The Analysis of Effectiveness for CPTED Project Using Weighted Displacement Quotient (WDQ)

1. Background and Purpose of Research

According to the 2018 Social Survey by Statistics Korea, 20.6% of Korean citizens felt that the main factor of anxiety was crime occurrence, which was significantly higher than other areas of social safety. Since the 2000s, the central and local governments have implemented CPTED projects under the Crime Prevention Through Environmental Design approach to reduce fear of crime and to prevent crimes. In particular, local governments have actively enforced CPTED projects following the 2012 Yeomni-dong Salt Way project by the Seoul Metropolitan Government.

Many studies have analyzed the effectiveness of CPTED projects and their widespread implementation. Research analyzing the effectiveness of CPTED projects can be largely divided into studies focused on changes in crime rate based on crime data in the target area, and surveys of fear of crime in the target area before and after the project. Due to difficulties involved in acquiring geographic crime data, most studies have analyzed fear of crime and change in disorder through residential surveys and 112 crime reports. Research on crime displacement, which refers to the relocation of crime to nearby areas under Crime Prevention Through Environmental Design, and the diffusion effect of crime control is hardly available in Korea.

This paper examines the crime prevention effect of CPTED projects using geographic crime data. It analyzes the actual crime prevention effect of CPTED projects, and relates effectiveness to crime type, namely, burglary and violence. In addition, it seeks to verify the crime displacement effect and diffusion effect of crime control, which is the reduction of crime in nearby areas as a result of CPTED projects.

1. Analysis of Effectiveness of CPTED Projects

According to the National Police Agency, the number of local CPTED projects increased from 305 in 2016 to 1,555 in 2018. Despite this continuous increase, the effectiveness of CPTED projects in crime prevention was not systematically verified due to the difficulties involved in acquiring data on crime occurrence. Most research analyzing crime prevention effectiveness was in the form of analysis of effectiveness in areas in which CPTED projects were implemented, and analysis of effectiveness of the CPTED approach, including the installation of CCTVs.

Kang et al. (2018) investigated trends of crime occurrence and changes in disorder in low-income residential areas where CPTED projects were implemented using crime statistics and 112 crime reports. The results showed that CPTED projects were effective in reducing burglary, but not sexual violence, violence, theft, and murder. In the case of disorder, the positive effects were not as distinct as that of the five major crimes, and differences were observed across regions.

Kim (2007) and Yun (2015) surveyed residents in CPTED project target areas, and analyzed project effectiveness based on residents’ fear of crime before and after implementation. Kim (2007) found that physical constituents such as crime prevention facilities and techniques lowered fear of crime, and strengthened social constituents such as social solidarity and reliability. The strengthened social constituents also helped to reduce fear of crime. That is, CPTED projects were effective in reducing fear of crime in target areas. Yun (2015) showed that CPTED projects lowered crime occurrence and reduced fear of crime. The fear of crime was more strongly influenced by social disorder than physical disorder. As such, efforts should be exerted to prevent social disorder when implementing CPTED projects.

Seo et al. (2018) analyzed the effectiveness of a CPTED project based on indirect factors such as a survey of residents in the target area, interviews with residents, changes in floating population before and after project implementation, and changes in perception of others. The results showed that the floating population increased in the target area, which can be interpreted as a positive effect of the project. However, the floating population decreased one year after the end of the project, which demonstrates that the effect of the project lacks sustainability. Similar results were observed in the survey and interviews. The CPTED project was effective in reducing fear of crime immediately after implementation, but quickly declines in effectiveness if not accompanied by follow-up measures.

Research analyzing the effectiveness of CPTED projects, which aim to reduce crime, is difficult to carry out without the cooperation of the National Police Agency, and most studies tend to rely on indirect measurements of effectiveness.

2. Crime Displacement and Diffusion Effect of Crime Control

The ultimate goal of CPTED projects is to suppress the occurrence of crimes by weakening perpetrators’ crime motivation and imposing constraints on criminal acts through changes in the physical environment of spaces vulnerable to crime and community enforcement. This intervention may result in crime displacement, which is the relocation of crime to nearby areas. As such, the effectiveness of CPTED projects should be evaluated with consideration of the crime displacement effect and the diffusion effect of crime control.

