You are on page 1of 17

Locke Science Publishing Company, Inc.

SAFE COMMUNITIES FOR URBAN RENTERS: RESIDENTS' PERCEIVED SAFETY, PHYSICAL


TERRITORIALITY, AND SOCIAL TIES IN URBAN APARTMENT PROPERTIES
Author(s): Suk-Kyung Kim and Andrew D. Seidel
Source: Journal of Architectural and Planning Research, Vol. 29, No. 2 (Summer, 2012), pp.
133-148
Published by: Locke Science Publishing Company, Inc.
Stable URL: https://www.jstor.org/stable/43030966
Accessed: 05-12-2019 16:07 UTC

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/43030966?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Locke Science Publishing Company, Inc. is collaborating with JSTOR to digitize, preserve
and extend access to Journal of Architectural and Planning Research

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 133

SAFE COMMUNITIES FOR URBAN RENTERS:

RESIDENTS' PERCEIVED SAFETY, PHYSICAL


TERRITORIALITY, AND SOCIAL TIES IN URBAN
APARTMENT PROPERTIES

Suk-Kyung Kim
Andrew D. Seidel

This study suggests ways of creating safer residential environments for urban renters, explores
whether physically gated and fenced residential environments affect residents' actual and
perceived safety , and investigates determinants that would improve residents 'perceived safety in
their residential environments. The major data collection methods were site visits and a
questionnaire survey of residents in targeted properties . Apartment communities were divided
into three categories based on gate control and fences: gated communities with fully controlled
gate systems , perceived-gated communities with fences and gates that are not fully controlled '
and non-gated communities with neither fences nor controlled gates. Results suggested
residents feel safer in gated communities and perceived-gated communities than in non-gated
communities. These results support the importance of territoriality for improving residents'
perceived safety in apartment communities. Residents ' perceptions of safety in perceived-gated
communities, however, are statistically similar to those of residents in fully controlled gated
communities. This finding suggests exclusive territoriality, achieved by disconnecting a
community from neighboring ones, is not necessary for guaranteeing residents ' safety.

Copyright © 2012, Locke Science Publishing Company, Inc.


Chicago, IL, USA All Rights Reserved

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 134

INTRODUCTION

Since Newman (1973) introduced neighborhood design considerations an


space theory as a means of reducing residents' fear of crime in public housin
the 1970s, a number of researchers have explored the relationship betwe
environments of residential properties and safety from crime (e.g., Brunson,
Moran and Dolphin, 1986; Normoyle and Foley, 1988; Taylor, et al ., 1984; We
Anderson, 1982). Defensible space theory provided a model for residential env
prevent crime by creating the physical expression of a social fabric tha
(Newman, 1973:3).

One contemporary topic related to Newman's works is the issue of gated com
cause gated and fenced territories are known to provide defensible space
gated community is defined as "a subdivision or neighborhood, often surround
to which entry is restricted to residents and their guests" (Pickett, 2006). Defin
gated community is a residential area with restricted access, fenced territory, p
nal roads blocked from unknown visitors, and public spaces within the gated te
be shared among residents (Blakely and Snyder, 1997; Tijerino, 1998). Due to t
istics, gated communities are regarded by neighborhood designers and comm
as a solution to control unwanted visitors and traffic in residential developm

According to the United States Census Bureau American Housing Survey, in 2


seven million households were secured communities with walls and fences (U.
reau, 2002), and that number increased to over 10 million in 2007 (ibid., 2007).
secured multi-units also increased from 5.3 million in 2001 to 6.8 million in 2
popular among the wealthy in the U.S. beginning in the 1800s, gated commun
became available to members of nearly every income level, including those re
apartment properties. This popularity turns various discussions about gated c
issues of safety, urban segregation, and community cohesion (e.g., Atkinson,
and Snyder, 1998; Goix, 2003; Wilson-Doenges, 2000). Among those issues,
highlighted because of the unique environments of these communities cr
fences, and privatized public spaces.

Halberg (2001) explored why individuals move into gated communities and fo
people perceived them as being more secure. The perception of greater security
fenced territory was also identified by Newman (1996:69), who came to a simil
an earlier study concerning residents' association of increased security w
fenced territory as free from vandalism. Some studies, however, have rejecte
between residents' perceptions of increased security and gates/fences. W
(2000) explored residents' sense of community and fear of crime in gated
communities and ultimately dismissed a correlation between residents' percei
community's gates and fences. Similarly, Blakely and Snyder (1997) and Fowl
(1986) found no relationship between actual crime rates and gated territories

Beyond the ongoing debate concerning whether a correlation exists between


ceived security and gated living areas, previous studies exploring perceptions
gated residential properties may be critiqued on the basis of their research des
odologies. Most divided their treatment of the architectural features of their
either gated or non-gated communities, focusing mostly on residents' percep
within gated or non-gated community boundaries, and merely compared
concerning the two community types. However, further discussions of intrica
characteristics created by gates and fences are needed to adequately explore t
between the levels of territoriality created by these physical artifacts. Previ

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 135

also failed to account for people's current crime experiences in gated and no
nity areas. Therefore, it is necessary to define more appropriate research sett
more fully the relationship(s) between gated and fenced residential territor
perceived and actual safety.

