Examining the Effects of Physical Environment and Structural Characteristics on the Spatial Patterns of Crime in Daegu,South Korea

Abstract

The current study examines the relationship between the features of physical and social environments and neighborhood crime in a large Korean city. We utilized the 112 Crime Calls Data from May 1, 2018 to April 30, 2019, aggregated at the ¼-mile egohood level. We estimated a series of negative binomial regression models to test the effects of social and physical environmental features on crime rates. Furthermore, we examine potential moderating effects between the measures of physical and social environments. The results indicate that incorporating the physical and social environmental features based on the theoretical framework of criminal opportunities and social disorganization can be useful for understanding the spatial patterns of crime in Korean context.

Keywords

crime neighborhood physical environment structural characteristics

Introduction

A burgeoning body of studies has found that crime is not randomly distributed over the space but rather spatially concentrated. Previous research has demonstrated that physical environmental attributes are important factors for understanding the structure of criminal opportunities—the probability of the convergence of motivated offenders, suitable targets with the absence of capable guardians at the same time and place. This is because such physical settings can largely shape the number and type of foot traffic visiting the area, and thus criminal opportunities in place. These studies often operationalized physical settings in various ways: the number and types of businesses (Bernasco & Block, 2011; Bernasco et al., 2016), housing type and age (Hipp et al., 2019; Kinney et al., 2008), the proportions of different types of land uses (Kurtz et al., 1998; Stucky & Ottensmann, 2009; Taylor et al., 1995), or land use mix (Taylor, 1988; Wo, 2019a).

In contrast, other studies have highlighted the importance of structural characteristics (Sampson et al., 1997; Sampson & Groves, 1989; Weisburd et al., 2012). These studies are based on the theoretical framework of social disorganization and informal social control. According to the theory, certain neighborhood structural characteristics such as socioeconomic status, residential stability, or racial/ethnic heterogeneity can impede the formulation of social ties and cohesion among residents, and thus weaken the capability among residents to monitor and regulate community problem including crime and disorder.

Although a large body of research has shown the importance of the physical environment and structural characteristics for understanding the location of crime, they were mostly conducted in Western countries where incident-level crime data are more readily available. However, relatively less attention has been paid to whether such neighborhood-level association still remains similar in non-Western (or non-US) context. Some studies have looked at the East-Asia region (Jiang et al., 2013; Jung et al., 2022; Liu et al., 2016; Roh et al., 2010); yet these studies focused exclusively on testing social disorganization theory while paying less attention to criminal opportunities engendered by physical environment. Moreover, only a handful of studies have looked at criminal opportunities provided by physical environment such as alcohol outlets, accommodation service businesses, or abandoned housings in Korean context (Lee et al., 2016; Shin, 2019; Yeom, 2018, 2019b). Therefore, drawing on both criminal opportunities and social disorganization theories, the current study focuses on various physical environmental features and structural characteristics to answer why some areas seem to have more crime in the City of Daegu, South Korea.

In sum, many studies (mostly done in the US) have examined the effects of the structural characteristics and physical environments on the spatial patterns of crime. Yet, there is a dearth of empirical work to confirm the generalizability of such association in other regions of the world. Thus, the current study extends the literature by analyzing the relationship between the features of structural characteristics, physical environment, and neighborhood crime in the City of Daegu, South Korea. To do so, we utilized the incident-level 112 Crime Calls Data from May 1, 2018 to April 30, 2019 and the Korean Census Data in 2018, aggregated at the ¼-mile egohood level. 112 Crime Call is a type of computer aided dispatch (CAD) system that corresponds to 911 system in the US. Call takers in the central office dispatch the closest police officers to the emergency location with the aid of geographic information system, record management system, and an automatic location system.

We contribute to the literature by doing the following: (1) check if theoretical perspectives of criminal opportunities and social disorganization in the US context are generalizable in a large Korean city context; (2) employ the most exhaustive incident-level crime data (112 Crime Calls Data) in South Korea to analyze a comprehensive measure of crime, which to our knowledge has not been utilized in previous studies; (3) test the effects of various measures of physical environment (e.g., business facilities, housing age and type, etc.) and structural characteristics on different crime types; and finally (4) test possible interactions between the measures of physical environment and structural characteristics to see if any moderating effects exist.

Specifically, the current study posits the research questions as follows: (1) How are various measures of physical environment including type of business, housing type, or housing age associated with the spatial patterns of crime in Daegu, South Korea?; (2) How are structural characteristics such as socioeconomic status or residential stability associated with the spatial patterns of crime in Daegu, South Korea?; and (3) Do physical environment and structural characteristics exhibit possible moderating effects on the spatial patterns of crime in Daegu, South Korea? In the subsequent sections, we discuss our theoretical motivations for the current study. Then, we describe our data and the methods employed, followed by the findings and their implications.

Business Facilities and Crime in Place

Previous research has revealed that certain types of business facilities can impact the location of crime (Bernasco & Block, 2011; Kim & Hipp, 2021; Tillyer et al., 2021; Wilcox et al., 2004). These studies frequently employed the criminal opportunity perspective to explain how certain situational and contextual factors of physical environment can shape the spatial patterns of crime (Cohen & Felson, 1979; Felson, 1987; Felson & Boba, 2010). Particularly, crime pattern theory looks at how a set of physical settings (activity backcloth) may influence routine activities of individuals including potential offenders and victims (Brantingham & Brantingham, 1993, 1995). According to the theory, potential offenders and victims construct routinized spatial movement patterns based on daily social activities. Busier areas with more foot traffic tend to have more criminal opportunities given that there will be higher probability of having potential offenders and targets at the same place and time.

Business facilities, in particular, are important factors for understanding the spatial patterns of crime because they can largely determine the number and type of people coming into the area. Certain types of business facilities are theorized to be crime generators (e.g., shopping centers, malls, schools, hotels, etc.) given that they draw large inflow of population visiting the area including potential offenders and targets. Although crime generators do not necessarily have criminal reputation favorable to potential offenders, they can engender criminal opportunities by attracting more foot traffic in the area. Other facilities are seen to be crime attractors (e.g., drug markets, half-way houses, bars, etc.) given their known criminal reputations for weakly guarded targets, and therefore, motivated offenders find them more attractive for committing crime.

Thus far, a handful of studies have examined the association between the presence/type of business facilities and neighborhood crime in Korean context. For example, Yeom (2018) found that the alcohol outlet density was associated with an increased number of violent crime in a large Korean city. Also, Lee et al. (2016) found that the accommodation businesses tend to increase sex crime in a Korean city. Cheong and Park (2010) also showed that the alcohol outlet and accommodation density exhibit crime-enhancing effect for homicide in 120 urban cities in Korea. Thus, we hypothesize that our measures of business will be associated with more crime: H1. The presence of more business facilities will be related to more crime. This effect can vary by different types of businesses.