Weisburd et al. (2006) defined crime displacement as “shift of crime either in terms of space, time or type of offending from the original targets of crime prevention interventions.” The six types of crime displacement are: spatial displacement in which perpetrators move from areas where crime is difficult to commit to another location where they face less risks, temporal displacement in which perpetrators shift the timing of their offenses to hours or days seen as less risky, target displacement in which offenders give up well-protected targets and focus on more vulnerable ones, tactical displacement in which perpetrators change the tactics they use to avoid an obstacle designed to thwart their efforts, perpetrator displacement in which other perpetrators take the place of perpetrators who are arrested or give up committing offenses, and crime type displacement in which perpetrators commit entirely different type of offenses when a certain crime is not possible (Barnes, 1995:96).

Meanwhile, crime displacement researchers observed effects that are the complete opposite of displacement. Clarke & Weisburd (1994) referred to the spread of crime reduction benefits beyond the target area as “diffusion of crime control benefits,” and defined it as “the spread of the beneficial influence of an intervention beyond the places which are directly targeted, the individuals who are the target of control, the crimes which are the focus of intervention or the time periods in which an intervention is brought.” The diffusion of benefits occurs when perpetrators are uncertain about the risks of detection and exposure when committing offenses, and when the rewards of certain offenses are weaker than their efforts (Clarke & Weisburd, 1994). Lee & Kim (2009) found that the diffusion of crime control benefits not only suppressed crimes in nearby areas, but also suppressed other types of crime, and could also sustain the effect of crime reduction.

When analyzing the actual effectiveness of crime prevention activities, the analysis should not simply be confined to the target area, but should take into account the possible effects of crime displacement and diffusion of benefits in nearby areas. Past studies on the subject by local and international researchers have mostly relied on the Weighted Displacement Quotient (WDQ) proposed by Bowers & Johnson (2003) to verify the effects of crime displacement and diffusion of benefits.

3. Understanding of WDQ

Bowers & Johnson (2003) developed WDQ as a tool to measure crime displacement effect and diffusion of benefits.

A quasi-experimental design was incorporated as the basic framework of research, and the effects of the experiment were evaluated using a nonequivalent control group design with pretest–posttest. The experimental results were compared for the target area, comparison area, and buffer area (Lee & Kim, 2009:260). The target area is the area in which crime prevention or a CPTED project was implemented. The buffer area is the area expected to be involved in crime displacement as a result of crime prevention activities in the target area. The comparison area, which serves as the control, is unrelated to changes in crime occurring in target and buffer areas. When setting buffer and control areas, it should be noted that the three areas must be homeogenous in terms of geographic properties, demographics, and criminal characteristics to overcome the problem of how variables cannot be controlled between groups (Lee & Kim, 2009:263). The control area must be larger than the target area as a smaller control area would cause severe fluctuations in measurements of crime change rate relative to changes in reference time period. Since the buffer area is set to be within a certain radius of the target area, it is generally larger than the target area (Lee & Kim, 2009:263-264).

Bowers & Johnson (2003) proposed three principles of WDQ. First, the crime change rate of the control area is compared to that of the target area and buffer area, and serves as a criterion in describing the crime displacement effect. Second, spatial displacement to be described by WDQ presumes that crime is displaced from the target area to the buffer area. Third, the occurrence of crime displacement means that the target area sees a decrease in crime due to crime prevention activities, while the buffer area experiences a higher increase in crime than the control area. The WDQ equation below reflects the three principles, and integrates spatial and temporal concepts.

Crime displacement and diffusion of benefits can be verified by comparing changes in the crime rate between the target area and control area, and at the same time, changes in crime rate between the buffer area and control area. According to the above equation, the denominator (A′) represents the effect of crime prevention activities or the CPTED project in the target area. The denominator (A′) generally assumes a negative (−) value when crime prevention activities are performed, which demonstrates the effectiveness of such activities in reducing crime. A positive denominator (A′) means that the decrease in crime rate in the target area is lower than that of the control area, indicating that crime prevention activities were not effective in reducing crime. The numerator (B′) measures the crime displacement effect and diffusion of benefits in the buffer area. A negative (−) value means that crime prevention activities resulted in diffusion of benefits, while a positive (+) value indicates that crime was displaced to nearby areas (Bowers & Johnson, 2003:285).