In addition, previous works have rarely considered middle- or low-income f


gated multifamily housing in urban areas. Most previous subjects, except th
Doenges's (2000) study, lived in single-family housing developments in
Nonetheless, low-income families or renters living in urban areas shoul
residential environments. According to the U.S. Department of Justice (2004
in the U.S. occurs more often for those living in rented properties and urb
fore, crime prevention may be more important in multifamily housing, and
understand how such families view gated communities as one tool to h
challenge.

To begin to address these gaps in the relevant literature, this study sought to identify ways to
create safer residential environments for urban renters by pursuing two primary research
objectives. The first was to explore whether physically gated and fenced multifamily residen-
tial environments affect residents' perceived and actual safety. We investigated residents'
crime experience in their current properties and their perceptions of safety from crime. The
second objective was to explore explanatory variables regarding residents' perceived safety in
their residential environments and investigate determinants that might help improve resi-
dents' perceived and actual safety in those environments.

REVIEW OF LITERATURE AND RESEARCH FRAME

Safety Issue in Residential Environments and Explanatory Variables

The issue of crime in neighborhoods has long been accompanied by residen


the actual and perceived safety of their living environments. In addition t
study, Weidemann and Anderson (1982) have explored residents' perception
and safety in multifamily housing. They highlighted safety in residential
critical predictor of residents' satisfaction with their housing sites.

Taylor, et al. (1984) examined effective physical and social factors for redu
block level and found both physical factors and social factors, such as socia
territorial attitudes, were significant in preventing crime. Normoyle and Fo
the correlation between fear and perceptions of the local crime problem am
dents living in high-rise public housing sites. They found the local crime r
recent victimization experience, residents' segregation status, whether resi
high- or low-rise building, and the percentage of the elderly residing within
ing site were related to residents' perceptions of safety. Rohe and Burby (
explanatory factors associated with fear of crime among public housing re
victimization experience, crime level, and social attachment were related to
tions of safety.

Holzman, et al. (1996) focused on physical attributes associated with residen


of safety from crime and found a relationship between the perceptions and
sign. Their study suggested perceived safety was closely associated wit
property, the presence of police or a private security guard, fence around
erty, visitor intercoms, and video cameras in public areas. Fisher and Nasar
the component of prospect and refuge, along with lighting, as important va
residents' fear of crime.

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 136

In addition to the variables indicated above, Blakely and Snyder (1999), Mor
(1986), and Newman (1996) have all emphasized territoriality in residential en
have suggested reducing through-traffic and providing physical territory to
residential areas. Wilson-Doenges (2000), however, did not strongly support
fenced territories to guarantee residents' safety on their properties.

Blakely and Snyder ( 1 999) offered more concrete crime prevention tactics f
ronments in urban areas, such as physical designs, managerial plans, police
actions. Many of their strategies repeated Newman's (1973) crime-free neig
considerations. In addition to the physical design interventions identi
Blakely and Snyder (1999) suggested managerial considerations (such a
guards) and social tactics (such as increasing residents' interactions with ne
ing education programs for inhabitants, and encouraging neighbors to be i
communities).

Among the demographic and socioeconomic variables affecting residents' perceived safety, a
number of studies have concluded gender is the most powerful explanatory variable in shap-
ing a person's perception of safety from crime (see Kim, 2006; Perkins and Taylor, 1996; Taylor,
et al., 1984). Investigators have also found other demographic and socioeconomic variables
to be related to perceptions of safety, including age (Brunson, et al. , 200 1 ; Taylor, et al. , 1 984),
number of children (Wilson-Doenges, 2000), length of residence (Taylor, et al., 1984; Wilson-
Doenges, 2000), and income (Kim, 2006; Newman, 1996; Taylor, et al., 1984). Level of social
interaction (termed social attachment, neighborhood attachment, or neighborhood level) has
also been found to be a significant variable in explaining individual residents' perceived
safety in their communities (see Funk, et al. , 2007; Kim, 2006; Newman, 1973; Rohe and Burby,
1988; Taylor, et al., 1984; Weidemann and Anderson, 1982).

This research uses a holistic approach to explore residents' actual and perceived safety in
urban multifamily housing communities. We included variables linked not only to physical
environments but also other characteristics, including a managerial dimension and residents'
social ties, in developing our integrated measurement design.

Gated Communities and Related Studies

To prevent crimes in housing projects and small neighborhoods, Newman (1996:48) suggested
gates be established at their entrances to block unwanted traffic from passing through the
sites. For example, Newman proposed gates for the entrances of each neighborhood in the
Five Oaks community of Dayton, Ohio. Gates were planned on roads to control vehicle access
and on pedestrian routes, but the pedestrian-oriented gates remained open.