The Role of Housing Type and Age for Understanding Crime in Place

In addition to business facilities, previous studies posited that housing type and age are also important physical environmental features as they can provide different criminal opportunities. For example, areas with mostly detached single family housing units will have different structure of criminal opportunities and guardianship capability, compared to those with multi-family housing units. Likewise, physical settings attached to a high-rise apartment complex may provide different opportunities for crime compared to small apartments or commercial-residential mixed units. For example, Hipp et al. (2019) found that street segments with more detached single family housing units are generally at lower risk of crime than other types of housing. For a specific crime type such as motor vehicle thefts, areas with more single family detached units are expected to be at lower risk given that a detached single-family housing is more likely to have a personal garage and an automobile parked and locked in the garage (Felson & Boba, 2010).

Physical features attached to housing units can also provide different levels of territoriality—a sense of separation between private and public space (Newman, 1972). Therefore, areas with high territoriality tend to have fewer crimes because it is clear who owns the place and performs the surveillance responsibility. In our study area context (Daegu, South Korea), for example, most high-rise apartment complexes are surrounded by fences and gates. Such features of guardianship clearly define the separation between the private property and public space. Moreover, these apartments typically have private security guards at each gate watching over the area 24/7 and admit only residents and their guests. Therefore, high-rise apartment complexes are expected to have lower risk of crime compared to other housing types. Therefore, we pose hypothesis: H2. The presence of more residential housing units will bring more guardianship capability, and result in less crime. This effect can vary by different housing types.

In terms of housing age and crime, filtering theory in housing economics posits that as older housings tend to be appreciated lower price than new housings, over time, residents with lower household income are more likely to move into older housings compared to those moving into newer houses (Rosenthal, 2014). Then, a consequence is that older housings are more likely to be physically deteriorated as the owners are less financially stable for timely maintenance, and thus the area will experience more physical disorder, which may lead to more serious crime (Perkins et al., 1992; Skogan, 1990, 2015). According to criminological theories such as broken windows, potential offenders can perceive physical deterioration of buildings as cues that residents are less actively and effectively regulate and monitor the community. Therefore, they find the areas more attractive for committing crime (Wilson & Kelling, 1982). Moreover, physical disorder from older housings and housing deterioration may reduce informal social control and collective efficacy among residents given that physical disorder can work as signs for incapability of surveillance and guardianship from other residents in neighborhood (Taylor, 1997). In Korean context, Yeom (2019b) found that the density of abandoned housings were positively associated with the number of violent crime using a sample of 637 census output areas in a Korean city. We therefore hypothesize that: H3. The presence of more older housing units will result in more crime.

The Role of Structural Characteristics for Understanding Spatial Patterns of Crime

Studies of neighborhood and crime have suggested that structural characteristics of neighborhood are important for understanding why some areas in a city have more crime than others (Hipp, 2007; Morenoff et al., 2001; Sampson et al., 1997; Sampson & Groves, 1989). Social disorganization theory underpins many of these studies. The theory posits that socially disorganized neighborhoods have lower capability of regulating the community by residents themselves to prevent crime and disorderly conducts. According to the theory, neighborhoods with certain structural characteristics (i.e., socioeconomic disadvantage, residential instability, and racial/ethnic heterogeneity) are at higher risk of crime because such structural characteristics impede the social ties and cohesion among residents, and thus reduce the level of informal social control.

Previous research has examined the applicability of social disorganization theory to Korean context (Cheong & Kang, 2013; Cheong & Kwak, 2008; Cheong & Park, 2010; Cho et al., 2021; D. S. Lee & Lee, 2009; Yoon, 2018). Given that South Korea is relatively more homogeneous in terms of racial/ethnic compositions, these studies have focused more on such neighborhood characteristics as socioeconomic status (Cheong & Kwak, 2008), divorce rates (Cheong & Park, 2010), or residential mobility (Cho et al., 2021) to predict crime or victimization rate. For example, in their Dong-level analysis,¹ Cheong and Kang (2013) found that neighborhoods with higher poverty rate are at higher risk of homicide, while residential mobility and population heterogeneity did not show statistically significant associations. Cheong and Park (2010) also found that divorce rate was positively associated with the homicide risk in Korea. We therefore hypothesize that: H4. The measures of structural characteristics will exhibit similar effects on neighborhood crime as shown in previous studies conducted in the US.

Moderating Effects

So far, we discussed how physical environment and structural characteristics can be important. We now theoretically discuss how they can collectively work together to produce different spatial patterns of crime. Previous studies have suggested that although each type of physical environment (e.g., businesses and housing) contributes to explaining crime in place, it may be contingent on other types of physical environmental features (Kim & Hipp, 2021) or structural characteristics of place (Wilcox et al., 2004). Therefore, it is more plausible to think that combinations of physical environment and structural characteristics can work collectively to increase or decrease foot traffic, social activities, and thus crime in place. Crime generators/attractors (business facilities) may have significant main effects on crime, yet they may moderate the effects of other types of physical environment (housing type and age). For example, the effect of the number of businesses may be so strong that variations in territoriality and guardianship capability constructed by more high-rise apartments abovementioned may not be impactful to the risk of crime. In contrast, it may weaken the crime reducing effects of certain physical environments. For example, areas predominantly high-rise apartments are likely to have higher levels of territoriality and guardianship capability, which may dampen the crime enhancing effects of the presence of certain types of businesses. An implication is that it is necessary to see how these physical environmental features and structural characteristics interact with each other for different spatial patterns of crime. Therefore, the current study examines potential moderating effects between the measures of businesses, housing, and structural characteristics.

Data and Methods

Our study area is the City of Daegu, South Korea located in the South-Eastern part of the Korean peninsula. It is the fourth-largest city in the country with population over 2 million. We combined two data sources for our analysis. First, we collected the 112 Crime Calls data in Daegu from May 1, 2018 to April 30, 2019. The 112 Crime Call is the Korean emergency call system that corresponds to the 911 emergency call in the US. Unlike the 911 call, the 112 call is operated specifically for crime-related emergencies. For instance, if a call has no bearing on police duties (e.g., fire or medical emergency), 112 dispatchers immediately contact other related agencies. Thus, the data mainly include emergency calls regarding crimes, disorders, traffic accidents, and other issues that can be addressed by the police. We also used the Korean Census Data in 2018 from the Statistics Korea (the Bureau of Census in Korea) for the measures of businesses, housing age and type, and structural characteristics.

