According to theoretical and empirical studies on causes of crime, regional or space characteristics have been considered utmost important. This research was based on the assumptions that crime has the nature of environmental opportunity and that it can be prevented from a point of view of risk assessment. The purposes of the current study were to define crime risk by likelihood and consequences of crime occurrence, to develop tools of crime risk assessment at specific region and/or space, to apply these tools to the region and/or space, and to test appropriateness of the tools. And, the study was also to suggest policy alternatives for the crime prevention in each region and space by making more systematic the whole process of diagnosis of crime, development of policy alternatives, and assessment of policy effects and feedback. In 2012, the first of 3-year research program was done. Particularly the goals of the first year research were to develop assessment tool of crime risks, to assess crime risks at low income residential areas, and to develop policy alternatives for crime prevention.
This study was based on theoretical tradition of environmental criminology from the Chicago School and on discussions concerning risk assessment and management in relation to discourse of risk society that has recently drawn much attention.
The works of the Chicago School for which the names of scholars such as have been well known Robert Park, Ernest Burgess, Clifford Shaw and Henry McKay were much popular in the 1930s to the 1940s, and they lost much support since 1940s due to criticism on prepositions and methodologies of the theories. However, in the 1970s when C. Ray Jeffrey's writing of “Crime Prevention through Environmental Design”(1971) and Oscar Newman's writing of “Defensible Space: Crime Prevention through Urban Design”(1972) were published, relation between crimes and environmental characteristics drew much attention again.
Publication of both books spawned a large volume of research in criminology, architectural engineering and urban engineering that investigated effects of the environment upon crimes. The focus of Jeffrey and Newman was to prevent crime from occurring, which has become integual component of environmental criminology literature. The environmental criminology investigates crime phenomenon from a point of view of constantly changing environment. Jeffrey made use of terminology of environmental criminology at his work of “Crime Prevention through Environmental Design” for the first time to emphasize needs of 'new idea' within the criminology. Jeffrey's new idea was to keep principle of classic criminology and to be focused not on each criminal but on environment where crime occurs. Here, the environment is to be conceptualized to include legal system, social system, physical design (architecture) and the built environment such as railway, use of the land and buildings, and etc.. Jeffrey contended that men should be considered as a part of the environment because men comprise a part of the environment and are constantly reacting to it. The theoretical approaches that were said to be within environmental criminology from a broad point of view included theories such as routine activity theory, geometric theory of crime, rational choice theory and pattern theory of crime, and etc..
On the other hand, in either criminology or criminal policy areas, it seems that conceptualization of risk has been made in three ways: First, an individual's likelihood of committing crime had been measured through risks and made into theory. The risk measurement for individuals had much relations with sentencing policy, corrections and probation policy and adopted psychological methodology in many cases. Second, likelihood of crime occurrence at specific environment and space could be quantified by using concept of risks with development of policy alternatives. This risk measurement, once propelled in environmental criminology, has increasing connection with various areas of sciences such as urban engineering, architecture, psychology and others. Third, acceptance and control of risks in police activities have been studied from a theoretical point of view of risk society proposed by Ulich Beck. The three types of approaches could be classified into microscopic level, middle level and macroscopic level.

In this study, weights were given to factors having influence upon likelihood of crime occurrence(such as crime characteristics factors, socio-demographic factors, and physical and situation factors) and factors having influence on consequences and effects of crime(such as fear of crime, tangible and/or intangible expenses of crime), making assessment tool of crime risks. The likelihood of crime occurrence factor consisted of 20 indexes, while the crime consequences and effect factor did 3 indexes. Here, the social disorganization theory, routine activity theory, situational crime prevention theory, collective efficacy theory, crime prevention and CPTED theory, and EN14383 European standard of CPTED were utilized to provide theoretical support for construction of assessment indexes. The criminologists, police science scholars, sociologists, urban engineering scholars, security experts, psychologists and others joined to examine official crime statistics, crime prevention environment designs and other materials, and decided upon indexes through more than 10 times of meetings and discussions.
The research team again looked into items collectable through the Government official statistics, population census, questionnaire survey and observation study, and made efforts to select the most objective and quantitative indexes. The team collected physical characteristics material by on-the-spot examination and some of data, not easy to collect otherwise, by questionnaire survey.
The research team was divided into two in the beginning of the site research to do for one week on-the-spot survey, analysis of maps, interviews with relevant persons, and reviews of building registers and other government office materials to examine appropriateness of indexes and collectibility of the materials. One team found that collection of materials concerning general as well as present situation on the crime possible by visiting the police precinct office. The other team examined demographic characteristics as well as situation and physical characteristics of the areas under study. The teams repeatedly examined indexes that were difficult to collect or to assess, and they decided whether to include or not some of the indexes in the study. Questionnaire survey, on-the-spot observation survey and official statistics were utilized. In this study factors and indexes of crime risk were not assumed to have the same weight but rather relative weight, and AHP (Analytic Hierarchy Process) was used to decide upon weight.
