The relationship between children’s fear of crime and pedestrian volume in school zones

ABSTRACT Focusing on reducing the fear of crime among elementary school students, this study examined the relationship between pedestrian volume and fear of crime in elementary school zones. For this, we i) identified why and where in the school zones students fear crime, by conducting a survey among fifth and sixth grade students of two elementary schools; ii) calculated the volume of pedestrian traffic across the school zones through local integration analysis of axial map, and iii) determined the relationship between fear of crime and pedestrian volume and the difference in patterns of fear of crime perceived by the students depending on the pedestrian volume. The results indicated a higher prevalence of fear of crime in places with low and high pedestrian volumes compared with those with an average volume. Notably, in high pedestrian traffic areas, despite improvement in natural surveillance, the major determinants of high fear of crime were human factors, such as strangers and peers who bully. The findings will help architects, building engineers, and urban planners effectively plan and design schools and public facilities in school zones. Moreover, it will enable policy makers develop effective and accurate guidelines regarding positioning patrol personnel or surveillance.


Background
Crime rate targeting children under the age of 12 has been increasing over time in South Korea.Major crime types against these children include sexual assault, physical assault, and theft.Moreover, there are many criminal incidents and antisocial behaviors including school bullying, cyber bullying, and cursing that are not counted as official crime cases.Nevertheless, children experiencing such crime or fear of being victims of these criminal incidents is a matter of concern.Moreover, considering the fact that majority of these incidents happen in schools or on the way to schools (Byun and Ha 2017), investigating where and why children are likely to be exposed to criminal incidents or feel fear of crime is critical.
According to the Korean Police Crime Statistics (2021) of the National Police Agency of South Korea (2021), the number of crimes targeting children under the age of 12 in South Korea increased steadily from 2017 (13,569 cases) to 2019 (15,763 cases).Furthermore, the school violence survey conducted by the Ministry of Education in the first half of 2019, among students from fourth grade of elementary school to third grade of high school (equivalent to 12th grade) showed that the overall share of students who acknowledged being victims of school violence, was 1.6%, an increase of 0.3 percentage point, year-onyear (YoY).The corresponding share of victims in elementary schools was 3.6% (increase of 0.8 percentage point YoY), which was significantly higher than 0.8% in middle schools (increase of 0.1 percentage point YoY), and 0.4% in high schools (no change) (Ministry of Education Korea 2019).Previous studies have reported that continuous exposure to social crimes has a negative effect on the development of adolescents and their ability to adapt (Attar and Guerra 1994;Fitzpatrick and Boldizar 1993;Jenkins and Bell 1994;Martinez and Richters 1993), with fear of crime being closely associated with physical vulnerability (Mirrlees-Black, Mayhew, and Percy 1996).Several studies indicate that students who have experienced violence tend to have a reduced sense of safety and poorer academic achievement (Bowen and Bowen 1999;Ozer et al. 2017).Nijs et al. (2014) found that students' negative perception of school safety was associated with their mental health and psychosocial risks in a large sample of secondary school students (N = 11,130, aged 11-19).Moreover, elementary school students' fear of crime was closely related to the surrounding environment on their way to schools (Wiebe et al. 2013).In a study on the pedestrian environment around elementary schools, the students experienced fear of crime regardless of whether they had been subjected to it (Byun and Ha 2017).
Research on crime and fear of crime, however, has largely focused on adults.Hart, Chataway, and Mellberg (2022) reviewed 547 articles focusing on fear of crime from 1996-2021 and found that adults were the most frequently used unit of analysis (65%) in the study, whereas only 7% of the studies involved young adults or children.Even the studies focusing on children are often conducted based on parents' fear of crime.The study by Francis et al. (2017) has reported that parents' fear regarding their children being targeted by strangers is exaggerated and related to their socioeconomic status (SES), with anxiety being higher in the low SES group.Such studies may not reflect the fear of crime directly experienced by children.Therefore, it is necessary to examine children's perspectives, in addition to that of teachers and parents.Conducting research with children may not be easy because they may face difficulties in understanding the survey questions as well as the geography.However, the findings of a large-scale school survey conducted by Nayak (2003) suggest that children show acute location-based sensitivity to various highly relevant community safety issues.As presented by Pain (2001), there has been a shift from the focus on fear of crime in older people, to that in younger people, with vulnerability and impact of victimization in the latter being more recognized.