Unit of Analysis

Our unit of analysis is ¼-mile egohood (N = 4,569). We created ¼ mile egohood using the Korean Census Output Area (COA) as a focal unit. The COA is the smallest unit in which Korean Census data are available, corresponding to the Census blocks in the US context (Jung et al., 2022). The egohood approach creates overlapping concentric circles that surround each block (Hipp & Boessen, 2013). Therefore, egohoods are not geographically discreate but they are rather spatially overlapping each other and thus are able to capture the spatial heterogeneity more properly within a spatial area. Figure 1 visualizes how an egohood looks like and how one egohood is spatially overlapping with another. As shown, each COA is a focal unit. We measured the distance from each focal COA centroid to other surrounding COA centroids. Then we selected all COAs within the ¼ mile distance from the focal COA and consider them (including the focal COA) as one neighborhood unit. Therefore, an egohood is not a discrete boundary but rather overlapping each other. The closer two focal units (i.e., COA) are, the more areas they will share with each other. Hipp and Boessen (2013) tested and found that structural characteristics measured at the egohood level explained more variation in crime compared to other models using conventional neighborhood units such as Census tracts and block groups in the US context. The full extent of the egohood approach is summarized by Hipp and Boessen (2013, pp. 294–295):

“When egohoods are constructed for all blocks in the city, a particular block is tied not only to the blocks in its own buffer but also to the buffers of these blocks. As such, the egohood of the focal block will contain portions of the buffers of all of the blocks within its own buffer.”

Dependent Variables

The dependent variables are the counts of crime calls. The 112 data include information such as type of crime, date and time, and the incident-level address where each case occurred. We geocoded the incident-level crime calls to the latitude and longitude points.² We then spatially aggregated the crime calls to the egohoods based on the geographic locations. We classified the crime calls into four categories: Violent crime (homicides, assaults, and robberies), property crime (burglaries and thefts), disorders (quarrels, misconducts, dine-and-dash, and all other disorders), and sex crime (forcible rape and dating violence).

Independent Variables

To account for the physical structural qualities of areas that might shape criminal opportunities, (Boessen & Hipp, 2015; Stucky & Ottensmann, 2009; Wo, 2019a, 2019b), we include various measures of business and housing. First, we include the number of business establishments for the following business types: Retail, Accommodation and food services, Public administration, Education, Health and social work, Arts/sports and recreation, and Repair and other personal services. We also include various measures of housing types (%): multi-family units, single-family units, high-rise apartments, small apartments, and commercial-residential mixed housing units. To capture the overall housing age in place, we include the measure of the average housing age. For the housing and structural characteristics measures, we converted the percent measures to counts first, aggregated from the COA to the egohood level, and then computed them back to percentages at the egohood level. For example, to make it proportional, the percent multi-family housing units at the COA level was first divided by 100. Then, we multiplied it by the total number of housing units to calculate the total number of multi-family housing units in COAs. Then we aggregated this count at the egohood level and divided it by the total number of housing units in egohoods and multiplied it by 100 to convert it back to the percentage measure.

To capture the structural characteristics, we used the Korean Census data in 2018. We include the percent residents with at least a bachelor’s degree to capture the general level of social status. We include the percent homeowners to approximate residential stability. To capture the level of social control or guardianship (Wilcox et al., 2003), we include the percent single-person households. We also include the percent 15 to 29-year-old residents as a proxy for the crime-prone age group in neighborhood. The percentage measures of structural characteristics were originally provided at the COA level. As abovementioned, these percent measures were converted into counts, aggregated, and then computed back to the percent measures at the egohood level.

Analytic Strategy

We estimated a series of models to test the effects of physical environmental factors and structural characteristics on crime rates. Our dependent variables are counts of various types of crime and disorder and their distributions are less likely to be normally distributed. Accordingly, we employ the negative binomial regression approach to effectively address the over-dispersion in the count outcomes (Osgood, 2000). We include the residential population as the exposure variable, which turns the outcomes interpretable as crime rates. The general form of the models is:

E (y) = e x p (α + B_{1} B + B_{2} H + B_{3} S + v)

where, y represents the dependent variable, $α$ is an intercept, B is a matrix of business measures, H is a matrix of housing measures, S is for a matrix of various measures of structural characteristics, and ν has a gamma distribution that captures the overdispersion. Table 1 describes the summary statistics of the dependent and independent variables used in the analyses. We also checked the variation inflation factor (VIF) scores and detected no evidence of multicollinearity as the VIF scores for all models did not exceed a value of 5.

Table 1.

Summary Statistics.

	M	SD	Minimum	Maximum
Crime
Violent	55.29	58.44	0	1,169
Property	30.49	22.72	0	284
Sex crime	7.22	8.24	0	131
Disorder	392.19	289.63	0	2,212
Number of businesses
Retail	132.76	176.74	0	2,984
Accommodation and food	98.09	73.16	0	568
Public administration	0.16	1.04	0	10
Education	33.26	38.14	0	259
Health and social work	14.98	16.29	0	130
Arts, sports, and recreation	13.56	16.80	0	133
Repair and other personal services	59.19	37.32	0	253
Housing
% Multi-family units	6.59	9.31	0	71.88
% Single-family units	19.28	21.75	0	100
% High-rise APT	70.35	29.75	0	100
% Small apartments	1.12	2.47	0	66.67
% Commercial-residential mixed units	0.93	1.18	0	22.08
Average housing age	20.93	6.43	0.85	44
Structural characteristics
% Person with at least a BA degree	24.27	9.63	0	100
% Homeowners	59.23	16.87	0	100
% Single person household	27.47	11.91	3.26	76.60
% Aged between 15 and 29	17.95	4.09	7.95	73.86

To test moderating effects, we estimated a set of models including the interaction terms into the models, respectively. Specifically, we chose a measure for each concept that can properly represent physical environment or structural characteristics in Korean context: The number of accommodation and food businesses (business facilities), the percentage of high-rise apartments (housing type), and the percentage of single person households (structural characteristics). We selected the number of accommodation and food businesses because they are identified to be more consumer-facing businesses in previous studies (Kane et al., 2016; Kim & Hipp, 2021) and thus have more direct relevance to criminal opportunities from the number and type of foot traffic visiting the area. We chose the percent high-rise apartments for the housing type because high-rise apartments are the most common housing type in Korea as more than half of Korean population live in high-rise apartments (KOSIS, 2020). Moreover, living in high-rise apartments reflects socioeconomic status in Korea. For example, in 2020, about 78% of people in the upper 20% of annual household income bracket lived in high-rise apartments, while only 32% of people in the lower 40% of annual household income did. Finally, we selected the percent single-person households given that rapid increase of single-person households is one of the most important social structural changes happening in Korea. About 9% of the total households were single-person occupied in 1990; but it went up to 24% in 2010 and 32% in 2020 (KOSIS, 2020). Then, we created multiplicative interaction terms for each pair of the three measures and included them to the models, respectively: % high-rise apartments X % single person households, % high-rise apartments X the number of accommodation and food businesses, or % single person households X the number of accommodation and food businesses.