Newman (1996:49) also proposed a fence that extended pedestrian gates to adjacent physical
buildings. Though the municipal government simplified Newman's original fence design, the
addition of gates and the fence to the neighborhoods consequently had a positive effect on
the residents. As the boundaries established control of internal streets and roads, children in
the neighborhoods began to play inside the gated territory.

Urban sociologists Macionis and Parrillo (2004: 127) suggested that recent research exploring
gated communities has addressed four basic value valences: sense of community, segregation
and protection from the outside, privatization, and resident homogeneity. Many architectural
researchers (see Blakely and Snyder, 1997; Halberg, 2001; Wilson-Doenges, 2000) have ex-
plored safety associated with the territorial characteristics of gated communities, as men-
tioned above, in addition to these four values.

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 137

METHODOLOGY

Site Selection

This research employed site visitation and a questionnaire survey mailed to residents living in
selected properties to explore whether gated and fenced multifamily residential environments
affect residents' perceived and actual safety. To prevent the crime rate itself from biasing
residents' perceptions of safety, we first investigated the crime rates of major Texas cities. The
city of Houston had a medium crime rate, including both violent and property crimes (i.e.,
5,505.4 crimes per 100,000 inhabitants), among representative Texas cities (Austin, Houston,
and Dallas) (U.S. Department of Justice, 2004c: 171-173). Thus, Houston was selected as the
subject city. Considering the distribution of multifamily housing in the city, the northwestern
area of Houston was designated as the target population for this research.

We constructed a list of 72 apartment communities in the northwestern area of Houston based


on a thorough review of a website for advertising rental units (apartmentguide.com). Depend-
ing on the presence of fences and the level of gate control, we divided our list into three types:
gated communities with fully controlled gate systems and fences, perceived-gated communi-
ties with fences and gates that are not fully controlled, and non-gated communities with
neither fences nor controlled gates. The second community type was labeled "perceived-
gated" because the residents in those communities regarded their communities as gated, but
their physical territories were not clearly closed because of open gates. Gated communities
fully control access from outside traffic while perceived-gated communities do not do so
effectively due to open gates. Non-gated communities do not control the access of traffic from
the outside at all.

After dividing our 72 apartment properties according to gate control systems and fences, we
identified the rental price, property size, and property amenities (including sports facilities and
outdoor environments) of all of the properties we studied. After obtaining all of the relevant
information from the apartment websites, we contacted each property via telephone and con-
firmed its site amenities. To select subject properties with similar sizes and conditions, we
determined the average rental price for a two-bedroom apartment unit. On this basis, we
selected 273.5 ± 91.9 two-bedroom apartment units with a rental price of $805 ± $171.50 as
subject properties for our survey. Finally, six communities were chosen from each community
type. General information for each of the 1 8 sample properties was collected through site visits
with a checklist to investigate their specific physical characteristics, including gating condi-
tions, fences, lighting, community facilities, and apartment unit floor plans.

Sampling

All residents living in the selected 18 properties were potential survey subjects. We con-
ducted a pilot study in February 2005 to check the efficacy of our questionnaire and the
appropriateness of our mail-survey procedure. Questionnaires were distributed to randomly
selected residents from the entire population of the 18 properties. Based on the assumption
that statistical analysis would be valid with 60 cases from each of the three types of properties,
the minimum sample size for the analysis was set at 1 80. Considering the return rate of the pilot
study (22%), a total of 900 resident addresses were selected: 300 from gated communities, 300
from perceived-gated communities, and 300 from non-gated communities. Questionnaires
were distributed three times to selected addresses over the course of four months. We identi-
fied invalid addresses one month after the initial survey distribution in March 2005. Additional
addresses were then selected, and questionnaires were sent to the new addressees. A total of
207 questionnaires were collected, with the final response arriving in March 2006, and all
responses were used for analysis. The overall return rate of the survey was 16.2%.

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 138

Public

(Roads, comm

FIGURE 1. Hie
based on N

Measurement

"Actual safety" refers to residents' personal experience of crime in their residential environ-
ments. Thus, we first investigated residents' crime experiences since moving to their current
property. The U.S. Department of Justice (2004b) classifies crimes into two categories: prop-
erty and violent. This study included only property crime experiences because it is difficult to
inquire about violent crime experiences without risking emotionally upsetting participants.
The detailed types of property crime were drawn from the U.S. Department of Justice (ibid.).

"Perceived safety" is more conceptual and personal than actual safety. The primary method
employed in past studies to assess residents' perceptions of their safety has been to obtain
their views regarding their environment through standardized questions (see Normolye and
Foley, 1988; Rohe and Burby, 1988; Weidemann and Anderson, 1982). We analyzed various
methods of measuring residents' perceptions of safety in communities for this study. Wilson-
Doenges (2000) determined residents' perceived safety by asking questions such as "How
safe would you feel being out alone in your community during the day (or at night)?" Brunson,
et al. (2001) used questions such as "How safe do you feel in near-home space in the front and
the back of your building during the day and at night?"

