Location : Home > Resource > Paper > Theoretical Deduction
Shan Yong | Mechanisms of crime attraction in high-density urban areas
2024-05-16 [author] Shan Yong preview:

[author]Shan Yong


Mechanisms of crime attraction in high-density urban areas

*Author Shan Yong

Professor, School of Law, Nanjing University

Abstract: With the rapid increase of high-density areas in cities,there has been a high concentration of crimes in these areas,hence an urgent need to clear up the doubts about the criminal attraction mechanism in high-density areas. In response to this problem,this article uses urban density to explain,assess and predict the attractiveness of crimes. Based on the assumption that criminal attraction is definitely related to the density of cities,it employs the correlation analysis method to identify the rule of urban density variables and criminal density variables,uses GIS ( Geographic Information System) -based crime mapping to demonstrate that crime hot points are scattered over high-density areas, applies the stepwise regression method and the geographically weighted regression method to screens urban variables affecting crime occurrence and distribution,builds a criminal attraction model,and reveals criminal attraction mechanisms in urban high-density areas,which consists mechanisms for space attraction,attached attraction,comprehensive attraction,hedge attraction and so on. The research on criminal attraction mechanism,which constitutes the contemporary evolution of the“Chicago Paradigm”,integrates the macro perspective and the micro perspective into the“Crime Field Theory”,outlines the knowledge pedigree of urban criminology,and provides theoretic support for the “forecasting,warning and prevention”of crime. It reforms the approaches to circuitous governance by optimizing governance policies and techniques,renews the ideas of reserve criminal policy through the transformation from the emergency criminal policy to the reserve type of criminal policy,and innovates the models of technical control by realizing the technological revolution marked by information technology and artificial intelligence.

Along with the transfer and concentration of population from underdeveloped areas such as rural areas to cities, the expansion of first-class cities and economically developed towns in various regions of China has been obvious, and high-density urban areas have sprung up. High-density spaces, which are contiguous and highly complex, and whose spatial quality needs to be upgraded, have accommodated a large influx of people and led to the gathering of crime. This situation has exacerbated the vulnerability of urban security and magnified the risks to social and national security. "Those who are wise change according to the times, and those who are knowledgeable make decisions according to the events." Social governance innovation needs to deepen the understanding of the laws of social operation and governance, and crime prediction, early warning and prevention cannot be separated from the exploration of the laws of urban operation and the internal mechanism of "urban crime attraction".

Defining crime attraction by urban density

Since Robert Parker of the Chicago School of Criminology put forward the thesis that "crime is a problem of the city", "urban attraction to crime" has become a classic topic in criminology. Scholars have examined the impact of urban microenvironment, community characteristics, human activities, psychological cognition and other factors on crime. However, the impact of high-density urban space on the occurrence and distribution of crime has received less attention, especially the lack of systematic monographs on the relationship between urban density and crime density. As emphasized in the "urban gravity model", there is a natural link between urban gravity and urban density, which is externalized in the siphoning effect of high-density urban areas on low-density areas. As the underlying law that affects the development of high urban density, urban gravity also includes urban crime gravity. Given that urban criminology emphasizes the spatial pattern behind the phenomenon of crime agglomeration rather than the phenomenon itself, exploring the crime gravity of high-density urban areas cannot be separated from the research perspective of urban criminology, and cannot be separated from thinking about urban density and spatial structure.

Urban density refers to the number or proportion of various urban variables in the surface space per unit area. Together, urban density, layout and morphology constitute urban spatial structure. The city is an organism made up of a variety of land use modes, residential forms, architectural forms, facility layouts, road networks and other elements, mosaic combinations in space. Various urban variables are combined into a specific spatial structure with a certain proportion and different mosaic methods, and different spatial structures shape different daily activities and generate very different crime opportunities. The physical environment conditions and spatial characteristics of the city affect the occurrence of crime; criminal hunting behavior, microscopic high-risk environment, residents' daily activities, various community characteristics, and even criminal governance activities are inseparable from the overall constraints, regular influence and basic limitations of the urban spatial mosaic structure. It can be said that urban spatial structure is the field variable that restricts the gravitational force of crime and the filter that affects the choice of crime. It is necessary and feasible to explore the gravitational force of crime from the perspective of "urban density-space structure":

First of all, the widespread emergence of high-density urban areas provides sufficient practical reasons for defining the gravitational pull of crime in terms of urban density. Along with the wave of urbanization, urban high-density areas with continuous distribution, compact structure, complex functions and high complexity have been formed. In these areas, there is a high degree of mixing of various land use types, high population density and mobility, active economic and social activities, a high concentration of buildings, roads and commercial outlets, and an intermingling of newly constructed commercial buildings, old open buildings, and farmers' houses in the middle of the city and on the edge of the city. This situation increases the vulnerability of urban security, magnifies the risk of social security, and causes cities to act as magnetic fields that attract a large number of crimes. In this regard, urban density can be defined as the gravitational force of crime, converting the grasp of high-density spatial mosaic structure into the spatial measurement of the density of different urban variables, analyzing the correlation between urban density and crime density, responding to the complexity of the social security risk in high-density areas of the city, and providing theoretical support for the updating and upgrading of crime management.