Results

We first check the spatial concentrations of crime to see whether the law of crime concentration still holds in Daegu, South Korea. Our findings suggest that crime spatially concentrates within the city. For instance, about 50% and 45% of violent and property crimes spatially concentrate in only 5% of all areas in the city of Daegu, respectively. Likewise, only 5% of all areas produce 47% and 55% of all disorders and sex crimes. These findings are consistent with previous studies and confirm that crime concentration pattern is robust in a large Korean city context.

We next check the distribution of crime presented in Figure 2. These are the maps of the study area with COAs colored according to the levels of crime types in which red color represents the areas with higher risk of crime while blue color is for lower risk. Crime counts were normalized by the size of COA for mapping purpose. First, we observe that older downtown areas in the city (the city center) generally have higher risk of crime. For example, East side of Seo-gu (middle to left in the map extent) and Northeast side of Dalseo-gu (bottom-left) are known for older housings with relatively poor security. We also see some spatial concentration of crime in mid-west side of Suseoung-gu (bottom-right), south-east side of Buk-gu (top-center), and northern part of Nam-gu (mid-bottom). These areas are known for “fun places” where many restaurants and pubs are located. Also, we observe that the college towns in the city such as north-west Dalseo-gu, south-east Buk-gu have more crime than other nearby areas as they tend to have more crime-prone age residents and single person households. Finally, we see that the sex crime exhibits different spatial pattern that the city center areas with relatively more aged population show lower risk for sex crime (e.g., east Seo-gu).

Main Effects

We next turn to our findings from estimated negative binomial regression models (Table 2). We begin with the coefficients for the business measures in the models. We observe that our business measures have crime-enhancing effects, in general. The number of accommodation and food business exhibits the strongest crime-producing effect for all types of crime and disorder. For instance, a one standard deviation increase in the number of accommodation and food business is associated with about 36% increase in the risk of violent crime (exp (β × SD) – 1). Likewise, there is about 11%, 19%, and 23% higher risk of property, sex crime, and disorder by one standard deviation increase in the number of accommodation and food business, respectively. We see a similar pattern for other types of businesses. For example, a one standard deviation increase in the number of Arts, sports and recreation service business leads to about 14%, 11%, 17%, and 6% increase in violent, property, sex crime, and disorder, respectively. The number of public administration office is positively associated with all types of crime as well. Specifically, a one standard deviation increase in the number of public administration office results in 3%, 7%, 2%, and 5% increase in violent, property, sex crime, and disorder, respectively. Interestingly, some business measures tend to reduce sex crime in place. For instance, a one standard deviation increase in the number of retail businesses (and health and social work businesses) results in about 3% (6%) reduction in sex crime.

Table 2.

Negative Binomial Models for Various Types of Crime in ¼ Mile Egohoods.

	Violent	Property	Sex crime	Disorder
Retail	0.000	0.000**	0.000**	0.000
Retail	0.000	0.000	0.000	0.000
Accommodation and food	0.004**	0.001**	0.002**	0.003**
Accommodation and food	0.000	0.000	0.000	0.000
Public administration	0.026**	0.063**	0.022*	0.047**
Public administration	0.008	0.006	0.010	0.007
Education	0.000	0.001**	0.000	0.001**
Education	0.000	0.000	0.000	0.000
Health and social work	0.001	0.003**	−0.004**	0.002**
Health and social work	0.001	0.001	0.001	0.001
Arts, sports, and recreation	0.008**	0.006**	0.009**	0.004**
Arts, sports, and recreation	0.001	0.001	0.001	0.001
Repair and other personal services	0.000	0.001^†	0.003**	0.001**
Repair and other personal services	0.000	0.000	0.000	0.000
% Multi-family units	−0.006*	−0.009**	−0.009**	−0.008**
% Multi-family units	0.003	0.002	0.003	0.002
% Single-family units	−0.011**	−0.013**	−0.009**	−0.013**
% Single-family units	0.002	0.002	0.003	0.002
% High-rise APT	−0.015**	−0.017**	−0.013**	−0.016**
% High-rise APT	0.002	0.002	0.003	0.002
% Small apartments	0.006	0.012**	0.038**	0.005
% Small apartments	0.004	0.004	0.006	0.004
% Comm-res mixed units	0.088**	0.119**	0.027*	0.118**
% Comm-res mixed units	0.011	0.009	0.014	0.009
Average housing age	0.008**	0.003*	−0.005*	0.004**
Average housing age	0.002	0.002	0.002	0.002
% Person with at least a BA degree	−0.005**	−0.003**	−0.010**	−0.007**
% Person with at least a BA degree	0.001	0.001	0.001	0.001
% Homeowners	−0.008**	−0.007**	−0.010**	−0.007**
% Homeowners	0.001	0.001	0.001	0.001
% Single person household	0.020**	0.014**	0.028**	0.013**
% Single person household	0.001	0.001	0.002	0.001
% Aged between 15 and 29	−0.004^†	−0.006**	0.028**	−0.004*
% Aged between 15 and 29	0.002	0.002	0.003	0.002
Intercept	−4.598	−4.643	−7.448	−2.148
Intercept	0.272	0.211	0.329	0.235
N	4,569	4,569	4,569	4,569

Note. Standard errors below coefficient estimates.

p < .05 (two-tailed test). **p < .01 (two-tailed test). ^†p < .05 (one-tailed test).

Next, we turn to our findings of housing measures. First, we begin by focusing on the effects for our measures of various housing types. We find that the percentage of multi-family housing units, percentage of single-family housing units, and percentage of high-rise apartments are negatively associated with all types of crime and disorder. A one standard deviation increase in the percent multi-family units in egohoods decreases 5% to 8% more violent, property, sex crime, and disorder, respectively. Likewise, a one standard deviation increase in the percent single-family units is associated with about 17% to 24% decrease in crime and disorder. Among all housing type measures, the percent high-rise apartments exhibits the strongest effect in magnitude for all types of crime. For instance, a one standard deviation increase in the percent high-rise apartments results in about 40% to 42% reduction in the risk of violent, property, sex crime, and disorder, respectively. In contrast, we observe that areas with more small apartments are at higher risk of property and sex crime given that there will be 3% and 10% increase in property and sex crime associated with a one standard deviation increase in the percent small apartments, respectively. Our measure of mixed land use shows a crime-enhancing effect, which is consistent with previous studies on land use mix and crime (Wo, 2019a, 2019b; Wo & Kim, 2022). For instance, a one standard deviation increase in percent mixed land use units is associated with about 8% to 10% increase in the risk of violent, property, sex crime, and disorder, respectively. For our average housing age measure, we detect a positive relationship between this measure and crime, except for sex crime. For example, a one standard deviation increase in the average housing age is associated with about 5%, 2%, and 3% increase in violent, property crime, and disorder, respectively. In contrast, the average housing age tends to reduce sex crime as a one standard deviation increase in average housing age entails about 3% reduction in the risk of sex crime.