To measure perceived safety, this study investigated residents' perceptions of safety in vari-
ous spaces in their multifamily housing communities. The level of residents' perceptions of
safety was measured using a five-point Likert-type response scale from one (not at all safe) to
five (very safe). We inquired into residents' perceptions of safety during the day and at night.
To measure residents' perceived safety, we used questions such as "Do you feel safe when
you walk alone through the parking lot during the day (or at night)?" and "Do you feel safe
when you exercise alone in the fitness center during the day (or at night)?"

The hierarchy of defensible space in apartment communities was considered when attempting
to understand residents' perceptions of safety in various spaces in their communities. To do
so, we drew on Newman (1973), who argued there are four categories of space in an apartment
complex: public, semipublic, semiprivate, and private. This taxonomy was applied to explore
residents' perceptions of safety in public, semipublic, and private spaces in multifamily hous-
ing communities. The space hierarchy outlined in Figure 1 is based on Newman's model
(1973:9).

This research also investigated residents' views concerning how their perceived safety might
be improved. Finally, we examined residents' demographic and socioeconomic characteristics
and levels of social interaction with neighbors to explore the relationship between perceptions
of safety and these variables.

FINDINGS AND DISCUSSION

Demographic Characteristics

Among the 207 respondents, 30.4% (n = 63) were living in gated communities,
resided in perceived-gated communities, and 44.9% (n = 93) called non-gated

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 139

home. Demographically, 63.6% of respondents were female, and 36.4% were


61% of respondents were in their 30s or younger, 48% of respondents wer
29.4% were African American. In terms of the residents of each unit, 41.5%
householders living alone, and 24.2% were female householders with childr
hold type varied significantly by type of community (p < .005). More fem
holders resided in non-gated communities.

Socioeconomic and Housing Characteristics

Approximately 97% of the respondents had at least a high school education;


them were college graduates or held higher degrees. This level of educa
allowed most respondents to understand the questionnaire without pro
employment rate of respondents was over 63%, including full-time employe
employed (5.4%). While their employment rate was similar to the national av
education level of the subjects was higher than the national average (83.9%
school graduates), based on the 2004 American Community Survey (U.
2004). Regarding annual income, 36 respondents did not disclose their i
reported an annual income under $30,000, and 13.5% reported earning more
median income of the responding group was approximately $40,000, lower
median income ($50,046).

A slightly larger share of the respondents (47.5%) lived in two-bedroom un


bedroom apartments (45.1%). Newman (1973) argued that the floor level on
dwell affects their perceptions of safety. This study delimited building he
story walk-up, however, so more than 50% of our respondents lived on the
building, while 35.8% resided on the second floor. Residents' average le
showed significant mean differences between the three types of community
munity residents had longer residence periods than gated community resi
length of residence was 20.4 months for gated community residents, 19.5
ceived-gated community residents, and 34 months for non-gated commun
4.406, p <. 05).

Crime Experience

The actual safety of residents was investigated through respondents' crime experiences while
living in their current properties. Table 1 shows that, out of 207 respondents, 40 (19.4%)
indicated they had had a crime experience, while 166 respondents had not. According to the
community type, 1 5.7% of perceived-gated community residents, 2 1 .7% of non-gated commu-
nity residents, and 19% of gated community residents reported experiencing crime. These
results suggest that non-gated community residents had relatively higher experiences of
crime than gated or perceived-gated community residents.

However, because the average length of residence was longer in non-gated communities, if we
estimated the crime rate across the average residence period for each type of community,
0.64% of non-gated community residents (= 21.7 34 months), 0.81% of perceived-gated
community residents (= 15.7 ^ 19.5 months), and 0.93% of gated community residents (= 19 ^
20.4 months) would experience crime each month. Another question assessing actual safety
was neighbors' crime experience. The number of residents who had heard about their neigh-
bors experiencing crime was 127 out of 207; the number was higher in gated communities than
in perceived-gated and non-gated communities. The difference was significant at the 0.005
level. These results do not support the general assumption that gated communities provide
safer residential environments than their non-gated counterparts.

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 140

TABLE 1 . Crime experience within apartment property territory.

Crime Experience Type of Community Total


Gated Perceived- Non-gated
gated

Cri^ Ye¡ 12 (19.0)* 8 (15.7) 20 (21.7) 40 (19.4)


experience No 51 (81.0) 43 (84.3) 72 (78.3) 166 (80.6)
Total 63 (100.0) 51 (100.0) 92 (100.0)** 206 (100.0)

Neighborhood Yes 48 (76.2) 33 (64.7) 46 (50.0) 127 (61.7)


crime No 15 (23.8) 18 (35.3) 46 (50.0) 79 (38.3)
experience Total 63 (100.0) 51 (100.0) 92 (100.0)** 206 (100.0)

Stolen Part of motor 27 (36.0) 21 (39.6) 24 (31.6) 72 (35.3)


items vehicle
Purse, wallet, cash 7 (9.3) 12 (22.6) 9 (11.8) 28 (13.7)
Electronics 9(12.0) 6(11.3) 10(13.2) 25(12.3)
Bicycle or parts 7(9.3) 2(3.8) 10(13.2) 19(9.3)
Computer-related 2(2.7) 1(1.9) 6(7.9) 9(4.4)
equipment
Jewelry, watch, keys 4(5.3) 2(3.8) 3(3.9) 9(4.4)
Clothing, luggage 3(4.0) 2(3.8) 2(2.6) 7(3.4)
Other 16(21.3) 7(13.2) 12(15.8) 35(17.2)
Total 75 (100.0) 53 (100.0) 76 (100.0) 204 (100.0)

* Numbers in parentheses indicate the frequency (%) of the crime experie


due to rounding; ** one respondent chose not to answer this question.