Secondly, respecting and following the law of urban development is the policy basis for defining the gravitational pull of crime by urban density and adopting governance measures.The concept of "compact city" was first mentioned at the Central Urban Work Conference in 2015, and the concepts of high-density, diversification, full coverage of the road network, mixed use of land, and systematic planning constitute the basic principles of urban spatial growth. Principles. In order to guarantee the high-density and safe development of cities, crime management should follow the law of urban development, and the relationship between cities and crime should not be viewed in isolation from the characteristics of spatial structure.

Finally, crime mapping techniques based on geographic information systems (GIS) provide scientific methods for analyzing the correlation between urban density and crime density and for measuring the gravitational pull of crime; GIS constitutes a basic tool for processing spatial data on crime and producing crime maps.

Crime mapping is the most direct field of application for thinking about crime from the spatial dimension, and along with the continuous improvement of GIS technology, its field of application is gradually developing towards complexity and integration. In crime mapping, crime hotspots are nothing but high value areas of crime density, and keywords such as crime density and city density constitute new coordinates for exploring crime attraction.

Analyzing ideas of crime attraction mechanism

1. Theoretical organization of "city attraction to crime"

In terms of analytical framework, existing studies have grasped the attraction of crime with probabilistic thinking, based on the correlation analysis of urban variables and crime, and examined the influence of micro-environment, commercial network, land use, road network, community and demographic characteristics and other factors on crime. For example, there is a significant positive correlation between urban public space and juvenile delinquency; changes in the travel patterns of juvenile offenders are explained from the perspective of urban architecture, land use, and road networks; community characteristics (unemployment, income inequality) have a significant impact on property crime; population density, total migration rate, and burglary are significantly and positively correlated; a mix of commercial and residential land uses increases burglary risk, and higher rates of burglary and Lower income levels, closer proximity to urban areas, and higher apartment plot ratios; spatially dilapidated locations in commercial districts and nodes of commercial activity are prone to automobile theft; and in the context of substantial crime reductions in Europe and the United States, commercial venues remain a strong attraction to crime. The above correlation analysis lays down the research framework of "city and crime" and provides reference for the screening of urban variables.

In terms of analytical methods, regarding descriptive statistics, Itoji's analysis of crime statistics and environment of various spaces in Japanese cities is a classic, which has had a profound impact on criminological research in China. For regression analysis, crime is generally set as the dependent variable, urban factors are set as independent variables, and regression models are selected according to the research purpose. For example, the influence of population and land use characteristics on the spatial distribution of violent crime is analyzed by multiple linear regression model; the influence of population density on crime is examined by spatial regression model; the correlation between road network and burglary is measured by hierarchical linear model; the spatial stability of hot and cold spots of crime is verified by grouped trajectory model; and the relationship between the places of daily activities of residents at different time periods and pickpocketing is grasped by spatial negative binomial regression model. The modeling is based on a spatial negative binomial regression model. For crime mapping, detecting crime hotspots with crime maps is a popular research method in the international arena, and scholars in China have made follow-up studies. For example, different economic function areas show different types of crime and crime concentration, the central business district, wholesale markets, etc. are prone to pickpocketing, smashing and breaking into cars, handbags and other cases; pickpocketing cases occur in commercial streets; bars and provocative cases have a strong correlation.

In terms of theory development, established research has promoted the extension of theories of daily activities to the micro-spatial level. Scholars have previously examined the correlation between daily activity variables and crime at both the macro and micro levels. Macro approaches mostly measure proxy variables such as macroeconomic and social structures, making it difficult to directly measure daily activities and only indirectly validate daily activity theories. Micro-analysis focuses on daily activities at the individual level, but does not fully address the spatial characteristics of daily activities and ignores daily activities at the "place level". Sherman et al. propose a "micro-spatial level" between the macro and micro levels, aiming at highlighting the spatial aggregation of crime and exploring the variables of daily activities affecting crime in hot spots. The extension of the theory of daily activities to the micro-spatial level shows that spatial measures have a broad application prospect in the analysis of the gravitational pull of crime.

In terms of coping strategies, based on the above research, the theoretical and practical sectors have proposed environmental prevention programs such as space defense, crime prevention through environmental design (CPTED), business district improvement program (BID), and situational prevention, which have achieved good social results. In the commercial districts of Philadelphia, United States, where the BID program was implemented, the property crime rate was significantly lower than in commercial districts where the program was not implemented. The success of environmental prevention coping strategies signals that the study of crime attraction mechanisms has significant application value.