Next, our findings of the structural characteristics are consistent with previous studies conducted in the US context. First, we observe that the percent person with at least a BA degree is negatively associated with crime and disorder. For instance, a one standard deviation increase in the percent person with at least a BA degree results in about 3% to 9% decrease in the risk of violent, property, sex crime, and disorder, respectively. The percent homeowners in egohoods shows a similar pattern. That is, a one standard deviation increase in the percent homeowners is associated with about 11% to 16% reduction in violent, property, sex crime, and disorder, respectively. We find that single person households in egohoods has a robust crime-enhancing effect for all types of crime and disorder. For instance, a one standard deviation (SD) increase in the percent single person households results in about 27% increase in the risk of violent crime. Likewise, a one standard deviation increase in the percent single person households enhances the risk of property crime, sex crime, and disorder about 17%, 39%, and 17%, respectively. Our findings of the percent person aged 15 to 29 are mixed. Whereas the percent aged 15 to 29 tends to decrease property crime and disorder, there is a crime enhancing effect for sex crime.

Moderating Effects

Given the robust findings of the main effects, we next assess whether the associations between the measures of structural characteristics, housing, and business establishments have moderating effects for each other. To do so, we include the interaction terms to the models: % high-rise apartments X % single person households, % high-rise apartments X the number of accommodation and food businesses, and % single person households X the number of accommodation and food businesses. We test these interactions one at a time by adding each interaction term in the models, respectively. We report the interaction coefficients in Table 3. Additionally, we plotted the predicted crime rates for these interactions in Figure 3. Specifically, we visually displayed the effect of each measure at varying levels of another measure (Low = −1 SD and High = +1 SD). The interaction plots show that the patterns were generally similar. Therefore, we select the findings for property crime as a representative of the plots for all crime types.³

Table 3.

Interaction Models.

	Violent			Property
	High-rise APT and single person HH	High-rise APT and accommodation and food	Single person HH and accommodation and food	High-rise APT and single person HH	High-rise APT and accommodation and food	Single person HH and accommodation and food
% Single person HH	−0.008**		0.020**	−0.009**		0.013**
% Single person HH	0.003		0.001	0.002		0.001
% Percent high-rise APT	−0.031**	−0.020**		−0.030	−0.020**
% Percent high-rise APT	0.003	0.002		0.002	0.002
# Accomodation and food		0.001**	−0.000**		0.001**	−0.000
# Accomodation and food		0.000	0.000		0.000	0.000
Interaction term	0.000**	−0.000**	0.000**	0.000**	−0.000**	0.000**
Interaction term	0.000	0.000	0.000	0.000	0.000	0.000
Intercept	−3.476**	−4.291**	−4.521**	−3.704**	−4.430**	−4.497**
Intercept	0.287	0.276	0.272	0.222	0.215	0.210
N	4,569	4,569	4,569	4,569	4,569	4,569

Note. Standard errors below coefficient estimates.

All other controls were included but not shown.

p < .05 (two-tailed test). **p < .01 (two-tailed test).

	Sex crime			Disorder
	Single person HH and high-rise APT	High-rise APT and accommodation and food	Single person HH and accommodation and food	Single person HH and high-rise APT	High-rise APT and accommodation and food	Single person HH and accommodation and food
% Single person HH	−0.015**		0.027**	−0.014**		0.013**
% Single person HH	0.004		0.002	0.003		0.001
% Percent high-rise APT	−0.039**	−0.018**		−0.031**	−0.020**
% Percent high-rise APT	0.003	0.003		0.002	0.002
# Accomodation and food		0.001**	−0.001**		0.001**	−0.000**
# Accomodation and food		0.000	0.003		0.000	0.000
Interaction term	0.001**	0.000**	0.000**	0.000**	−0.000**	0.000**
Interaction term	0.000	0.000	0.000	0.000	0.000	0.000
Intercept	−5.677**	−7.117**	−7.233**	−1.114**	−1.870**	−2.066**
Intercept	0.331	0.332	0.329	0.244	0.240	0.234
N	4,569	4,569	4,569	4,569	4,569	4,569

Note. Standard errors below coefficient estimates. All other controls were included but not shown.

p < .05 (two-tailed test). **p < .01 (two-tailed test).

As shown in Figure 3a, areas with more high-rise apartments are at lower risk of crime and such crime-reducing pattern is stronger in areas with fewer single person households. Therefore, areas with high high-rise apartments combined with low single person households are at the lowest risk of property crime. For example, for two areas with higher proportion of high-rise apartments, the one with low single person households has about 54% fewer property crime than does one with high single person households. In the models looking at % high-rise apartments X the number of accommodation and food businesses, although areas with more high-rise apartments have fewer property crime, there is little evidence of an interaction effect with the number of accommodation and food businesses; the pattern is plotted in Figure 3b. For the interaction of % single person households X the number of accommodation and food businesses, we observe a pronounced moderation effect. Figure 3c demonstrates that the areas with more single person households in combination with more accommodation and food businesses are at highest risk of property crime. For example, for two areas with high percent single person households, the one with low accommodation and food businesses has about 10% fewer property crime than does one with high accommodation and food businesses.

Discussion

Building on the theories of criminal opportunities and social disorganization, we examined how physical and structural characteristics of areas in the city of Daegu, South Korea are associated with different types of crime. In particular, we tested whether the theoretical propositions developed in the Western countries could be applied to Korean context by using a sample of Daegu, which is one of the largest cities in South Korea. We included measures of businesses and housings to capture physical environment, while accounting for the effects of structural characteristics. Our descriptive results suggested that about 50% of all types of crime were spatially concentrated in about 5% of areas within the city. This is consistent with previous studies that crime is not uniformly distributed over the space but rather spatially concentrated, which confirms the generalizability of the spatial concentration pattern of crime (Weisburd, 2015).

Next, we observed that various types of business facilities generally have crime-producing effects, consistent with previous studies. Areas with more nonresidential business activities are expected to have more criminal opportunities because such areas tend to have more foot traffic that increases the probability of convergence of potential offenders and targets at the same time and place. Moreover, in areas with high density of foot traffic, it is common for people to be unfamiliar one another (Brantingham & Brantingham, 1995; Sherman et al., 1989), which impairs guardianship capability to detect potential offenders (Reynald, 2010). Interestingly, we found that retail businesses and health and social work businesses have crime-reducing effects for sex crime. One possible explanation is that, in Korean cities, retail and health-related businesses often locate in areas where more aged people with higher length of residence tend to reside, which may result in enhanced territoriality and guardianship capability. For the same reason, we can expect reduced level of sex crime given that sex crime has been characterized as a type of young adults’ crime (Farrington, 1986).