Objectives
This study aimed to examine the relationship between pedestrian volume and fear of crime in elementary school zones.Since the school zone is a linear space used by an unspecified number of people, we considered pedestrian volume as an important factor influencing elementary school students' fear of crime.Therefore, we conducted a survey focusing on elementary school students and analyzed the pedestrian volume in the school zones through axial map analysis of space syntax to explore the differences in the pattern of fear of crime in places with high, medium, and low pedestrian volumes.
In this regard, we evaluated the following hypotheses: (1) Within the school zone, fear of crime will be higher in places with low and high volumes of pedestrian traffic compared to those with average volume.
(2) The reasons for fear of crime will show different patterns in places with low and high pedestrian volumes in the school zone.
(3) The pattern of fear of crime in a neighborhood with high-rise apartments will differ from that with low-rise multifamily residential apartments.
The results of this study will have implications for environmental designers, school officials, parents, and education administration experts, and will ultimately contribute to reducing the fear of crime that elementary school students experience in the school zone.

Definition of fear of crime
The definition of fear of crime differs across studies (Collins 2016;Ferraro and Grange 1987;Henson and Reyns 2015;Warr 2000).There have been many competing conceptualizations of fear of crime including perceived risk, worry, anxiety, and fear as the part of the experience of fear of crime (Etopio and Berthelot 2022).Furstenberg (1971) was one of the first researchers who conceptualized fear of crime.He argued that "fear" is different from "concern" because fear of crime is measured as one's perceived chances of victimization, whereas concern refers to one's estimated seriousness of the crime (p.603).While Furstenberg defined fear of crime as perceived risk, Garofalo (1981) distinguished between perception and emotion.Garofalo (1981) defined fear as an "emotional reaction characterized by a sense of danger and anxiety produced by the threat of physical harm" (p.840).Ferraro and LaGrange (1987) defined fear of crime systematically.They separated perceptions of crime into three categories: judgments (about the risk of crime), values (concern), and emotions (fear), and defined fear of crime as "a negative emotional reaction to crime or the symbols associated with crime" (p.72).Later, Rader (2004) argued that emotion (fear of crime), cognition (perceived risk of victimization), and behavior (constrained behaviors) comprise the larger construct "threat of victimization."Although elementary school students may not have an exact concept of victimization, they may feel unsafe or threatened when they walk to school.Therefore, the fear of crime in this study is a broad concept of feeling afraid and the opposite of sense of safety.

Determinants of fear of crime
There have been several approaches to understand fear of crime including the vulnerability model (Killias 1990;Lewis and Salem 1986), victimization model (Box, Hale, and Andrews 1988;Garofalo 1979), disorder model (Skogan and Maxfield 1981;Taylor and Hale 1986), and social integration model (Hartnagel 1979;Hunter and Baumer 1982;Rountree and Land 1996).Among these theories, this study focused more on the vulnerability and social integration models.
Although direct victimization is associated with fear, this relationship is weak (Balkin 1979).The relationship, however, becomes strong by controlling for the most important vulnerability variables (i.e., sex, age, race, and income; Skogan and Maxfield 1981).Fear of crime is a subjective feeling, that one may be the victim of a crime, and does not directly reflect the actual crime.However, it clearly has harmful effects on individuals or society.Fear of crime limits activities and their scope, increases stress, and interferes with neighborhood integration, thereby increasing the probability of actual crime (Newman 1972;Silberman 1978;Skogan 1986;Taylor 1989;Warr and Stafford 1983;Warr 1985;Wilson and Kelling 1982;Wilson 1975).Moreover, these adverse effects may be more pronounced in groups with lower social status or physical strength, such as low-income families (Lavrakas 1982), less educated population (Kennedy and Silverman 1985), older adults, or young women (Box, Hale, and Andrews 1988;Skogan and Maxfield 1981;Warr 1985).Additionally, a recent study conducted with 13-16-year-olds by Rišová and Sládeková Madajová (2020) showed that regardless of the time of day, girls felt less safe in public spaces than boys did.Therefore, additional research is needed on crime and fear of crime in socially vulnerable groups, including young children.
Fear of crime is also caused by poor social integration.It is higher in large cities characterized by heterogeneity, social diversity, and high residential density (Moser 1999); therefore, it is more common in urban areas (Miceli, Roccato, and Rosato 2004).This is even stronger when the population is large and the neighborhoods are poorly integrated.A lack of social interaction raises fear of crime as people feel they are less likely to get help from those around them.However, extended and intense neighborhood relations lower fear of crime by strengthening a sense of community, providing social and emotional support, and creating neighborhood attachments (Farrall et al. 2000;Gubrium 1974;Ross and Jang 2000).