2.Problems with established research

2.1 Insufficient localized exploration embedded in high-density urban areas

Existing domestic research has focused on the review of foreign criminological theories, with a strong color of theoretical discourse, empirical description, speculative analysis, and expression of opinions, and relatively lagging behind in the exploration of theories originating from empirical analyses. In the past, most of the relevant discussions relied on foreign cities, and there were few empirical studies based on high-density urban areas in China. In fact, the spatial structure of Chinese and foreign cities differs greatly, and the overall number, spatial scale, population density, degree of agglomeration, and speed of formation of high-density areas in domestic cities are much larger than those in Western countries. The standard for built-up areas in the United States is that an area with a core area of 386 people/square kilometer and a total population of 50,000 is an urbanized area. And the density of the built-up area of Beijing in 2015 was 23,800 (10,000 people/square kilometer). It can be seen that if we detach from the systematic empirical analysis of domestic urban areas and ignore the attention to the spatial structure characteristics of high-density urban areas, we can not rationally explore the general law of urban crime in China, and there is no way to provide theoretical support for three-dimensional social security prevention and control.

2.2 The screening of urban variables needs to be optimized urgently.

Under the quantitative social science paradigm, the establishment of the model to explain the attraction of crime cannot be separated from the screening of urban variables, and previous studies have certain problems in screening the independent variables such as commerce and population. Despite the fact that commercial attraction to theft has become a common understanding, previous studies have mostly used commercial factors as a holistic variable without further subdividing the commercial independent variables, and have fallen into the "cluster fallacy" of statistical generalization, resulting in a lack of precise elaboration of the specific mechanisms of commercial attraction to theft. The design of the dependent variable does not further categorize burglary crimes (e.g., daytime vs. nighttime burglary, burglary vs. street pickpocketing, etc.), resulting in a lack of clarity about which types of business factors correlate with which types of burglary cases and what specific correlation exists.

Additionally, certain urban variables are in dire need of integration. Previous studies have examined the relationship between population density, road networks and crime, but have not integrated population density with road factors. Roads are the axes of urban development, and population relies on them for mobility. Population density and road network in urban business districts are must-separate relationships that cannot be used separately in modeling. Statistically, population density and road network show strong covariance in regression analysis, and splitting them into two independent variables will seriously affect the scientific nature of statistics; it is more appropriate to integrate the two independent variables into one comprehensive independent variable. In addition, although some studies have paid attention to the specific influence of certain variables (such as bars, small inns and other commercial places or community characteristics), they have not placed a greater number and variety of urban variables in the same model for systematic analysis and comprehensive consideration.

2.3 Large hierarchy of geographical units

Urban geographic units can be categorized into "city-street-community-grid or road section" according to the area from large to small. Most of the studies in China choose "street" as the meso-unit. For example, a study using streets as the geographic unit used multiple regression analysis to detect no significant correlation between the proportion of commercial land and motor vehicle theft and burglary; a study using streets as the geographic unit neglected to explore in depth the correlation between commercial districts and crime, even though crime hotspots were detected through crime mapping. It is possible that the reason why the above studies conclude that there is no correlation between business and burglary or neglect to examine this correlation is because of the large scale of the geographic unit used by the researchers, thus affecting the scientific nature of the statistics.

From the viewpoint of urban spatial structure, the city is formed by the organic combination of various elements, from the local combination to the whole, from the micro to determine the macro. "Social governance is the art of fine ...... delineation of small governance units, according to the respective characteristics of small units, targeted measures, the precise extension of social governance to every corner, so that the new model of small unit governance to support the big pattern of fine governance". According to the crime map, the distribution of crime has obvious spatial heterogeneity, and certain crime hot spots are adjacent to crime cold spots around them. Only from the street meso-scale crime spatial analysis is too loose, should also dive into the grid, road sections and other micro-level exploration of the city's crime gravity.

3. Correlation analysis between urban density and crime density

Based on the foregoing reflection, the study of crime gravity should be rooted in the high-density areas of the city, dive into the urban microscopic, and carry out the correlation analysis between urban density and crime density. In the correlation analysis, crime mapping and regression analysis methods are combined to screen and optimize urban variables, design the crime attraction model, and explore the inner law of attracting crime in urban high-density areas.

First, the logical clue of correlation analysis. In this regard, the logic of "what - why - how" can be followed to grasp the relationship between urban density and crime density. With regard to the "what", it is necessary to verify whether crime hotspots are mainly distributed in high-density urban areas. This is the prerequisite assumption for the study of crime gravity. For "why", it is necessary to explore which urban variables attract crime to what extent, what specific crime attraction mechanisms exist, and whether there is a general crime attraction model. As for "how to do", it is necessary to examine how to utilize crime attraction mechanisms to help innovation in crime prediction, early warning and prevention in high-density urban areas.

