas does
     less information than other ways of depict­        not help identify places where nuisance
     ing area hot spots.                                abatement would be useful. Alternatively,
                                                        a point map is too specific for implement­

10
MAPPING CRIME: UNDERSTANDING HOT SPOTS

ing street reconfigurations or neighbor­              attention is needed. At the other extreme,
hood redevelopment efforts.                           focusing attention on point locations when
                                                      the problem is at the area level focuses
The consequences of using the wrong                   attention at too precise an area and sug­
type of map are not equal. Point maps are             gests action that is too focused.
more forgiving than street or areas maps.
Dot maps allow the user to see the under­             Maps convey powerful messages to their
lying pattern of crime and determine                  readers, most of whom are not knowledge­
whether to go up a level. However, maps               able about the technicalities of crime map­
of hot streets or hot areas often do not              ping. These messages are conveyed in
show the hot places, thus place concen­               symbols, as shown in exhibit 3. Dots (A)
tration can remain hidden. This suggests              draw attention to specific places and sug­
that crime mapping should start at the                gest that places without dots can be
lowest level and work upward to avoid                 ignored. A point conveys the message that
overlooking low-level concentrations                  the hot spot is located at this exact location
where effective action can be taken.                  and should be the focus of police efforts. A
                                                      shaded street segment (B) suggests that
                                                      the chances of crime are roughly equal
Conclusion                                            along the entire segment and police efforts
                                                      should focus along this line, but not along
Different kinds of hot spots, which devel­            other lines. A shaded area (C), such as one
op from different causes, require different           used in a choropleth map, also suggests
kinds of police action. For crime mappers,            equivalent risks of crime throughout the
this means that the visual display of the             area with a dramatic reduction in risk at the
crime pattern on the map should be consis­            border. It suggests that police activity
tent with the type of hot spot and possible           throughout the area is appropriate. An area
police action. Plotting area maps when the            covered by a gradient (D), such as that
hot spots are addresses is not useful to the          depicted in isoline maps, implies that a cen­
police officers using the map because the             ter of high-crime activity exists and that
map is imprecise. It directs their attention          criminal activity tapers off gradually from
to large areas where little effort needs to be        that center. It directs police attention to the
expended and away from the places where               center and its surroundings. Each way of

Exhibit 2. Concentration, mapping, and action

Concentration                Hot spot depiction          Action level              Action examples

Place—at specific          Points                      Place, corner            Nuisance abatement, hot
addresses, corners or                                                           spot patrols
other places
Among victims              Points, lines, and areas    High-risk targets and    Developing networks
                           depending on the nature     potential victims        among potential victims,
                           of concentration                                     repeat victimization
                                                                                programs
Street—along streets       Lines                       Streets, highways        Concentrated patrolling of
or block faces                                                                  specific streets, traffic
                                                                                reengineering
Area—neighborhood          Ellipses, shaded areas,     Large areas              Community partnerships,
areas                      and gradients                                        neighborhood redevelopment

                                                                                                             11
SPECIAL REPORT / AUG. 05

     depicting hot spots is connected with use­                  3. Some disagreement exists over the geographic
     ful theories, each of which suggests differ­                scope of neighborhood theories of crime. Most
                                                                 researchers refer to areas covering several square
     ent types of police action. Recognition of                  blocks, although Taylor (1997, 1984) makes a strong
     these links in mapping practice will lead to                case for the relevant area being about the size of a
     better use of crime maps.                                   single block. Clearly, the difference between a large
                                                                 place and a small neighborhood is blurry.

     Notes
     1. Although one could have hot spots of anything that
                                                                 References
     can be geographically distributed—a hot spot of auto­       Block, R., and C. Block. 1995. “Space, Place and
     mobile dealerships, for example—usage of this term          Crime: Hot Spot Areas and Hot Place of Liquor-
     is restricted to crime or disorder. So unless other­        Related Crime.” In J.E. Eck and D. Weisburd (eds.),
     wise specified, hot spots means hot spots of crime          Crime and Place (vol. 4, pp. 145–184). Monsey, NY:
     or disorder.                                                Criminal Justice Press.