Next, we found that different types of housing matter for understanding the spatial patterns of crime, consistent with prior studies. Specifically, multi-family units, single-family units, and high-rise apartments were negatively associated with crime. These findings are consistent with previous studies that residential land use generally exhibits a crime-reducing effect (Boessen & Hipp, 2015). In terms of the effect size, the high-rise apartments show the strongest crime-reducing effect. In South Korea, high-rise apartment complexes are mostly gated communities with the private features within the gates or walls as well as the internal governance and management of the community. They are residential areas with restricted access for residents and their guests with designated perimeters (fences or walls). Such physical and social features enhance the separation between the private and public space—territoriality (Newman, 1972), and thus reduce the ambiguity who is responsible for the surveillance of the area. Moreover, private security guards are at the controlled gates, and the areas are monitored by CCTVs, which may promote guardianship capability in high-rise apartment complexes and surrounding areas.

In contrast, we found that the proportion of small apartments (i.e., Korean style “one-room” studio apartments) was positively associated with crime. In our study area context, “one-room” studio apartments have been massively constructed for single female college students and workers since 1993 (Choo et al., 2014). The structural vulnerability of one-room apartments has been well documented in previous Korean studies that they have insufficient security and easiness of intrusion (An, 2018; Jung & Lee, 2018; Hong & Lee, 2020; Hwang et al., 2013). We suspect that such features of one-room apartments provide more criminal opportunities that potential offenders may perceive the areas more favorable for committing crime as they see higher target suitability and lack of guardianship capability. We also observed that commercial-residential mixed units have crime-producing effect. Given that mixed-use implies a diversity of activities, areas with a high concentration of mixed-use buildings should enhance the amount of pedestrian foot traffic including residents and nonresidents while being unfamiliar each other. This may impairs social cohesion in the area (Sampson & Raudenbush, 1999; Taylor, 1988; Weisburd et al., 2012) and undermines the ability to detect suspicious activities including crime and disorder (Reynald, 2010), thereby increasing crime.

Our findings of structural characteristics are consistent with our expectations. The percentage of people who have at least a bachelor’s degree and the percentage of homeowners were negatively associated with all types of crime and disorder. Particularly, the percentage of single person household was positively associated with all types of crime and disorder. A large proportion of single person household includes college students or workers who temporarily stay in the areas for attending schools and part time jobs. Therefore, it is possible that areas with more single person households may have high residential turnover. Indeed, Jung and Lee (2018) found that more than 80% of single person household members in their sample reside less than 2 years in their current residences. Also, only 32.6% of single person household members own a house. Such conditions may collectively weaken social ties, cohesion, and thus informal social control in neighborhood, and consequently increase crime.

Our findings for the interaction effects between the measures of physical and social environments suggest an importance of simultaneously accounting for them for understanding the spatial patterns of crime. For example, the crime-reducing effect of the percentage of high-rise apartment was amplified in areas with lower level of single person household. Whereas more high-rise apartments are consistently associated with fewer crime, this effect was accentuated when combined with low single person households in place. Also, the number of accommodation and food businesses moderates the crime-enhancing effect of the percentage of single person household to the direction that amplified the crime rates. Although the proportion of single person household enhances the level of crime, this effect might be strengthened when combined with a sufficient level of potential foot traffic from business activities. These findings highlight the need for further research to examine how housing type, business activities and structural characteristics affect the criminal opportunities and guardianship in place; yet the results of the moderating effects confirmed the importance of examining different physical and social environmental features at the same time and how they collectively work together to produce different spatial patterns of crime.

In addition to these theoretical implications, we see that our findings have some practical implications. First, risky facilities (e.g., entertainment areas) and vulnerable residential housings (e.g., one-room apartments) may require more monitoring effort with proper security methods such as streetlights (Welsh & Farrington, 2008, 2009) or CCTVs (Piza, 2018; Yeom, 2019a) to reduce crime. Such crime prevention tactics can be employed to reduce crime in high-crime areas identified in the current study. Also, given the empirical evidence (including the current study) that structural characteristics were associated with crime in Korean context, practitioners should find ways to enhance social cohesion, informal social control, and collective efficacy among residents. For example, community policing accompanied by foot patrol and community meetings may be considered to increase informal social control (Lombardo & Donner, 2018).

As with all research, the current study is not without limitations. First, although we provided possible explanations for the results, we were not able to test a specific mechanism such as social ties, foot traffic, informal social control, and guardianship capability in place because our dataset does not include survey data. Therefore, a natural extension for future work would be studying the mechanism employing survey data or a qualitative research method. Another limitation is that the current study was cross-sectionally designed. One challenge with examining the effects of ecological measures on neighborhood crime is a possible temporal endogeneity where levels of crime in previous time points may shape the current social and physical environments (Hipp, 2010; Hipp et al., 2019). For instance, Hipp et al. (2019) suggest that higher violent and property crime in neighborhoods are associated with both business failure and relocation. Therefore, we recommend that future work to employ a longitudinal analysis to account for the possible changes of structural characteristics and physical environment over time. Finally, there is always concern of under-reporting of crime when using official crime data reported to police agencies. The 112 crime calls data employed in the current study are not confirmed crime cases but initial reports from victims and/or witnesses. Such nature may bring about potential bias in the data if there is systematic under-reporting. Although one study found less evidence of systematic bias when reporting the serious types of crime (Baumer, 2002), the concern remains. Also, we were not able to include socioeconomic variables (i.e., poverty or unemployment rate) due to data availability. Given that they are the key factors of structural characteristics, future research should consider including them and examine whether they have distinct effects on neighborhood crime in Korean context.

In conclusion, the principal objective of the current study was to understand how physical environment and structural characteristics may shape the spatial patterns of crime in a large Korean city context and to check the generalizability of the theoretical framework of criminal opportunities and social disorganization. Our results suggest that various types of businesses generally exhibit a crime-producing pattern, although retail businesses have crime-reducing effect for sex crime. We also found that housing types are important for understanding the spatial patterns of crime. Our findings of structural characteristics were consistent with our expectations and previous studies conducted in the US context. Furthermore, our interaction results highlighted that it is necessary to examine the different types of physical environment and structural characteristics simultaneously as suggested in previous studies (Kim & Hipp, 2021, 2022a,b). These results indicated that incorporating the physical and social environmental features based on the theoretical framework of criminal opportunities and social disorganization can be useful for understanding the spatial patterns of crime in Korean context. We encourage future studies to examine whether place-based explanations for crime is applicable in other regions of the world as it is important to “make further progress in understanding the generalizability of the link between community social mechanisms and crime rates” (Sampson, 2006, p. 162).