WDQ can be interpreted in seven ways that are shown in Table 1 depending on the combination of denominator (A′) and numerator (B′) values. First, a WDQ larger than 1 means that crime prevention activities were effective in reducing crime in the target area, and led to diffusion of benefits. That is, crime prevention activities were largely successful. Second, a WDQ close to 1 means that the diffusion of benefits in the buffer area was similar to the crime prevention effect in the target area. Third, a WDQ between 0 and 1 means that the crime prevention effect in the target area was greater than diffusion of benefits in the buffer area. Fourth, a WDQ of 0 occurs when the numerator (B′) is 0, which means that there was no crime displacement or diffusion of benefits. Fifth, a WDQ between 0 and -1 means that crime prevention activities were effective in reducing crime in the target area, but that the buffer area was affected by crime displacement. Sixth, a WDQ close to -1 means that crime prevention activities reduced crime in the target area, but the net effect was neutral considering the crime displacement to the buffer area. Lastly, a WDQ smaller than -1 means that crime prevention activities had the reverse effect of increasing crime (Bowers & Johnson, 2003: 285-286).

Table 1.

Interpretation of the weighted displacement quotient

4. Literature Review and Distinctiveness of Research

Following the introduction of WDQ by Bowers & Johnson in 2003, many studies have applied WDQ to the analysis of crime displacement and diffusion of benefits (Bowers & Johnson, 2003; Park et al., 2011; Lim et al., 2016; Cho & Jung, 2017; Haberman & Stiver, 2018; Park et al., 2018; D. Kim et al., 2019).

1. Setting of Target Area

This study analyzed the effectiveness of CPTED projects in consideration of crime displacement and diffusion of benefits. For this purpose, the target area was set as the area in which Project B was implemented by District A.

Seoul has a higher number of crimes than other regions, and many CPTED projects were implemented following the 2012 CPTED project in Yeomni-dong. District A has the highest proportion of single-person households in Seoul, and has seen a high number of burglary and violence due to break-ins. In 2016, District A enacted the Ordinance of Urban Environment Design for Crime Prevention to help prevent crime and lower fear of crime. Under the ordinance, the district has implemented various CPTED projects, and even received the grand prize in the 2017 Korea Crime Prevention Awards for its systematic and efficient efforts.

The target area of Project B implemented by District A in 2016 is largely comprised of single-person households in multiplex housing, and shows high homogeneity in physical environment across areas. Most buildings are less than five stories high, and sidewalks are not separated from roads. Visibility is limited by the narrow streets and alleys between buildings. The area had a high proportion of residential burglary and violence before the implementation of Project B. The residents have a high fear of crime, but are less aware of crime prevention methods.

This study selected a target area where the CPTED project had been implemented in a residential area, allowing analysis of effects before and after the project, and where crimes had occurred to an extent such that changes in crime rate can be examined for the period between 2015 and 2017.

2. Status of Target Area and Project Details

This study selected Project B, implemented by District A of Seoul in 2016, to analyze crime displacement and diffusion of benefits. The target area has a high density of multiplex housing, and the narrow, dimly lit alleys make it more vulnerable to crime. The park in the target area may become a high-intensity crime area due to its poor accessibility and few night-time users. The CPTED project was implemented to improve spaces more prone to crime.

Under the CPTED project, various elements shown in Figure 1 were installed. Building security was improved with security windows, mirror sheets, safe fences, and fluorescent substance on gas piping, and alleys with hand rails on uphill sections, patrol number on telegraph poles, shelters, town information maps, patrol point signs, and lights. In addition, to enforce natural surveillance in the park, extra lights, crime prevention facilities, and a shelter were installed.

Figure 1.

Applied CPTED elements at target area

The crime prevention facilities in the target area were installed based on CPTED principles such as surveillance, access control, territoriality, activity support, legibility, and maintenance and management that are shown in Table 2. Various spatial environment design techniques were also employed to improve the visibility of crime prevention facilities.

Table 2.

Principles and elements of CPTED

3. Setting of Buffer and Control Areas

The buffer and control areas that are shown in Figure 2 were set as described below to analyze crime displacement and diffusion of benefits.

Figure 2.