In this study, at the 1st stage, researchers and outside advisory members uated relative importance of major factors such as crime environment characteristics, socio-demographic characteristics, physical and situation characteristics, and consequences and effects of crime. Then they again uated relative importance of indexes within each factor. At the 2nd stage, all the research members got together to review outcomes of the 1st stage and to get rid of indexes having very low weight or being overlapped. As a result, the socio-demographic characteristic factor tuned out to have the highest importance among three factors with importance of about 33.8%, assuming that a total of importance for all three factors is 100%. The physical and situation characteristics were given importance of about 30.8%, which was lower than socio-demographic characteristics, yet considered much important for uating crime risks. And, the crime environment characteristics were found to have importance of no more than 6.1%, to be still needed for crime risk assessment but not considered as core factor. Among a total of 23 indexes considered, 'fear of crime' had the highest importance, followed by cohesion among residents, intangible crime cost, types of housing, and stability of the residence.
The regions to be studied were selected in the following way: At the 1st stage, cities of Seoul and Incheon and Gyeonggi-do province were selected among other regions since those regions had the largest population and the highest crime occurrence. Then, with a 2*2 cross-tabulation using two variables of crime rates and payment of property tax in 2010 in many local municipal governments, also in consideration of regional characteristics, accessibility, and appropriateness for research, four local governments were selected respectably for Seoul and Gyeonggi-do, making a total of eight areas: for instance, Gangnam-gu, Songpa-gu, Jungrang-gu and Eunpyeong-gu at Seoul, and Bucheon, Seongnam, Guri, and Dongducheon in Gyeonggi-do.
The regions(‘gu’) as a whole were not appropriate for assessing crime risks.
Therefore, after careful examination of statistical data on land use and major housing types in each area, the sub-administrative units of ‘dong’ were selected for 8 areas each with one dong of the highest ratio of conventional housing types and one dong of the highest ratio of apartment complex areas. This then made up a total sample of 16 areas for focused research. Then questionnaire survey and on-the-spot observation survey were done for 16 areas by using crime risk assessment tool. A social research company conducted questionnaire survey for one month from July 23, 2012 to August 22, 2012, visiting each household for interviews. The original research team divided into three did on-the-spot observation survey and in-depth interview from June 25, 2012 to July 20, 2012 by visiting each area of 16 dongs.

The findings were: The 16 dongs had average crime risk of 1.91 and could be classified into three classes, that is to say, high risk, middle risk and low risk. The high risk dongs consisted of B-dong, Bucheon, B-dong Eunpyeong-gu, B-dong Songpa-gu and B-dong Seongnam, and middle risk regions did B-dong Jungrang-gu, B-dong Gangnam-gu, B-dong Dongducheon, A-dong Bucheon, A-dong Guri and A-dong Eunpyeong-gu, A-dong Gangnam- gu, and A-dong Jungrang-gu, and low risk dongs did A-dong Songpa-gu, A-dong Seongnam and A-dong Dongducheon.
Dividing the 16 dongs into conventional housing areas and apartment complex areas by housing types reveals that the conventional housing areas had higher crime risk than the apartment complex areas. Among conventional housing areas, B-dong Bucheon, B-dong Eunpyeong-gu and B-dong Songpa-gu had high crime risk, while B-dong Gangnam-gu and B-dong Dongducheon had low crime risk. And, among apartment complex areas, A-dong Bucheon and A-dong Guri had high crime risk, while A-dong Seongnam and A-dong Dongducheon had low crime risk. The 16 dongs in Seoul and Gyeonggi-do were divided into two groups by housing types. Then, the conventional housing areas in Gyeonggi-do had the highest crime risk, followed by conventional housing areas at Seoul, apartment complex areas at Seoul and apartment complex areas in Gyeonggi-do. So, it appears to confirm that conventional housing areas were more exposed to crime risk and that the former was vulnerable to the crime. It also seems that conventional housing areas in Gyeonggi-do were more vulnerable to crime risk than the ones at Seoul. When it comes to household income, low income areas showed the highest crime risk, followed by middle income areas and high income areas.
This study investigated four regions carefully:
First, the 4 areas under case study had different results depending upon demographic factors and crime victim factors, and housing type had influence upon those factors. And, obsoleteness of the region that was thought to have great influence upon crime risk had less relation with the risk. Not only one-person household ratio but also foreigner ratio had influence upon crime risk.
Second, it appeared that crime inducing factor had more influence than crime control factor. The crime inducing factor included number of 112 police calls, number of habitual criminals in the area and victimization experience: And, conventional housing areas had higher crime risk than apartment complex areas. The conventional housing areas had more policemen, more people working for crime prevention and CCTV than the apartment complex area. So, it seemed that crime inducing factor had more influence upon crime risk than crime control factor.