Crime prevention through environmental design (CPTED), focusing on natural surveillance and pedestrian environments
Crime prevention through environmental design (CPTED) is a crime prevention strategy aiming to reduce the opportunity of crime and fear through natural, mechanical, and procedural means.Jacobs (1961) suggested that the presence of "eyes on the street" would discourage criminal activities, Newman (1972) proposed the idea of "defensible space" and confirmed the relationship between physical design features and fear of crime, and the term CPTED was first formally outlined by Jeffery (1971).
Many studies have supported that CPTED is effective in preventing crime in the community.The four basic dimensions of CPTED are natural surveillance, access control, territoriality, and maintenance (Cozens, Saville, and Hillier 2005;Cozens 2002;Parnaby 2007).The purpose of surveillance is to reduce the likelihood of committing a crime by maximizing the visibility of a physical place through people or mechanical and electrical devices, thus making potential criminals aware that they are being observed.According to previous studies, natural surveillance is "the ability of everyday users to be able to see into and across space by virtue of physical design and site layout" (p.22) (Kitchen and Schneider 2007).It gives the impression to potential criminals that they may be exposed in the area, making them feel that they are being observed, thereby discouraging their criminal intentions (Cozens, Saville, and Hillier 2005).
In her book, The Death and Life of Great American Cities, Jacobs (1961) stated that shop owners and residents of apartments facing the street kept an eye on sidewalk activities, offering a safe urban environment for children and adults in mixed-use, crowded, and varied settings like Greenwich Village, her neighborhood in the 1960s.Wood, Frank, and Giles-Corti (2010) tested the "eyes on the street" hypothesis in association with active travel.After controlling for sociodemographic characteristics, they discovered strong and positive relationships between neighborhood design and safety perceptions.They emphasized the association between density of commercial floor space, a sense of community, and walking behaviors.Based on their findings, hospitality and social interactions in the urban space seem conducive to active travel.This aligns with the concept of New Urbanism, suggesting that high quality pedestrian environments enhance a sense of community by encouraging local social interactions (Lund 2002).A recent study by Jing et al. (2021) suggested that the green spaces used by many residents can lower fear of crime.They tested the relationships between street-view greenery and fear of crime in the neighborhood and the results demonstrated that a higher level of neighborhood street-view greenery was associated with a lower fear of crime.This is because outdoor spaces with more trees are consistently used by residents than spaces without trees.

Space syntax and axial map
Space syntax, developed in the late 60s by Hillier and Hanson (1984), refers to a group of theories that examine the social use of space.Axial map is one of the main analysis techniques of space syntax, often used to analyze urban spaces.The depth, connectivity, and integration of a space can be analyzed based on the axial line, which is the longest straight line with minimum number of spaces.Depth refers to the number of spaces that one must pass through to access a certain space, and connectivity refers to the number of other spaces directly connected to a unit space.Thus, a low depth indicates high accessibility, whereas a high connectivity indicates easy access to adjacent spaces.Integration is an index indicating the accessibility of each unit space, in the overall spatial distribution.It indicates how close the original space is to all other spaces and can be viewed as a measure of relative depth.Generally, when the degree of integration is greater than 1, it indicates a space with a high degree of integration, that is, it is a central space with good access to the entire area being analyzed.
In this study, a radius-3 integration analysis of the axial map was performed to examine pedestrian volume.Radius-3 integration is also called "local integration," which allows the integration to be calculated only up to three lines away from each line in every direction.The pedestrian densities of local areas can be best predicted via radius-3 integration (Hillier 2007); thus, it can be used to calculate the pedestrian volume for analysis.