Crime involving motor vehicles was the most frequent proper


(Table 1). The most frequently stolen items were parts of mot
or cash (13.7%); electronics (12.3%); and bicycles or bicycl
stolen less frequently: computer-related equipment (4.4%); je
and clothing and luggage (3.4%). The "other" category in Tabl
other than those listed above, such as toys, plants, and recrea

Perceptions of Safety

The reliability of residents' perceptions of safety was te


Alpha values for perceptions of safety during the day and at
tively. Table 2 shows residents' perceived safety in priva
units), semipublic spaces (stairs of apartment buildings),
swimming pools, fitness centers, and outdoor mailbox space
each space varied according to the type of community. Resi
showed higher mean values in gated communities than in
communities. In particular, residents' perceptions of safety
lots and fitness centers) showed a statistical difference depen
Gated community residents felt safer in the public spaces th
community residents did.

These differences were more fully analyzed with Tukey's po


Tukey's test appear in Table 2 and suggest residents' percept
public, and private spaces of gated communities were signif
residents of non-gated communities. These results seem to s
of fully controlled gates and fences to residents' perceived sa

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 141

TABLE 2. Residents' perceptions of safety in near-home environments during the day and at

Perceptions Spaces Community Type n Mean3 F-valueb Tukey's Post-


Hoc Test

In the parking Gated 63 4.43 11.18 Group 1


lot Perceived-gated 51 4.27 Group 1
Non-gated 92 3.77 Group 2
Total 206 4.10

In the outdoor Gated 60 4.40 9.14 Group 1


swimming Perceived-gated 47 4.09 Groups 1 & 2
pool Non-gated 71 3.75 Group 2
Total 178 4.06
Perception of
safety during In the fitness Gated 5 5 4.51 18.36 Group 1
the day center Perceived-gated 45 4.07 Group 2
Non-gated 5 8 3.47 Group 3
Total 158 4.00

In the stairs Gated 60 4.35 9.66 Group 1


of apartment Perceived-gated 45 4.07 Groups 1 & 2
buildings Non-gated 86 3.71 Group 2
Total 191 3.99

In the current Gated 60 4.57 14.23 Group 1


community Perceived-gated 51 4.18 Group 1
Non-gated 90 3.72 Group 2
Total 201 4.09

At home Gated 63 4.30 10.74 Group 1


Perceived-gated 51 3.98 Group 1
Non-gated 92 3.54 Group 2
Total 206 3.88

In the parking Gated 62 4.00 10.41 Group 1


lot Perceived-gated 50 3.42 Group 2
Non-gated 92 3.16 Group 2
Total 204 3.48

In the outdoor Gated 5 7 3.86 9.11 Group 1


swimming Perceived-gated 46 3.22 Group 2
pool Non-gated 66 3.02 Group 2
Total 169 3.36
Perception of
safety at night In the fitness Gated 5 5 3.91 13.19 Group 1
center Perceived-gated 48 3.19 Group 2
Non-gated 54 2.87 Group 2
Total 157 3.33

In the stairs Gated 5 8 3.95 9.61 Group 1


of apartment Perceived-gated 45 3.47 Groups 1 & 2
buildings Non-gated 88 3.15 Group 2
Total 191 3.47

Around the Gated 61 3.87 8.93 Group 1


outdoor mail- Perceived-gated 50 3.34 Groups 1 & 2
box space Non-gated 91 3.02 Group 2
Total 202 3.36

In the current Gated 61 4.08 12.82 Group 1


community Perceived-gated 51 3.33 Group 1
Non-gated 90 3.10 Group 2
Total 202 3.46

a 1 = not at all safe, 2 = unsafe, 3 = neutra

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 142

TABLE 3. Explanatory variables for residents' perceptions of safety.

Category Subitem Pearson's Correlation Coefficients with Residents'


General Perceptions of Safety in Their Properties
During the Day At Night

Socioeconomic Length of residence -.114 -.069


variables Education level .147* .091
Annual income .286** .346**
Family size -.160* .290*

Architectural Dwelling floor level .149* .020*


variable Level of territoriality .354** .326**

Social interaction Neighborhood attachment .190** .265**

Perceptions of Parking lots .715** .759**


safety in specific Swimming pools .593** .727**
spaces Fitness centers .522** .728**
Outdoor mailbox spaces - .766**
Stairs of apartment buildings .734** .700**

* Correlation is significant at the 0.05 level; ** correlation is significant at the 0.01 level.