     2. Level does not indicate superiority or rank in this      Block, R., M. Felson, and C.R. Block. 1985. “Crime
     instance. A high-level hot spot is not better or worse      Victimization Rates for Incumbents of 246
     than a low-level hot spot. Rather, higher level hot spots   Occupations.” Sociology and Social Research, 69,
     can be composed of groups of lower level hot spots.         442–451.
     In this sense, level refers to level of aggregation.

     Exhibit 3. Messages in crime maps

                                                                                          Chance of a crime site
          Chance of crime being                                                           being part of the pattern is
          part of pattern is high                                                         equal anywhere along
          at site, but zero at                                                            street segments, and zero
          other places.                                                                   elsewhere.

              Everywhere outside                A                 B
              the chance of being a
              crime site that is part
              of the pattern is zero.

          Everywhere inside is
          equally likely to be a                    C
          crime site that is part
                                                                                    D
          of the pattern.

                                                                                  Gradient indicates that the
                                                                                  probability of being a crime
                  Chance of a site being part
                                                                                  site that is part of the pattern
                  of the pattern changes
                                                                                  varies from low near the edge
                  radically at the edge.
                                                                                  to high near the center.

12
MAPPING CRIME: UNDERSTANDING HOT SPOTS

Brantingham, P.L., and P.J. Brantingham. 1981. “Notes     Menard, S. 2000. “The ‘Normality’ of Repeat
on the Geometry of Crime.” In P.J. Brantingham and        Victimization From Adolescence Through Early
P. L. Brantingham (eds.), Environmental Criminology       Adulthood.” Justice Quarterly, 17(3), 543–574.
(pp. 27–54). Beverly Hills, CA: Sage Publications.
                                                          Sampson, R.J., S.W. Raudenbush, and F. Earls. 1997.
Brantingham, P.L., and P.J. Brantingham. 1995.            “Neighborhoods and Violent Crime: A Multilevel
“Criminality of Place: Crime Generators and Crime         Study of Collective Efficacy.” Science, 227, 918–924.
Attractors.” European Journal of Criminal Justice
Policy and Research, 3, 5–26.                             Shaw, C.R., and H.D. McKay. 1931. Social Factors in
                                                          Juvenile Delinquency (vol. 2, no. 13). Washington,
Bursik, R.J., and H.G. Grasmick. 1993. Neighborhoods      DC: U.S. Government Printing Office.
and Crime: The Dimensions of Effective Community
Control. New York, NY: Lexington Books.                   Shaw, C.R., and H.D. McKay. 1969. Juvenile
                                                          Delinquency and Urban Areas (revision of 1942 ed.).
Eck, J. 2000. “Policing and Crime Event Concentration.”   Chicago, IL: University of Chicago.