Settings of buffer area and control areas

The buffer area in this study had to be as close as possible to the target area, but no larger than two times the target area in order to be consistent with criteria used in past research (Hamilton-Smith, 2002:36, Lee & Kim, 2009:279). The boundary of the buffer area was 50 meters away from the target area. To ensure a similar physical and social environment, the selected buffer area had similar building use, land use, and crime patterns. That is, the area had a high-density of multiplex housing, and saw a similar proportion of burglary and violence caused by housebreaking.

The control areas had similar environmental characteristics with the target area, and met the criteria of the buffer area. They were areas not targeted by the CPTED project. Two control areas were selected to ensure consistency in the analysis. To ensure a similar environment, the control areas were areas within the same administrative unit as the target area and had a high density of multiplex housing.

4. Setting of Hypotheses

The following hypotheses concerning crime displacement or diffusion of benefits were set based on crime counts for burglary and violence, which occurred at higher rates among the five major crimes, in the selected target area, buffer area, control area 1, and control area 2.

• Hypothesis 1. The project will have positive effects on the target area and buffer area.
  1.1 The number of crimes in the target area will decrease.
  1.2 The buffer area will see a diffusion of benefits.
• Hypothesis 2. Consistent results will be seen for burglary, but not violence.
  2.1 WDQ results will be the same for burglary regardless of control area.
  2.2 WDQ results will be different for violence depending on control area.

1. Changes in Crime Rate

The crime count in the target area was reviewed as shown in Table 3 before analyzing crime displacement or diffusion of benefits under the CPTED project. First, the cases of five major crimes in the administrative district containing the target area of Project B increased by 12 cases from 397 in 2015 to 409 in 2017. Burglaries decreased by 28 cases (187 cases → 159 cases), while violence increased by 45 cases (194 cases → 239 cases) in the same period from 2015 to 2017. The change in crime count of burglary and violence in the target area was similar to that of the administrative district. Burglaries decreased by 2 cases (18 cases→ 16 cases), while violence increased by 10 cases (14 cases→ 24 cases) in the same period.

Table 3.

Crime count in study area

This indicates that the CPTED project is more effective in preventing premeditated crimes, which is consistent with past research. The buffer area saw a similar decrease in burglaries (5 cases, 24 cases→ 19 cases), and an increase in violence (3 cases, 19 cases→ 22 cases). Unlike the target area and buffer area, control area 1 experienced 12 more cases of burglaries (27 cases → 39 cases) and 6 more cases of violence (14 cases→ 20 cases). Control area 2 showed a similar trend with 6 more cases of burglaries (10 cases→ 16 cases) and 10 more cases of violence (6 cases→ 16 cases).

2. WDQ Results

The results of the WDQ analysis for burglary and violence are shown in Table 4. The change in burglary rate in the target area is -0.54 relative to control area 1 and -0.80 relative to control area 2. The values were less than 0 regardless of area, meaning that the change in crime rate in the target area is greater than that of the control area regardless of control area number. Similar results were observed for the change in crime rate in the buffer area relative to control area. The change in burglary rate in the buffer area was -0.85 relative to control area 1 and -1.21 relative to control area 2, which are larger than values obtained for the target area. This is reflected by the WDQ values being greater than 1 for both control area. In other words, the CPTED project can be seen as having greater diffusion of benefits than a direct effect, which translates to a positive net effect.

Table 4.

Result of WDQ

For violence, WDQ values were different depending on control area. This is consistent with the findings of Park et al. (2011) in their study on crime displacement by WDQ analysis.

The change in crime rate was greater than 0 for both the target area and buffer area relative to control area 1. This shows that there was a higher number of crimes in the target area relative to the control area after project implementation. On the other hand, the change in the buffer area was smaller than 0 at –0.26, meaning that there was a smaller number of crimes in the buffer area relative to the control area after project implementation. Theoretically, measuring the crime displacement effect is meaningless if the target area has seen an increase in crime rate. The change in crime rate of the target area relative to control area 2, which is represented by the denominator in WDQ, was smaller than 0 (-0.83). The change in crime rate of the buffer area was –1.79, which is greater than that of the target area. As a result, the WDQ was larger than 1 at 2.15. While violence increased in the target area, the increase was smaller than in the control area. Since crime count decreased in the buffer area, the CPTED project can be seen as having a positive net effect. However, due to the different WDQ values for control area 1 and control area 2, crime displacement and diffusion of benefits cannot be measured.