Space syntax and crime
The studies in the past decade focused on the effect of configurational properties on crime.These studies revealed correlations between measures of space syntax and crime in residential neighborhoods (Hillier and Shu 2000).
Jones and Fanek (1997) conducted a study on the "correlations between syntax values and crime rates."They observed the effect of spatial configuration on crime in Austin, Texas.Four pairs of tracts with each pair having similar income, poverty rates, population, and racial composition were selected.The Axman software was used to calculate the radius-3 integration, radius-10 integration, and connectivity values for each tract.Correlations between syntax values and crime rates were examined.The results demonstrated that pairs with higher integration values had lower crime rates.Moreover, three tracts with higher mean radius-3 integration and connectivity values had lower crime rates.The authors suggested that better connected streets encourage greater pedestrian movement, thereby increasing surveillance.
The results of the abovementioned studies differ from the hypotheses of this study -fear of crime will be higher not only in places with low pedestrian volume, but also in those with high pedestrian volume, compared to those with average volume.We expected different results because this study was based on fear of crime, not actual crime, and focused on elementary school students instead of adults.

Study area
Seoul, the South Korean capital, comprises 25 autonomous districts.Songpa-gu, with a high crime rate and a large number of elementary school students, was selected as the study area.According to Seoul Statistics 2019, Songpa-gu had the largest number of elementary school students among the 25 autonomous districts, with a total of 32,208 out of 422,293 students in Seoul (Seoul Metropolitan Government 2019a).Additionally, the number of crimes in Songpagu was 5,698 as of 2019, which was significantly higher than the average of 4,146.72 across the capital (National Police Agency of South Korea 2021).Therefore, Songpa-gu was considered suitable for research on fear of crime around elementary schools.
Among 40 elementary schools in Songpa-gu, we selected elementary school A in a neighborhood with high-rise apartments and elementary school B in a neighborhood with low-rise multifamily residential apartments (Figure 1).High-rise apartment complexes and low-rise multifamily residential apartments are representative of different residential environments in Seoul.Since there were differences in the arrangement of the buildings and the school zones, as shown in the satellite images in Figure 1, these two sites with different characteristics were selected to increase the reliability of the study, test its generalizability, and observe differences in patterns of fear of crime.Based on the 2019 data of Statistics Korea, major housing types in South Korea were apartment houses, detached houses, low-rise multifamily houses, and houses in commercial buildings (Statistics Korea 2019).According to Seoul Statistics 2019, the total number of households in Seoul were 3,738,598.The major housing types in Seoul were apartment houses (42.4%), followed by detached houses (27.9%) and lowrise multifamily houses (21.2%) (Seoul Metropolitan Government 2019b).The Songpa-gu area selected for this study is characterized by a mixture of high-rise apartments and low-rise multifamily residential apartments, and schools representing the neighborhoods with each housing type were selected as the study sites.Elementary school A covers a 1,500-unit apartment complex built in 1980 and a 6,860-unit apartment complex built in 2008.Elementary school B covers an area with low-rise multi-family residential apartments, each comprising 5-40 households.Figure 2 highlights the difference between high-rise apartment complex and low-rise multifamily residential apartments in Songpa-gu.