However, our results showed residents' perceptions of safety in perceived-ga


ties, particularly during the day, were not significantly different from those
gated communities. The results of residents' perceptions of safety in perceive
nities suggest the significance of fences and open gates as physical symbols
in residential properties. Even though the gates were not fully controlled, res
in perceived-gated communities than in non-gated communities, which prov
symbol of the property's territory. These results address the significant relat
territoriality provided by fences and gates in communities and residents' pe

Open gates providing less exclusive territoriality, however, functioned similar


trolled gates in terms of residents' perceptions of safety. This finding sugges
necessary to create exclusive territoriality by disconnecting communities fr
ones to secure residents' safety.

Explanatory Variables for Residents9 Perceived Safety

Including territoriality by the level of gates and fences, this study determined
lated with perceptions and sought to test correlations among demographic, soc
architectural variables addressed by previous studies of safety in rental prope

Table 3 provides the results of a correlation analysis of residents' genera


safety in their current apartment properties and several independent variab
suggest the gender of the respondent and whether the respondent was a hea
were significant in explaining residents' perceived safety during the day (F =
F = 4.4,/? < .05 respectively) and at night (F = 14.6,/? < .001 and F = 7.85,/? < .

Among the socioeconomic variables tested, residents' education level, ann


family size were significant variables in explaining their perceived safety. In p
size correlated negatively with residents' perceived safety during the day, wh
positively with it at night. This variable was also related to the number of ch
12 in the units. During our site visits, children were observed playing on outd
or roads during the day without parents watching. Nonetheless, parents stil
cerns about their safety.

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 143

TABLE 4. Step-wise multiple regression models explaining residents' general perceived safety
territories.

Model W- Adjusted R 2 ß

Dependent variable: General perceived safety in apartme


Modell .111 .105
Territoriality level .333****
Model 2 .165 .155
Territoriality level .338****
Neighborhood attachment .233***
Model 3 .205 .190
Territoriality level .252***
Neighborhood attachment .232***
Annual income .219**

Dependent variable: General percei


Modell .105 .101
Territoriality level .326****
Model 2 .167 .156
Territoriality level .334****
Neighborhood attachment .245***
Model 3 .197 .182
Territoriality level .257***
Neighborhood attachment .246***
Annual income .191*

Note. Territoriality level: 1 = n


standardized coefficients beta;

Territoriality was defined


nity, 2 = semi-exclusive te
fully controlled gated com
territoriality and resident
(p < .01) at night. The dw
during the day, but the co
subjects to three-story wa

Neighborhood attachment
bors, was also verified a
safety. Residents with st
properties during the day
the same correlation. Res
strongly correlated with
centers, and the stairs of
higher than .5 for all corr
properties and the perce
residents' perceived saf
significantly influences t
safety in public and semi
perceptions of safety in t

To determine the relation


safety in their apartmen
results in Table 4 illustrat
nual income have more
education level, family
safety, both during the d

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 144

statistically significant ß (standardized coefficients beta) values. With the


tachment variable, the models to explain residents' perceptions of safety du
night had higher R 2 values. These results show that physical territoriality
attachment among residents together improve residents' perceived safety
communities.

Additional Considerations for Improving Residents 9 Perceptions of Safety

We also considered nine additional items often offered to improve residents' perceptions of
safety: patrol service, bright lighting at night, a gate control system, fences around the prop-
erty, visual access to the local police, direct emergency buttons, 24-hour maintenance ser-
vices, open visual access to public spaces in the property, and other. The three most important
factors indicated by all respondents were guard service by a private patrol company (61.7%),
bright lighting at night (53.9%), and a gate control system at the main entrance (37.4%). Gated
community residents embraced patrol service (74.6%), bright lighting at night (41.3%), and
fences (31.7%). Perceived-gated community residents emphasized patrol service (62.7%),
bright lighting at night (56.9%), and a gate control system (45.1%). Non-gated community
residents highlighted bright lighting (60.9%), patrol service (52.2%), and a gate control system
(39.1%). These results suggest gate control systems and fences should be accompanied by
various additional steps such as managerial services for residents (e.g., patrol service), safety
improvement support from outside the property (e.g., access to police, direct emergency
buttons), and physical interventions for safety improvement (e.g., lighting at night, open
visual access to public spaces in the property) in order to improve residents' perceptions of
property safety.

IMPLICATIONS AND CONCLUSION

Safety should be a critical issue for urban renters because the apartment p
they live characteristically contain more people who are strangers to one
generally more vulnerable to crimes (Newman, 1996:25). This study targete
properties as its main subject and examined various ways that have been p
renter residents' perceived and actual safety in those developments.

Because much of the literature indicates that perceptions of safety and cri
fundamentally related to territoriality, our initial research question was
residents felt safer in gated communities that provide exclusive territorial
trolled gates and fences than in those that do not. After considering the le
and traffic controls, we proposed three types of apartment properties -
gated, and non-gated communities - and examined the correlation be
territoriality and residents' perceptions of safety.