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Young-An Kim

Joonggon Kim

Notes

Author Biographies

Young-An Kim is an Assistant Professor in the College of Criminology and Criminal Justice at Florida State University. His research interests focus on various areas such as neighborhoods and crime, criminology of place, immigration and crime, and geo-spatial analysis. Besides criminology, he is interested in urban sociology and quantitative research methods.

Joonggon Kim is an Assistant Professor in the Department of Police Administration at Keimyung University. His research focuses on social reaction to delinquent behaviors, developmental change of criminality, and hate crime.

References

(2018). Study on the fear of crime and the actual condition investigation of residential environmental of a single person’s households in the university town—focused on a university town in a local small-sized city. Journal of the Korean Housing Association, 29(4), 63–71.

Baumer

E. P.

(2002). Neighborhood disadvantage and police notification by victims of violence. Criminology, 40(3), 579–616.

Bernasco

Block

(2011). Robberies in Chicago: A block-level analysis of the influence of crime generators, crime attractors, and offender anchor points. Journal of Research in Crime and Delinquency, 48(1), 33–57.

Bernasco

Ruiter

Block

(2016). Do street robbery location choices vary over time of day or day of week? A test in Chicago. Journal of Research in Crime and Delinquency, 54(2), 244–275.

Boessen

Hipp

J. R.

(2015). Close-ups and the scale of ecology: Land use and the geography of social context and crime. Criminology, 53(3), 399–426.

Brantingham

(1995). Criminality of place: Crime generators and crime attractors. European Journal on Criminal Policy and Research, 3(3), 1–26.

Brantingham

P. L.

Brantingham

P. J.

(1993). Nodes, paths and edges: Considerations on the complexity of crime and the physical environment. Journal of Environmental Psychology, 13, 3–28.

Cheong

J. S.

Kang

(2013). Causal relationship between structural characteristics of metropolitan neighborhoods and homicide. The Journal of the Korea Contents Association, 13(3), 152–161.

Cheong

J. S.

Kwak

D. H.

(2008). A study on the effects of community conditions on crime rates: Longitudinal analyses using the growth curve model. Criminological Review, 19(3), 251–290.

10.

Cheong

J. S.

Park

H. H.

(2010). A study on the effects of structural covariates on homicide: Nationwide analysis using negative binomial regression model. Korean Criminological Review, 21(1), 91–119.

11.

Cho

Lee

Y. H.

Harper

S. B.

(2021). Testing the systemic model of social disorganization theory in South Korean neighborhoods: A latent class growth analysis approach to specifying pathways to homicide. Homicide Studies, 25(2), 139–163.

12.

Choo

Jeong

Lee

K. H.

(2014). Feasibility analysis of a studio-type apartment investment: Focused on Daegu region. Korea Real Estate Academy Review, 59, 164–172.

13.

Cohen

L. E.

Felson

(1979). Social change and crime rate trends: A routine activity approach. American Sociological Review, 44(4), 588–608.

14.

Farrington

D. P.

(1986). Age and crime. In Tonry

Morris

(Eds.), Crime and justice: An annual review of research (pp. 189–250). University of Chicago Press.

15.

Felson

(1987). Routine activities and crime prevention in the developingmetropolis*. Criminology 25(4), 911–932.

16.

Felson

Boba

(2010). Crime and everyday life. SAGE.

17.

Hipp

J. R.

(2007). Block, tract, and levels of aggregation: Neighborhood structure and crime and disorder as a case in point. American Sociological Review, 72(5), 659–680.

18.

Hipp

J. R.

(2010). A dynamic view of neighborhoods: The reciprocal relationship between crime and neighborhood structural characteristics. Social Problems, 57(2), 205–230.

19.

Hipp

J. R.

Boessen

(2013). Egohoods as waves washing across the city: A new measure of “neighborhoods”. Criminology, 51(2), 287–327.

20.

Hipp

J. R.

Kim

Y.-A.

Kane

(2019). The effect of the physical environment on crime rates: Capturing housing age and housing type at varying spatial scales. Crime & Delinquency, 65(11), 1570–1595.

21.

Hong

Lee

(2020). Analysis of vulnerable spots of urban small scale multi-family housing against residential burglary through the analysis of crime investigation records. Journal of Community Safety and Security by Environmental Design, 11(1), 61–110.

22.

Hwang

Kang

Park

(2013). A research study on crime prevention for one-person households by type of houses. Journal of Korean Housing Association, 24(4), 9–17.

23.

Jiang

Land

K. C.

Wang

(2013). Social ties, collective efficacy and perceived neighborhood property crime in Guangzhou, China. Asian Criminology, 8, 207–223.

24.

Jung

Chun

Kim

(2022). Modeling crime density with population dynamics in space and time: An application of assault in Gangnam, South Korea. Crime & Delinquency, 68(2), 253–283.

25.

Jung

Lee

Y. H.

(2018). The impact of the physical status of urban life-type housing and the characteristics of single-person households on the anxiety of crime in the housing: Focused on single households in one-room type house in Seoul urban life-type housing. Korea Institute of Ecological Architecture and Environment, 18(6), 51–60.

26.

Kane

Hipp

J. R.

Kim

J. H.

(2016). Los Angeles employment concentration in the 21st century. Urban Studies, 55(4), 844–869.

27.

Kim

Y.-A.

Hipp

J. R.

(2021). Density, diversity, and design: Three measures of the built environment and the spatial patterns of crime in street segments. Journal of Criminal Justice, 77, 101864.

28.

Kim

Y.-A.

Hipp

J. R.

(2022a). Both sides of the street: Introducing measures of physical and social boundaries based on differences across sides of the street, and consequences for crime. Journal of Quantitative Criminology, 38, 75–103.

29.

Kim

Y.-A.

Hipp

J. R.

(2022b). Small local versus non-local: Examining the relationship between locally owned small businesses and spatial patterns of crime. Justice Quarterly, 39, 983–1008.

30.

Kinney

Brantingham

Wuschke

Kirk

Brantingham

(2008). Crime attractors, generators and detractors: Land use and urban crime opportunities. Built Environment, 34, 62–74.

31.

Korean Statistical Information Service. (2020). Population census 2020. Author.

32.

Kurtz

E. M.

Koons

B. A.

Taylor

R. B.

(1998). Land use, physical deterioration, resident-based control, and calls for service on urban streetblocks. Justice Quarterly, 15(1), 121–149.

33.

Lee

D. S.

Lee

K. H.

(2009). A study on the relationship between crime rates and characteristics of urban areas: Focusing on demographic and socio-economic factors. National Association of Korean Local Government, 26(5), 189–206.

34.

Lee

Jin

Kim

Choi

(2016). A study on characteristics of the spatial distribution of sex crimes: Spatial analysis based on environmental criminology. Journal of the Korean Geographical Society, 51(6), 853–871.