Survey on fear of crime
We conducted a survey among fifth-and sixth-grade students (aged 11-13) from the elementary schools A and B to investigate the places that caused fear of crime within each school zones.A school zone is a special protection area set up around schools since children mostly commute to schools through this area.In South Korea, the area within 500-meter radius of the front gate of elementary schools is designated as school zones.Also, a previous study on fear of crime in street environments nearby elementary schools showed that most elementary students experienced fear of crime within a certain distance (500-meter) of their schools (Byun and Ha 2017).
To determine where and why students felt fear of crime, the survey was conducted exclusively with students without adults.The reason for limiting the survey to fifth-and sixth-graders in elementary schools was to ensure that respondents could understand the purpose and content of the survey and respond without the help of their parents.The survey was conducted by collecting the questionnaires that were distributed by mail to households that agreed in advance.The study was approved by Chung-Ang University Institutional Review Board (approval no.1,041,078-202,003-HRSB-079-01).The survey was conducted from May to June, 2020.The students were asked to mark the spots where they experienced fear of crime on the satellite image of the school zones around their schools.In previous research (Lee, Ryu, and Ha 2012;Svatonova and Rybansky 2014), children stated that they were wellaware of the satellite image of the neighborhood.Considering this, we presented the satellite image around their schools and asked them to mark the areas where they felt fear of crime.Regarding fear of crime, we clarified that fear, in this context, was feeling unsafe and it did not include road safety such as fear of getting hit by a car.Then, students were asked to write down why they felt fear of crime at each spot.We limited the number of spots to three to prevent any respondent's opinion from becoming too dominant or insignificant.In elementary school A, of the 200 questionnaires that were sent out, 108 were received from 52 male and 56 female students.In elementary school B, of the 200 questionnaires that were sent out, 102 were received from 50 male and 52 female students.All questionnaires were answered faithfully and thoroughly, and the responses were valid to use for analysis.A sample questionnaire response is shown in Figure 3.

Axial map analysis
To examine the relationship between pedestrian volume and fear of crime in the school zone, we used an axial map, which is a space syntax methodology often used to identify the movement patterns of people in urban spaces.Radius-3 integration analysis of the axial map was conducted using Depthmap on the area within a 500 meter radius from each elementary school.Consequently, 219 and 132 lines were identified in elementary schools A and B, respectively.The integration value ranged from 0.333 to 3.099 in elementary school A, and from 0.728 to 3.523 in elementary school B. The color image of the axial map shows that the lines closer to red and farther away from blue have a higher degree of integration, as well as higher pedestrian volume and density (Figure 4).

Places with high fear of crime
A total of 128 and 124 spots for fear of crime were indicated around elementary schools A and B, respectively, as shown in Figure 5 (created with ArcGIS 10.5).The areas covered by the questionnaire were displayed on the axial map.The darker the pink dot on the map, the higher the frequency of indicating it as a spot for fear of crime in the survey.In both types of residential areas, places behind a building, spaces between buildings, parks near schools, and intersections were found to be the main spots where students experienced fear of crime (Figures 6 and 7).

Frequency of axial lines
As a result of radius-3 integration analysis of the axial map, 219 and 132 lines were identified in elementary schools A and B, respectively.Also, the integration value ranges from 0.333 to 3.099 in elementary school  A, and from 0.728 to 3.523 in elementary school B. We recoded the integration value with an ordinal variable ranging from 1-10 and calculated the frequency of axial lines at each level of integration.Figures 8 and 9 are the graphs showing the frequency of the axial lines in recoded integration in elementary schools A and B, respectively.The distribution was positively skewed in school A and negatively in school B.

Density of fear of crime by integration level and effect of pedestrian volume
Figures 10 and 11 represent the frequency of fear of crime by integration level at elementary schools A and B, respectively.Similar to the previous analysis, we recoded the integration value with an ordinal variable ranging from 1-10 and calculated the frequency of  fear of crime at each level of integration.The distribution was negatively skewed in both schools.In both schools, fear of crime was greater in places with higher integration.
Next, to examine the relationship between integration and fear of crime, we conducted a regression analysis for each school, with the density of fear of crime for each level of integration as the dependent variable and integration level as the independent variable."The density of fear of crime by integration" was calculated by dividing "the frequency of fear of crime by integration level" by "the frequency of the axial lines by integration level."This regression model was suitable with F = 14.476 (P < 0.01) at elementary school A and F = 7.705 (P < 0.05) at elementary school B. As indicated in Table 1, both schools showed higher R 2 (coefficient of determination) values in the quadratic model than in the linear model.Elementary school A, R 2 = 0.805, showed an explanatory power of 80.5%, whereas elementary school B, R 2 = 0.688, showed an explanatory power of 68.8%.The shape of the secondary trend model showed that both schools had a U-shaped curve (see Figures 12  and 13).Therefore, fear of crime was high in places with low and high pedestrian volumes.