Second, variable screening for correlation analysis. The city is a comprehensive organism in which various urban variables are blended together, and the attraction of high-density urban areas to crime is also a comprehensive attraction of a specific spatial mosaic combination pattern. This study takes the grid as the geographic unit, and uses the number of commercial outlets (the overall commercial variable), the number of hotels, the number of food and beverage outlets, the number of stores along the street, the number of leisure and entertainment outlets, the number of community police offices, whether the residence belongs to the old building district, the plot ratio of the district, the weighted value of the public transportation station, whether the district is closed and managed, the number of health care institutions, the number of schools, and the degree of integration of the road as the independent variables for regression analysis. The density value of factors such as theft crime (overall crime variable), burglary, pickpocketing, street shoplifting, theft of property in motor vehicles or bicycles, electric vehicles, daytime theft and nighttime theft in each grid as the independent variable in the regression analysis.

Third, the methodology of correlation analysis was integrated. The correlation analysis between urban density and crime density mainly relies on the integration of the methods of "crime mapping + regression analysis" and "stepwise regression + geographically weighted regression". Based on ArcGIS software, the spatial distribution of crime hotspots is detected to verify whether the hotspots are located in high-density areas of the city; then, regression analysis is used to grasp which urban variables affect the occurrence and distribution of crime to what extent. The integration of crime mapping and regression analysis is the basic approach to correlation analysis, while regression analysis is realized through two steps: stepwise regression and geographically weighted regression. Stepwise regression analysis is dedicated to introducing significant variables one by one and eliminating non-significant variables in order to obtain an optimal regression model. Stepwise regression belongs to the overall regression analysis, which is insufficient to consider the spatial heterogeneity, so it is necessary to further introduce geographically weighted regression as a local regression analysis. Geographically weighted regression analysis is based on exploring spatial non-stationarity, and local coefficient estimation is carried out using sub-sample data, and the parameters of the model can be changed with space. Because the spatial heterogeneity of various urban variables in different grids is fully considered, geographically weighted regression can form a useful complement to stepwise regression analysis.

Validation of crime clustering in high-density urban areas

The exploration of the spatial relationship between urban density and crime density is a process of starting from the crime map from the shallow to the deep, and pulling out the cocoon. The fact that crime is clustered in high-density urban areas signals the possible existence of crime gravity in the city. This is the premise for designing the crime gravity model. For the validation of crime gathering in urban high-density areas, it can be divided into two questions: whether crime hotspots (areas with high values of crime density) are distributed in urban high-density areas, and whether the higher the crime density is, the higher the urban density is as well; and whether there are areas in urban high-density areas with high urban densities and relatively low crime densities.

The researcher chooses S district of H city in Z province as the study area for crime mapping verification and regression analysis.H city is a new first-tier city in China, and S district is the first area, the central city, the center of trade and tourism, and the city with the highest GDP per unit in H city, and also an old city with a dense population, mature municipal construction, and well-developed public services.S district belongs to the typical high-density area of megacities. This study chooses theft crime as the object of study for two reasons: first, the absolute number and proportion of theft cases rank first in both the court of first instance and the public security criminal filing stage, and the governance performance of theft crime and the formation of the inflection point of crime can be said to be closely related. Second, empirically, theft is very closely related to commercial and demographic factors, and this type of crime is susceptible to the urban spatial environment.

From 2009-2015, the total number of burglary offenses in the S District was 1,601. The locations of the crimes are identified on a crime map based on the descriptions in the criminal indictments. The previous study found that crimes showed an aggregated distribution in District S, and crime hotspots had significant stability. Therefore, crime density can be mapped in a hierarchical subdivision, and the two areas with the highest crime density of about 10% of the area from the study area are selected and set up as primary and secondary crime hotspots; these two hotspots together constitute the crime high-density area in S-area (Fig. 1). Crime mapping found that Level I crime hotspots covering 9.7% of the study area covered 56.8% of all crimes; Level II crime hotspots covering 10.6% of the study area covered 23.5% of all crimes; and after combining the two levels of hotspots, crime hotspots covering 20.3% of the study area covered 80.3% of all crimes.

A comparative urban density analysis (Table 1) was conducted to compare Tier 1 crime hotspots with other areas and Tier 1 + Tier 2 crime hotspots with other areas, comparing the density of commercial outlets (pcs/km2 ), population density (assessed value), average distance to the CBD (landmark building) (km), road integration, density of bus stops (pcs/km2 ), weighted value of bus stops, density of community police offices (pcs/ square kilometer), average plot ratio of the neighborhood, and other 8 urban variables (variable setting and preprocessing are detailed later).