Our results support the idea that people's perceived safety and crime exp
mentally related to territoriality, as the literature has indicated. Gated com
perceived themselves as safer than non-gated community residents d
differences during the day and at night were statistically significant.

However, general perceptions of safety in apartment properties were not s


ent between residents of gated and perceived-gated communities. Residen
safety in public, semipublic, and private spaces were also not significantly
these two resident groups (gated and perceived-gated). These findings sug
territoriality provided by gated communities does not improve resid
safety. Our assessment of residents' experiences with crime in their commu
supported this claim. Residents of gated communities had higher crime rate

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 145

residents of perceived-gated and non-gated communities. Therefore, we co


controlled exclusive territoriality is unnecessary to improve the actual and p
residents.

Instead of exclusive territoriality, this study's results suggest that even without gates and
fences, apartment communities can provide perceived territoriality. Controlling unnecessary
traffic to and through apartment properties could be accomplished by narrow community
entrances and internal roads that provide advance warning of vehicles approaching from
outside. Cul-de-sacs in the internal roads of apartment communities can prevent pass-through
traffic, as Newman (1996) earlier suggested. Instead of steel fences, wooden fences surround-
ing an apartment property can provide both territoriality and a feeling of openness to resi-
dents.

The statistical evidence outlined here supports the conclusion that perceived safety in public
and semipublic areas should be guaranteed for improving residents' general perceptions of
safety in their apartment territories. That is, one way to ease residents' fears of crime in their
apartment complexes is to first address their fears of crime in the public and semipublic spaces
of their buildings.

Another powerful variable affecting residents' perceived safety was neighborhood attach-
ment. Newman (1973) and Taylor, et al. (1984) also indicated the importance of residents'
socialization in reducing their fear of crime. Blakely and Snyder (1999) noted the important role
of social interaction with neighbors for easing residents' fear of crime in their communities.
Apartment community dwellers are usually renters, not owners. This fact typically makes them
less informed about community issues and their neighbors than owners would be. However,
residents' neighborhood attachment proved very significant in preventing crime and improv-
ing perceived safety in the apartment communities we examined. Residents in apartment com-
munities thus need to pay attention to community issues and neighbors' needs. They can
enhance social contacts with their neighbors by being involved in community meetings or
events. Management groups should actively work to provide residents with opportunities to
meet and communicate with their neighbors and to encourage continuing social contacts
among them.

Other considerations in addition to physical territoriality and neighborhood attachment


should never be overlooked in creating safe apartment properties for residents. These include
patrol services, bright lighting at night, direct emergency buttons or visual access to the local
police, and visual access to public spaces on the properties.

As a general finding, we suggest that, to improve residents' perceived and actual safety in
multifamily housing properties in urban areas, comprehensive approaches aimed at imple-
menting architectural interventions, adopting managerial efforts, and securing sustained resi-
dent interaction are necessary. The design considerations and suggestions highlighted in this
study may also be applied to different types of residential settings, such as single-family
neighborhoods and planned communities. In fact, past studies have identified a relationship
between people's safety (both perceived and actual) and explanatory variables in single-
family neighborhoods and public housing authorities (Blakely and Snyder, 1997; Brunson, et
al. , 200 1 ; Newman, 1 996). This study confirms those earlier findings and their primary recom-
mendations.

REFERENCES

Atkinson R (2003) Fortress UK? Gated communities, the spatial revolt of th


space trajectories of segregation. Paper presented at the Gated Com

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 146

Building Social Division or Safer Communities? conference. Glasgo


September).

Blakely EJ, Snyder MG (1997) Divided we fall: Gated and walled communiti
States. In N Ellin (Ed.), Architecture of fear. New York: Princeton A
Press, pp. 101-114.

Blakely EJ, Snyder MG (1998) Forting up: Gated communities in the United
Architectural and Planning Research 15(l):61-72.

Blakely EJ, Snyder MG (1999) Fortress America : Gated communities in th


Washington, DC: Brookings Institution Press.

Brunson L, Kuo FE, Sullivan WC (2001) Resident appropriation of defensibl


housing: Implications for safety and community. Environment
33(5):626-652.

Fisher B, Nasar JL (1992) Fear of crime in relation to three exterior site f


refuge, and escape. Environment and Behavior 24(l):35-65.

Fowler FJ, Jr., Mangione TW (1 986) A three-pronged effort to reduce crime


The Hartford experiment. In DP Rosenbaum (Ed.), Community crim
Does it work? Newbury Park: Sage Publications, pp. 87-1 18.

Funk LM, Allan DE, Chappel NL (2007) Testing the relationship between
perceived neighborhood safety. Environment and Behavior 39(3):33

Goix RL (2003) "Gated communities" as territories apart in southern Califo


level of socio-spatial discontinuities created by walls and gates. Pa
at the Gated Communities: Building Social Division or Safer Comm
ference. Glasgow, UK (18-19 September).