Reasons for fear of crime
As mentioned earlier, students were asked to mark spots on a map of the school zone where they experienced fear of crime and felt unsafe except for the fear related to road safety, such as getting hit by a car, and write the reasons for each spot.Additionally, the number of responses was limited to three to prevent the     influence of any respondent's opinion from becoming too dominant or insignificant.
The reasons for the fear of crime were divided into low (30%), medium (40%), and high (30%) levels based on integration levels of 1-3, 4-7, and 8-10, respectively, that is, places with low, average, and high pedestrian volume to identify the differences in the causes of fear of crime.The results of the two study sites were individually synthesized, as shown in Tables 2 and 3. To check the generalizability of the results, we first examined whether there was a difference caused by the characteristics of the two residential areas.While there were differences in details, the results were similar in general.
Considering that the number of spots with medium integration was the highest as indicated by the frequency of the axial lines in Section 4.2, which showed normal distribution, the number of reasons in the questionnaire was the highest with medium levels of integration (see Figures 8 and 9).Nevertheless, as shown in Section 4.3, the density of fear of crime was high in places with low and high pedestrian volumes (see Figures 12 and 13).Therefore, rather than focusing on the number of answers for each spot, the results were interpreted with a focus on which reasons were higher in each level of integration and how they differed depending on the integration level.2), located in the high-rise apartment neighborhood, human factors were higher than space factors in all integration levels.In the high level of integration (30%), human factors accounted for 60% and space factors for 40%, indicating a large gap.In this level, among human factors, peers who bullied accounted for 30% and strangers for 13%.Overall, 43% of students answered that they were afraid of crime because of the presence of other people in the school zone.Although responses classified as "direct criminal experience" accounted for a small proportion, they needed attention and are mentioned separately.In a particular case in the high integration level (30%), one respondent said, "I almost got assaulted by a senior," and in another in the medium integration level (30%), said, "Someone followed me."These experiences were not of actual crimes, but those of feeling threatened.Space factors were highest at 47% in the low integration level (30%), with darkness being the greatest fear at 33%.Darkness was also the highest (23%), among space factors in both the medium (40%) and high (30%) levels of integration.
The findings of elementary school B (Table 3), located in a low-rise multifamily residential neighborhood, revealed that at the low (30%) and medium (40%) levels of integration, space factors were higher than human factors, at 57% and 52%, respectively.By contrast, at the high integration level (30%), human factors accounted for 70% and space factors for 30%, indicating higher human factors.Among the human factors in this section, peers who bully accounted for 37% and strangers for 20%.Overall, 57% of the students stated that they feared crime owing to the presence of other people in the school zone.Answers classified as "direct criminal experience" accounted for a small proportion, with one answer in the low integration level (30%): "A menacing senior demanded money." Space factors were highest at 57% at the low integration level (30%).Among the reasons, darkness was the highest at 43%, and it was also the highest at 35% and 20% among space factors in the medium (40%) and high (30%) levels of integration, respectively.
Combining the results of the two study sites, shown in Table 4, in the low (30%) and medium (40%) integration levels, space and human factors appeared in almost the same proportion.However, in the high integration level (30%), human factors were higher at 64% and space factors were lower at 36%.The biggest among the human factors in all integration levels was the presence of peers who bully.In the low (30%) and  medium (40%) integration levels, this reason was followed by less people, whereas in the high (30%) integration level, it was followed by strangers.Moreover, the biggest among space factors across levels was darkness.