35.

Liu

Song

Xiu

(2016). Spatial patterns of violent crimes and neighborhood characteristics in Changchun, China. Australian & New Zealand Journal of Criminology, 49(1), 53–72.

36.

Lombardo

R. M.

Donner

C. M.

(2018). Can community policing increase residents’ informal social control? Testing the impact of the Chicago alternative policing strategy. Police Practice and Research, 19(5), 427–442.

37.

Morenoff

J. D.

Sampson

R. J.

Raudenbush

S. W.

(2001). Neighborhood inequality, collective efficacy, and the spatial dynamics of urban violence. Criminology, 39(3), 517–559.

38.

Newman

(1972). Defensible space (p. 264). New York: Macmillan.

39.

Osgood

D. W.

(2000). Poisson-based regression analysis of aggregate crime rates. Journal of Quantitative Criminology, 16(1), 21–43.

40.

Perkins

D. D.

Meeks

J. W.

Taylor

R. B.

(1992). The physical environment of street blocks and resident perceptions of crime and disorder: Implications for theory and measurement. Journal of Environmental Psychology, 12(1), 21–34.

41.

Piza

E. L.

(2018). The crime prevention effect of CCTV in public places: A propensity score analysis. Journal of Crime and Justice, 41(1), 14–30.

42.

Ratcliffe

J. H.

(2004). Geocoding crime and a first estimate of a minimum acceptable hit rate. International Journal of Geographical Information Science, 18(1), 61–72.

43.

Reynald

D. M.

(2010). Guardians on guardianship: Factors affecting the willingness to supervise, the ability to detect potential offenders, and the willingness to intervene. Journal of Research in Crime and Delinquency, 47(3), 358–390.

44.

Roh

Kim

Yun

(2010). Criminal victimization in South Korea: A multilevel approach. Journal of Criminal Justice, 38(3), 301–310.

45.

Rosenthal

S. S.

(2014). Are private markets and filtering a viable source of low-income housing? estimates from a “repeat income” model. American Economic Review, 104(2), 687–706.

46.

Sampson

R. J.

(ed.). (2006). Collective efficacy theory: Lessons learned and directions for future inquiry. In Cullen

F. T.

Wright

J. P.

Blevins

K. R.

(Eds.), Taking stock: The status of criminological theory (pp. 149–167). Transaction Publishers.

47.

Sampson

R. J.

Groves

(1989). Community structure and crime: Testing social-disorganization theory. American Journal of Sociology, 94(4), 774–802.

48.

Sampson

R. J.

Raudenbush

S. W.

(1999). Systematic social observation of public spaces: A new look at disorder in urban neighborhoods. American Journal of Sociology, 105(3), 603–651.

49.

Sampson

R. J.

Raudenbush

S. W.

Earls

(1997). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277(5328), 918–924.

50.

Sherman

Gartin

Buerger

(1989). Hot spots of predatory crime: Routine activities and the criminology of place. Criminology, 27(1), 27–56.

51.

Shin

S. H.

(2019). Spatial features affecting the occurrence of hidden camera crime. The Journal of Police Science, 19(1), 59–83.

52.

Skogan

W. G.

(1990). Disorder and decline: Crime and the spiral of decay in American neighborhoods. Free Press.

53.

Skogan

W. G.

(2015). Disorder and decline: The state of research. Journal of Research in Crime and Delinquency, 52(4), 464–485.

54.

Stucky

T. D.

Ottensmann

J. R.

(2009). Land use and violent crime. Criminology, 47, 1223–1264.

55.

Taylor

R. B.

(1988). Human territorial functioning: An empirical, evolutionary perspective on individual and small group territorial cognitions, behaviors, and consequences. Cambridge University Press.

56.

Taylor

R. B.

(1997). Social order and disorder of street blocks and neighborhoods: Ecology, microecology, and the systemic model of social disorganization. Journal of Research in Crime and Delinquency, 34(1), 113–155.

57.

Taylor

R. B.

Koons

B. A.

Kurtz

E. M.

Greene

J. R.

Perkins

D. D.

(1995). Street blocks with more nonresidential land use have more physical deterioration: Evidence from Baltimore and Philadelphia. Urban Affairs Review, 31(1), 120–136.

58.

Tillyer

M. S.

Wilcox

Walter

R. J.

(2021). Crime generators in context: Examining ‘place in neighborhood’ propositions. Journal of Quantitative Criminology, 37, 517–546.

59.

Weisburd

(2015). The law of crime concentration and the criminology of place. Criminology, 53(2), 133–157.

60.

Weisburd

Groff

E. R.

Yang

S.-M.

(2012). The criminology of place: Street segments and our understanding of the crime problem. Oxford University Press.

61.

Welsh

B. C.

Farrington

D. P.

(2008). Effects of improved street lighting on crime. Campbell Systematic Reviews, 4(1), 1–51.

62.

Welsh

B. C.

Farrington

D. P.

(2009). Public area CCTV and crime prevention: An updated systematic review and meta-analysis. Justice Quarterly, 26(4), 716–745.

63.

Wilcox

Land

K. C.

Hunt

S. A.

(2003). Criminal circumstance: A dynamic multicontextual criminal opportunity theory. Aldine de Gruyter.

64.

Wilcox

Neil

Debra

T. C.

Shayne

(2004). Busy places and broken windows? Toward defining the role of physical structure and process in community crime models. The Sociological Quarterly, 45(2), 185–207.

65.

Wilson

J. Q.

Kelling

(1982). The police and neighborhood safety: Broken windows. Atlantic Monthly, 127, 29–38.

66.

J. C.

(2019a). Mixed land use and neighborhood crime. Social Science Research, 78, 170–186.

67.

J. C.

(2019b). Understanding the differential effects of land uses on crime: An examination across Philadelphia neighborhoods. British Journal of Criminology, 59(6), 1432–1454.

68.

J. C.

Kim

Y.-A.

(2022). Neighborhood effects on crime in San Francisco: An examination of residential, nonresidential, and “mixed” land uses. Deviant Behavior, 43(1), 61–78.

69.

Yeom

(2018). Spatial regression analyses on the relationship between alcohol outlet density and violent crime rates. The Journal of Police Science, 18(3), 109–132.

70.

Yeom

(2019a). Crime-deterring effect of CCTV: Application of a dose-response model for continuous treatment variable. Koran Journal off Criminology, 31(2), 203–233.

71.

Yeom

(2019b). Examination of relationship between abandoned house and theft, violent crime with geographically weighted regresssion model. Korean Criminological Review, 30(4), 139–173.

72.

Yoon

O. K.

(2018). An exploratory study on the relationship between welfare spending and crime. Korean Journal of Correctional Discourse, 12(1), 1–27.