In L.W. Kennedy, R.F. Meier, and V.F. Sacco (eds.),
The Process and Structure of Crime: Criminal Events       Sherman, L.W. 1997. “Policing for Crime Prevention.”
and Crime Analysis. New Brunswick, NJ: Transaction        In L.W. Sherman, D. Gottfredson, D. MacKenzie, J.
Press.                                                    Eck, P. Reuter, and S. Bushway (eds.), Preventing
                                                          Crime: What Works, What Doesn’t, What’s Promising
Eck, J.E. 1994. Drug Markets and Drug Places: A           (pp. 8-1–8-58). Washington, DC: U.S. Department of
Case-Control Study of the Spatial Structure of Illicit    Justice, National Institute of Justice.
Drug Dealing. Unpublished Ph.D. dissertation,
University of Maryland.                                   Sherman, L.W., P.R. Gartin, and M.E. Buerger. 1989.
                                                          “Hot Spots of Predatory Crime: Routine Activities
Eck, J.E. 1998. “What Do Those Dots Mean?                 and the Criminology of Place.” Criminology, 27(1),
Mapping Theories With Data.” In D. Weisburd and           27–55.
T. McEwen (eds.), Crime Mapping and Crime
Prevention (vol. 8, pp. 379–406). Monsey, NY:             Skogan, W.G. 1990. Disorder and Decline: Crime
Criminal Justice Press.                                   and the Spiral of Decay in American Neighborhoods.
                                                          New York, NY: Free Press.
Eck, J.E., J.S. Gersh, and C. Taylor. 2000. “Finding
Crime Hot Spots Through Repeat Address Mapping.”          Spelman, W. 1995a. “Criminal Careers of Public
In V. Goldsmith, P.G. McGuire, J.H. Mollenkopf, and       Places.” In J.E. Eck and D. Weisburd (eds.), Crime
T.A. Ross (eds.), Analyzing Crime Patterns: Frontiers     and Place (vol. 4, pp. 115–144). Monsey, NY: Criminal
of Practice (pp. 49–64). Thousand Oaks, CA: Sage          Justice Press.
Publications.
                                                          Spelman, W. 1995b. “Once Bitten, Then What?
Eck, J.E., and D. Weisburd. 1995. “Crime Places in        Cross-Sectional and Time-Course Explanations of
Crime Theory.” In J. E. Eck and D. Weisburd (eds.),       Repeat Victimization.” British Journal of Criminology,
Crime and Place (vol. 4, pp. 1–33). Monsey, NY:           35, 366–383.
Criminal Justice Press.
                                                          Stedman, J., and D.L. Weisel. 1999. “Finding and
Farrell, G., and K. Pease. 1993. Once Bitten, Twice       Addressing Repeat Burglaries.” In C.S. Brito and
Bitten: Repeat Victimization and Its Implications for     T. Allen (eds.), Problem-Oriented Policing: Crime-
Crime Prevention (vol. 46). London: Her Majesty’s         Specific Problems, Critical Issues and Making POP
Stationery Office.                                        Work (vol. 2, pp. 3–28). Washington, DC: Police
                                                          Executive Research Forum.
Harcourt, B.E. 1998. “Reflecting on the Subject: A
Critique of the Social Influence Conception of            Taylor, R.B. 1997. “Social Order and Disorder of Street
Deterrence, the Broken Windows Theory and Order           Blocks and Neighborhoods: Ecology, Microecology
Maintenance Policing New York Style.” University          and the Systematic Model of Social Disorganization.”
Michigan Law Review, 97, 291–389.                         Journal of Research in Crime and Delinquency, 34(1),
                                                          113–155.
Harries, K. 1999. Mapping Crime: Principle and
Practice. Washington, DC: U.S. Department of
Justice, National Institute of Justice.