Third, the 4 areas under case study showed that conventional housing areas had higher risk of physical environment than apartment complex areas. The apartment complex areas differed much from the conventional housing areas in access control, maintenance and negative use of land. This was because apartment complex areas with integrated control could manage crime risk more effectively than conventional housing areas with individual control.
Fourth, income level and economic conditions had much influence upon crime risk, and residents had various perception about crime prevention and sensitivity to the crime depending upon residential period of time.
In sum, it seemed that one-person household ratio and foreigner ratio had influence upon crime risks, and that crime inducing factor had more influence upon crime risks than crime control factor. Housing type, access control, maintenance and negative use of land also had influence upon crime risk. Residents' income level and residential period of time appeared to have sense relation with fear of crime and crime victimization cost.

The followings are policy alternatives that residents suggested through survey.
First, residents at conventional housing areas where low income citizens lived thought that hotels, bars, game rooms, amusement places, children's playground and empty spots could create potential crime. Children's playground and empty spots not controlled at residential area could produce harmful environment to induce teenagers to commit crime and to create fear to residents. On the other hand, residents at apartment complex areas thought that commercial facilities could be harmful. This was because land at Seoul and Metropolitan area was developed for mixed use of both residential and commercial.
Second, installation of CCTV and streetlight were thought to be main crime prevention measures that many residents commonly asked for. The residents did not think much of access control, territoriality, residents' autonomy and community based prevention. It is quite interesting that high income area residents preferred private security service, and that low income residents did better landscape and crossing gate for vehicles, while middle income residents did transparent walls and employment of security guards. Such trend appeared to have close relation with housing types and thus crime prevention measures should be taken depending upon income levels and/or housing types.
Third, installation of CCTV drew the most attention among public crime prevention measures, followed by bright lighting at public places, and keeping clean. The residents showed low preference for better environment by public designs, sports and leisure facilities for young people, better environment of the villages, and other new public crime prevention strategies such as residents' participation in community development. Nonetheless, residents at apartment complex areas in Gyeonggi-do who were middle income residents showed some interest in the new crime prevention strategies. And they were generally more interested in crime problems than residents at high income and/or low income areas.
Fourth, overall, the residents had negative opinion on future crime prevention plan except for 112 police calls. This was probably because the residents thought that crime risk was not community problem but individuals' problem. The residents at low income conventional housing areas seemed to seek to solve problems at individual level, while residents at middle income showed more interest in participation in crime prevention activities. The residents at high income areas preferred private security service and preventive methods against potential disorder. It may be that the low income residents, while being always exposed to crime risks, have no psychological and economic capabilities so that they were forced to prevent crime risk by physical monitoring. So, resident at high income regions may prefer to prevent crime risks by joining community activities. As such, effectiveness of public crime prevention policies might vary depending upon region.
This study suggested three stages of policy alternatives for applying assessment tool of crime risk; that is to say, diagnosis and assessment of crime risk, control and management of crime risk, and institutionalization of crime risk management.
First, police agency, local governments and policy authorities should recognize importance of crime risk assessment for diagnose and assessment of crime risk and develop scales proper to situations in the nation. The United States, the UK and Australia have developed objective assessment of crime risk to attain goals of commercial use as well as public interest. In Korea, measurement tools should be developed to assess crime risk with high precision and reliability for effective implementation of crime prevention policy. And assessment results should be shown to be important in deciding upon policies' priority and resource allocation. First of all, scales should be developed in accordance with situations in Korea so that they can be applied to practice and that assessment of crime risk can also be put into practice. The crime risk assessment should be done not just at one time but continuously for a long period of time.
Second, crime risk should be managed and controlled after diagnosis. The crime risk could be controlled from point of view of urban planning, risk control and resident participation. The urban planning includes use of the land, street and public facilities, and architecture and landscape and should introduce strategies linked with crime risk assessment. The risk control could select and apply management techniques considering likelihood of crime, results and effects. The risk control has three types; avoidance, acceptance and transfer. CPTED factor could be a type of risk control approach through improvement of physical environment. From point of view of the resident participation, crime occurrence could come not from physical disorder but from less monitoring and less participation of the residents. The crime prevention strategies from this point of view have been developed into various kinds of programs, and partnership among public administration, criminal justice agencies and local government has become more and more important.
Third, last stage of crime risk assessment is for institutionalization. This study suggested legal institutionalization, administrative institutionalization and economic institutionalization: Most of all, through legal institutionalization, assessment of crime risk should be continued and managed in stable ways. As in the case of industrial accident compensation insurance, the economic institutionalization to compensated for property losses from various risks should be also considered.