Halberg M (2001) Gated communities: Do they raise residents' expectat


liability for associations? Journal of Community Association Law
www.caionline.org/articles/index.cfm?RLID=3029. Site accesse
2004.

Holzman HR, Kudrick TR, Voytek KP (1996) Revisiting the relationship bet
architectural design: An analysis of data from HUD's 1994 sur
housing residents. Journal of Policy Development and Research 2(

Kim SK (2006) The gated community: Residents' crime experience and perce
behind gates and fences in the urban area. Unpublished PhD disser
A&M University, College Station.

Macionis JJ, Parrillo VN (2004) Cities and urban life. Upper Saddle River, N
tion, Inc.

Moran R, Dolphin C (1986) The defensible space concept: Theoretical and operational explica-
tion. Environment and Behavior 18(3):396-416.

Newman O (1973) Defensible space: Crime prevention through urban design. New York:
Macmillan.

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 147

Newman O (1996) Creating defensible space, http://www.huduser.org/publi


defensi.html. Site accessed 1 June 2004.

Normoyle JB, Foley JM (1988) The defensible space model of fear and elderly public housing
residents. Environment and Behavior 20(l):50-74.

Perkins DD, Taylor RB (1996) Ecological assessments of community disorder: Their relation-
ship to fear of crime and theoretical implications. American Journal of Commu-
nity Psychology 24( 1):63-107.

Pickett JP (Ed.) (2006) Gated community. The American heritage dictionary of the English
language , 4th edition, http://dictionary.reference.com/browse/gated commu-
nity. Site accessed 5 September 2009.

Rohe WM, Burby RJ (1988) Fear of crime in public housing. Environment and Behavior 20(6):
700-720.

Taylor FB, Gotteredson SD, Brower S (1984) Block crime and fear: Defensible space, local
social ties, and territorial functioning. Environment and Behavior 21(4):303-33 1 .

Tijerino R (1998) Civil spaces: A critical perspective of defensible space. Journal of Architec-
tural and Planning Research 15 (4): 32 1-337.

U.S. Census Bureau (2002) American housing survey for the United States: 2001 . http://www
.census.gov/prod/2002pubs/h 150-01. pdf. Site accessed 1 December 2005.

U.S. Census Bureau (2004) 2004 American community survey, http://www.census.gov/acs/


www/. Site accessed 1 December 2005.

U.S. Census Bureau (2007) American housing survey for the United States: 2007. http://www
.census.gov/housing/ahs/files/ahs07/hl 50-07.pdf. Site accessed 15 October 2009.

U.S. Department of Justice (2004a) Additional crime facts at a glance. http://www.ojp.usdoj


.gov/bjs/gvc.htm. Site accessed 21 February 2004.

U.S. Department of Justice (2004b) Crime and the nation s households , 2003 (NCJ Publication
no. 206348). Washington, DC: U.S. Department of Justice.

U.S. Department of Justice (2004c) Crime in the United States: 2003. http://www.fbi.gov/about-
us/cjis/ucr/crime-in-the-u.s/2003/toc03.pdf. Site accessed 30 March 201 1.

Weidemann S, Anderson JR (1982) Residents' perceptions of satisfaction and safety: A basis


for change in multifamily housing. Environment and Behavior 14(6):695-724.

Wilson-Doenges G (2000) An exploration of sense of community and fear of crime in gated


communities. Environment and Behavior 32(5):597-61 1 .

Additional information may be obtained by writing directly to Dr. Kim at School of Planning,
Design, and Construction, 20 IB Human Ecology Building, Michigan State University, East
Lansing, MI 48824, USA; email: kimsk@msu.edu.

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms
Journal of Architectural and Planning Research
29:2 (Summer, 2012) 148

ACKNOWLEDGMENTS

The authors wish to acknowledge the contributions of Dr. Charles Graham, Dr. Robin Abram
Giusti for their advice and comments on the primary work that was a foundation of this art

AUTOBIOGRAPHICAL SKETCHES

Dr. Suk-Kyung Kim is Assistant Professor of Interior Design in the School of Planning, Design,
at Michigan State University. She received a PhD in Architecture from Texas A&M Univer
gated community and people's perceptions of safety have been her research interest for a long
several publications and presentations both in Korea and the U.S. relevant to this topic. Her re
includes environmental behavioral studies focusing on place attachment to home environment
for children and seniors, new urbanism, and interior space planning and analysis.

Andrew D. Seidel is Professor and Chair of the School of Environmental Planning at the Univer
British Columbia. He has taught at exceptional universities from Harvard to Yonsei in Seoul. H
editorial work over nearly 37 years have produced more than nine books and 50 articles. This w
interdisciplinary and bridging among fields from policy research to graphic design, is best known
of the first refereed scholarly research journal in architecture and linking architecture and planni
wayfinding, knowledge and research utilization (currently referred to as evidence-based practic
areas.

Manuscript revisions completed 18 April 2012.

This content downloaded from 1.23.108.87 on Thu, 05 Dec 2019 16:07:46 UTC
All use subject to https://about.jstor.org/terms

You might also like