                                                                                                                    13
SPECIAL REPORT / AUG. 05

     Taylor, R.B. 2000. Breaking Away from Broken            Uchida (eds.), Drugs and Crime: Evaluating Public
     Windows: Baltimore Neighborhoods and the                Policy Initiatives. Thousand Oaks, CA: Sage
     Nationwide Fight Against Crime, Grime, Fear, and        Publications.
     Decline. Boulder, CO: Westview Press.
                                                             Weisburd, D., and L. Green. 1995. “Policing Drug Hot
     Taylor, R.B., S.D. Gottfredson, and S. Brower. 1984.    Spots: The Jersey City Drug Market Analysis
     “Block Crime and Fear: Defensible Space, Local          Experiment.” Justice Quarterly, 12(4), 711–736.
     Social Ties, and Territorial Functioning.” Journal of
     Research in Crime and Delinquency, 21, 303–31.          Wilson, J.Q., and G.L. Kelling. 1982. “Broken
                                                             Windows: The Police and Neighborhood Safety.”
     Weisburd, D., and L. Green. 1994. “Defining the         Atlantic Monthly, March, 29–38.
     Street Level Drug Market.” In D.L. MacKenzie and C.

14
Chapter 2. Methods and Techniques for
Understanding Crime Hot Spots
Spencer Chainey, Jill Dando Institute of Crime Science, University College London

This chapter presents a number of meth­           these points will be aggregated to small
ods and techniques to understand and              area geographies such as beats or census
describe patterns of hot spots in crime           blocks to demonstrate the applicability of
data. It explains the advantages and disad­       certain techniques. The methods dis­
vantages of certain techniques, focusing          cussed and tested on these data are per­
on methods that are easy to understand            fectly applicable to analysts’ own data.
and practical to apply. Four data sets are
used to test the methods. The results help
evaluate how the methods improve under­           Preliminary global
standing of crime patterns. This chapter
does not attempt to find the optimal
                                                  statistical tests
method. Rather, it presents a procedure           A number of simple-to-use global statisti­
for applying a number of complementary            cal tests can be used to help analysts
methods that can help analysts under­             understand general patterns in the crime
stand hot spots in the data.                      data presented here. These tests
                                                  include—
The four data sets are from the London
Metropolitan Police Force’s Crime Report          ■   Mean center.
Information System for Hackney Borough
Police for the period June 1999 through           ■   Standard deviation distance.
August 1999. This chapter contains tests
                                                  ■   Standard deviation ellipse.
developed and conducted to help analysts
understand hot spots in representative
                                                  ■   Tests for clustering.
samples of crime data sets of—

■   All crime (9,972 records) and three           Mean center
    subsets.
                                                  The mean center point can be used as a
■   Street robbery (588 records).                 relative measure to compare spatial distri­
                                                  butions between different crime types or
■   Residential burglary (1,068 records).         against the same crime type for different
                                                  periods of time (i.e., for measuring spatial
■   Vehicle crime (1,747 records).                shifts in the same crime type). For exam­
                                                  ple, exhibit 1 shows the mean center
The geographic area for these data is             points for all crime, street robbery, resi­
insignificant to explain the different meth­      dential burglary, and vehicle crime. The all-
ods for trying to understand crime hot            crime mean center can be used as a
spots. The data should be treated merely          control to compare against crime type
as point events of crime within a geo­            subsets. Residential burglary has a mean
graphical boundary area. In some cases,           center that is more north than all crime

                                                                                                  15
SPECIAL REPORT / AUG. 05

     and the two other crime subsets. The                    tance, the more dispersed are the crime
     mean center for vehicle crime is the far­               data. These results show that vehicle
     thest south of all the crime types, and                 crime is the most dispersed; robbery is
     street robbery is nearly as far north as that           the least.
     for residential burglary but slightly farther
     west. These mean centers can be used to                 Standard deviation ellipses
     generally indicate that residential burglary
     and robbery offenses show a greater ten­                Levels of dispersion also can be presented
     dency to occur in the northern part of the              using standard deviation ellipses. The size
     borough and that vehicle crime affects the              and shape of the ellipse help explain the
     southern areas of the borough more.                     degree of dispersion, and its alignment
                                                             helps to explain the crime type’s orienta­
                                                             tion. Exhibit 3 shows standard deviation
     Standard deviation distance                             ellipses for the four crime types. The sub­
     Measures of standard deviation distance                 tle differences between the ellipses help
     help explain the level and alignment of dis­            describe the relative differences in disper­
     persion in the crime data. These statistics             sion and alignment of the four crime
     are best used as relative measures, com­                types’ patterns. The ellipse with the small­
     paring crime types against each other or                est area (robbery) is the least dispersed of
     the same crime types for different periods              the crime types. The position of the rob­
     of time. Exhibit 2 shows the standard devi­             bery ellipse farther north of all crime and
     ation distances for the four crime types.               vehicle crime, but slightly farther south of
     The greater the standard deviation dis­                 the residential burglary ellipse, reflects its

     Exhibit 1. Mean center points for the four crime categories
                    All crime
                    Street robbery
                    Residential burglary
                    Vehicle crime
                    Borough boundary