Discussion
This study aimed to explain the effect of pedestrian volume on fear of crime in two different elementary school zones and identify the difference in patterns of fear of crime experienced by elementary school students depending on pedestrian volume.To this end, we first identified the spots where fear of crime was experienced in the school zones by conducting a survey among students in two schools with different environments (elementary schools A and B).We also implemented the site as an axial map.Then, we measured the frequency of fear of crime according to pedestrian traffic volume represented by integration values at the axial map.Next, to examine the relationship between pedestrian volume and fear of crime, we conducted a regression analysis for each school.The results showed high fear of crime in elementary school zones with low and high pedestrian traffic volumes; fear of crime due to bullying peers and strangers in high pedestrian volume areas; and fear of crime caused by darkness and less people in low pedestrian volume areas.
First, we confirmed the hypothesis that fear of crime is higher in places with low and high pedestrian volumes than those with an average volume.This result is different from previous studies on the relationship between fear of crime and natural surveillance which mentioned increased natural surveillance is expected to reduce criminal activity by discouraging potential criminals' intentions and reduce fear of crime in tern (Cozens, Saville, and Hillier 2005).However, the present study revealed that, despite improvement in natural surveillance, in addition to places with low volume, fear of crime was high in places with high pedestrian volume.
Second, the reasons and patterns of fear of crime were expected to differ in places with low and high pedestrian volumes.This hypothesis was inspired by a previous study by Lee and Ha (2015) who conducted a survey among elementary school students and studied the relationship between visibility and fear of crime in the external space of an elementary school building using visibility graph analysis.In this study, fear of crime was high in both low-and high-visibility spaces, mainly owing to human factors.High visibility can enhance natural surveillance; by contrast, anonymity (blending into a crowd) can increase fear of crime.Similarly, in the present study, we found that in lowvolume places, factors associated with a decline in natural surveillance, such as darkness and less people, were most likely to cause fear of crime.In contrast, in high-volume places, fear of crime was high despite the improvement in natural surveillance, mainly due to human factors in both study sites.The presence of peers who bullied and strangers were the major determinants of fear of crime for the students.Previous studies have demonstrated that social interaction may lessen fear by giving social and emotional support in dealing with criminal apprehension (Gubrium 1974).Although the presence of strangers can be a threat to adults as well, it is particularly noteworthy that, for young elementary school students, the presence of seniors, not just strangers, poses the greatest threat.The findings of this study are consistent with that of previous studies reporting that young people feel threatened by other young people as well as adults (Pain, Williams, and Hudson 2000;Zeng, Mao, and Wang 2021).This is also consistent with the study conducted with 12-15-year-olds by Nayak (2003) reporting that "children are both the perpetrators and victims of crime."Therefore, it is necessary to pay attention to the fact that young elementary school students can also be afraid of crime owing to peers or seniors, who they socialize with, rather than strangers from whom social and emotional supports are not expected.In this regard, different approaches may be required to reduce the fear of crime among elementary school students in places with low and high pedestrian volumes in the school zone.
Regarding the determinants of fear of crime, space factors were relatively high in places with low pedestrian volume and human factors in places with high pedestrian volume.Among the space factors, darkness and among human factors, the presence of others, such as peers who bullied and strangers, were the main determinants.Based on these findings, it seems necessary to increase lighting in places with low pedestrian volume, and to maximize surveillance by people or electronic devices in the physical environment.Nevertheless, in places with high pedestrian volume, in addition to procedural means such as regular patrol by the police and school officials, preparation of inter-school programs that strengthen exchanges between neighboring elementary, middle, and high school students may be considered to naturally increase social ties.Third, the patterns of fear of crime were expected to differ between neighborhoods with high-rise apartments and those with low-rise multifamily residential apartments.Although human factors were higher than space factors in the high-rise apartment neighborhood whereas space factors were slightly higher than human factors in the low-rise multifamily residential neighborhood, in a larger context, there was no significant difference.In both elementary schools A -located in a high-rise apartment neighborhood, and B -located in a low-rise multifamily residential neighborhood, fear of crime was higher in places with low and high pedestrian volumes than in places with average volumes.

Strengths and limitations
This study is meaningful in that it provided information about the location of the fear of crime experienced by young students that teachers and school officials may be unaware of.It can be used as a reference while designing the environment in the school zone.Furthermore, it has attempted to identify the relationship between pedestrian volume and the fear of crime by quantitative analysis using space syntax.
This study was conducted among elementary school students and the survey was conducted with fifth and sixth grade students to increase the reliability of the questionnaire responses.However, there may be limitations in the ability of young students to precisely mark the spots on a map.Therefore, the accuracy of the data should be further improved by using a global positioning system in future studies involving similar respondents.In addition, this study involved only two schools and 210 students.In follow-up studies, a larger sample size should be included to improve the reliability of the study.Additionally, more detailed research results can be obtained by increasing the sample size as well as asking the participants to rate the spots according to the degree of fear of crime with reasons.Moreover, adding a question about places where one feels safe along with places that arouse fear of crime can help identify the characteristics and differences between these places, thereby providing clarity.We plan to strengthen the contents of the questionnaire in this direction in a follow-up study.