16
MAPPING CRIME: UNDERSTANDING HOT SPOTS

                                                            assume that crime distributions are clus­
Exhibit 2. Standard deviation distance results for          tered, and whether clusters exist or not,
the four crime categories                                   some can be identified from random crime
                              Standard deviation distance   distributions. Testing for clustering is the
Crime type                             (meters)             first step in revealing whether data has hot
                                                            spots of crime.
All crime                                1,807.94
Robbery                                  1,749.94           Several approaches can be applied to test
                                                            for clustering in crime distributions. Most
Residential burglary                     1,806.28
                                                            methods incorporate the basic principles
Vehicle crime                            1,820.85           of hypothesis testing and classical statis­
                                                            tics, in which the initial assumption is that
                                                            the crime distribution is one of complete
mean center. Its north-west, south-east                     spatial randomness (CSR). By setting the
orientation also helps to describe the gen­                 CSR assumption as the null hypothesis,
eral direction toward which robbery crimes                  the crime distribution can be compared
have a tendency to be patterned.                            against a set significance level to accept or
                                                            reject the null hypothesis. Some tests for
                                                            clustering are the nearest neighbor index
Tests for clustering
                                                            and the spatial auto correlation tests.
The fourth and probably most useful of the
preliminary global statistics are those that                Nearest neighbor index (NNI). The NNI
test for clustering. Crime analysts often                   is a simple and quick method to test for

Exhibit 3. Standard deviation ellipses for the four crime categories

                  All crime
                  Robbery
                  Residential burglary
                  Vehicle crime

                                                                                                            17
SPECIAL REPORT / AUG. 05

     evidence of clustering. The NNI test com­       The results also show the differences
     pares the actual distribution of crime data     between the NNI results for a minimum
     against a data set of the same sample           bounding rectangle area and the actual
     size, with random distribution. It can be       catchment area of the crime points. When
     applied when the user has access to data        the actual catchment area is known or can
     in which each point relates to individual       easily be calculated, it should be used in
     crime events (irrespective of whether           the calculation of the NNI. If the area is
     some of these events are mapped on top          not known, a minimum bounding rectan­
     of each other at exactly the same loca­         gle around the crime distribution often is
     tion). The NNI method is explained in the       used to calculate the area representing
     following steps:                                the crime data’s catchment zone.

     ■   For each point, calculate the distance to   However, minimum bounding rectangle
         the nearest neighbor point.                 areas used for NNI tests often can create
                                                     misleading results for describing point dis­
     ■   Sum the nearest neighbor distance for       tributions. To test and show evidence of
         all points and divide by the number of      this, analysts purposely designed a regular
         points in the data. This value is the       distribution of points and applied three dif­
         observed average nearest neighbor           ferent types of bounding areas: minimum
         distance.                                   bounding rectangle, bounding convex hull,
                                                     and true boundary area. After applying NNI
     ■   Create a random distribution of the         tests, the bounding convex hull method
         same number of crime points covering        and true boundary area revealed similar
         the same geographic area, and for each      results, correctly describing the point
         point calculate the distance to each
         nearest neighbor point.

     ■
                                                     Exhibit 4. Nearest neighbor analysis results for the
         Calculate the sum of the nearest neigh­
                                                     four crime data sets
         bor distances for all these randomly
         distributed points and divide by the        Crime type and
         number of points in the data. This value    bounding area*                      NNI           Z-score
         is the average random nearest neighbor
                                                     All crime
         distance.
                                                     Bounding rectangle area             0.32          –129.11
     ■   The NNI is then the ratio of the            True boundary area                  0.46          –103.14
         observed average nearest neighbor dis­
                                                     Robbery
         tance against the average random near­
         est neighbor distance.                      Bounding rectangle area             0.59           –19.14
                                                     True boundary area                  0.80            –9.20
     If the result generated from the NNI test is
                                                     Residential burglary
     1, then the crime data are randomly dis­
     tributed. If the NNI result is less than 1,     Bounding rectangle area             0.57           –27.14
     then the crime data show evidence of            True boundary area                  0.74           –16.46
     clustering. An NNI result that is greater       Vehicle crime
     than 1 reveals evidence of a uniform pat­
                                                     Bounding rectangle area             0.52           –38.73
     tern in crime data.
                                                     True boundary area                  0.72           –22.16
     Exhibit 4 shows the NNI results for the
     four crime data sets. All four sets show        *Crime distribution is clustered for all areas.
     evidence of clustering in their distribution.

18
You can also read