Conclusion
Exposure to crime and fear of crime have negative effects on the development of children.Moreover, elementary students' fear of crime is closely related to the pedestrian environment on their way to schools and they may experience it regardless of being subjected to crime.It is important to reduce the fear of crime that they may experience in daily life.In this regard, we tried to determine where in the school zones is fear of crime experienced along with reasons for this fear.Thus, this study aimed to analyze the associations between pedestrian volume and fear of crime and identify the patterns of fear of crime in elementary school zones.Major findings of this study were: fear of crime is high in elementary school zones with low and high pedestrian traffic volumes; fear of crime is mostly caused by bullying peers and strangers in areas with high pedestrian traffic volume while it is caused by darkness and less people in areas with low pedestrian traffic volume.These findings are valuable in that few previous studies directly reflected young elementary students' opinion regarding fear of crime and identified different patterns of fear of crime depending on pedestrian volume.Based on the results of this study, different approaches may be needed to reduce the fear of crime among elementary school students in places with low and high pedestrian volumes in school zones.The results indicate the necessity to develop a program that helps increase the social and communal ties in places with high pedestrian volume, while focusing more on improving the physical environment, through natural surveillance, in places with low pedestrian volume.Elementary school students were markedly afraid of their seniors; however, this could not be improved by modifying the physical environment alone.We expect that the results of this study will contribute to the efforts to reduce children's fear of crime by connecting various parties such as the police, public administrators, policy makers, education administrators, guardians, as well as environmental designers.The reasons why students face fear of crime in specific areas can be identified, and police or public administrators can expand the police patrol or install streetlights and closed-circuit televisions (CCTVs) in appropriate places.The education administrators or guardians of the students can identify the places with high fear of crime and take precaution about those areas.Moreover, a school-wide support might be needed to promote interaction among students from different grades or from nearby schools.Designers, architects, building engineers, urban planners, and policy makers can identify the features of areas with high fear of crime and suggest design and operational guidelines to improve the present environment or consider it in future school zone planning.

Figure 1 .
Figure 1.Satellite imagery: elementary school A (left) and elementary school B (right).

Figure 2 .
Figure 2. Two different residential environments (left: high-rise apartment complex, right: low-rise multifamily residential apartments).Photographed by the first author, 2021

Figure 5 .
Figure 5. Spots with high fear of crime on the axial map (left: elementary school A, right: elementary school B).

Figure 6 .Figure 7 .
Figure 6.Images of the spots with high fear of crime around elementary school A. Photographed by the first author, 2021

Figure 8 .
Figure8.Frequency of axial lines by integration level (elementary school A).Integration is a normalized distance measure from the original space to all other spaces in the system(Hiller and Hanson, 1984).

Figure 9 .
Figure9.Frequency of axial lines by integration level (elementary school B).Integration is a normalized distance measure from the original space to all other spaces in the system(Hillier and Hanson 1984).

Figure 10 .
Figure10.Frequency of fear of crime by integration level (elementary school A).Integration is a normalized distance measure from the original space to all other spaces in the system(Hillier and Hanson 1984).

Figure 11 .
Figure 11.Frequency of fear of crime by integration level (elementary school B).Integration is a normalized distance measure from the original space to all other spaces in the system (Hiller and Hanson, 1984).

Figure 12 .Figure 13 .
Figure 12.Scattergrams plotting of the values of the relationship between density of fear of crime and integration level (elementary school A).Abbreviations: FoC, Fear of crime

Table 1 .
R 2 value in linear and quadratic models.

Table 2 .
Reasons for fear of crime by integration level (elementary school A).

Table 3 .
Reasons for fear of crime by integration level (elementary school B).

Table 4 .
Reasons for fear of crime by section (elementary schools A and B).