ABSTRACT 
 
 
 
Title of thesis:             TRANSFORMING VACANT LOTS: INVESTIGATING AN  
ALTERNATIVE APPROACH TO REDUCING FEAR OF 
CRIME  
 
 
Alaina M. De Biasi, Master of Arts, 2015 
 
 
 
Thesis directed by:       Professor Jean McGloin 
 Department of Criminology and Criminal Justice 
 
 
 
Vacant lots deserve criminological attention insofar as their disorderly conditions 
create opportunities for a host of negative outcomes including “fear of crime.” The 
present study considers whether incorporating fundamental standards of Crime 
Prevention through Environmental Design (CPTED) into traditional urban greening 
practices of vacant lots provides added value with regard to fear of crime above and 
beyond the traditional endeavor. This study conducted an experiment (N=523) from a 
sample of undergraduate students.  Research participants were asked to report their level 
of fear of crime in regards to one of three randomly assigned computer-adjusted images: 
1) A disorderly lot; 2) A traditional greened lot; and 3) A CPTED lot. This study found 
that on average participants who viewed a CPTED lot had lower levels of fear of crime 
than all other participants. This study discusses the implications of this finding for future 
research. 
 
 
 
TRANSFORMING VACANT LOTS: INVESTIGATING AN ALTERNATIVE 
APPROACH TO REDUCING FEAR OF CRIME   
 
 
 
 
 
 
By 
 
Alaina M. De Biasi  
 
Thesis submitted to the Faculty of the Graduate School of the University of Maryland, 
College Park in partial fulfillment of the requirements for  
the degree of Master of Arts  
2015 
 
 
 
Advisory Committee:  
Professor Jean McGloin, Chair  
Assistant Professor Lauren Porter 
Associate Professor Min Xie 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
© Copyright by 
Alaina M. De Biasi  
2015 
 
 
 
 
 
 
 
 
 
 
 
 
ii 
 
Acknowledgements 
I would like to thank Dr. Jean McGloin for her support and guidance and for always 
giving me the best advice. I would also like to thank Dr. Min Xie and Dr. Lauren Porter 
for their guidance and helpful critiques throughout this process. In addition, my cohort 
has been a constant source of support. The individuals of my cohort have played an 
instrumental role in my graduate school experience. I am thankful for the opportunity to 
have studied and grown as a researcher alongside these amazing individuals.  Finally, I 
would like to thank my friends, family members and, in particular, James Podges for their 
unwavering love and support.    
 
 
 
 
 
 
 
 
 
 
 
 
iii 
 
Table of Contents 
 
Chapter 1: Introduction ....................................................................................................... 1 
Chapter 2: Literature Review .............................................................................................. 6 
Chapter 3: Methods ........................................................................................................... 30 
Chapter 4: Results ............................................................................................................. 57 
Chapter 5: Discussion & Conclusion ................................................................................ 60 
Appendix A ....................................................................................................................... 68 
References ......................................................................................................................... 79 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
iv 
 
List of Tables  
Table 1- CPTED Standards ........................................................................................... 13 
Table 2-  Fear of Crime Descriptive Statistics  ............................................................. 38 
Table 3- Polychoric Correlation Matrix ........................................................................ 43 
Table 4- Eigenvalues ..................................................................................................... 45 
Table 5- Fear of Crime Index Descriptive Statistics ..................................................... 48 
Table 6- Adjustments  ................................................................................................... 50 
Table 7- Lumpkin's Abbreviated Locus of Control Scale ............................................. 54 
Table 8- Indivdual-level Charcteristics Descriptive Statistics ...................................... 55 
Table 9- Assesment of Equivalence  ............................................................................. 57 
Table 10- OLS Regression ............................................................................................ 58 
Table 11- OLS Regression with No Controls ............................................................... 93 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
v 
 
List of Figures 
Figure 1 – Orignal Image of the Vacant Lots ................................................................ 69 
Figure 2- Template Image   ........................................................................................... 69 
Figure 3-  Disorderly Lot .............................................................................................. 70 
Figure 4- Greened Lot  .................................................................................................. 70 
Figure 5- CPTED Lot .................................................................................................... 71 
Figure 6- Scree Plot  ...................................................................................................... 71 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1 
 
Chapter 1: Introduction  
Vacant lots, defined as unkempt parcels of land with no buildings on them, are 
common features of city landscapes in the United States (U.S.) (Bowman & Pagano, 
2000; Burkholder, 2012; United States Environmental Protection Agency, 2013). In fact, 
on average 15% of a city’s land in the U.S. is vacant (Bowman & Pagano, 2000).1 Within 
cities, vacant lots are typically not only dominant features within their inner cores, but 
also in surrounding neighborhoods. For example, in some neighborhoods in Detroit, 
Michigan, such as Westwood Park, vacant lots comprise a greater proportion of land area 
than they do in the inner-city (Data Drive Detroit, 2015).  
Vacant lots deserve criminological attention; their disorderly conditions create 
opportunities for a host of negative outcomes including “fear of crime” (Spelman, 1993; 
Wilson & Kelling, 1982). Fear of crime has been demonstrated to negatively influence an 
individual’s social, mental and physical health as well as, on a larger scale, the vitality of 
his or her community (Green, Gilbertson & Grimsley, 2002; Morrall, Marshall, Pattison 
& Macdonald, 2010; Ross & Mirowsky, 2001; Wilson & Kelling, 1982). In light of the 
detrimental outcomes associated with vacant lots as well as their prevalence in the U.S, 
understanding the connection between vacant lots and fear of crime is an important 
endeavor (Bowman & Pagano, 2000). Research on this connection must develop more 
comprehensive methods to manage the disorderly conditions of vacant lots and to 
capitalize on the opportunity to reduce fear of crime. 
                                                          
1  In order to minimize definitional disparities, Pagano and Bowman (2000) defined vacant land as 
including “publicly-owned and privately-owned unused or abandoned land or land that once had structures 
on it, but also the land that supports structures that have been abandoned, derelict, boarded up, partially 
destroyed, or razed” (p. 2).  Although this figure does not solely focus on vacant lots, as defined by this 
study, it is able to provide a rudimentary understanding of their prevalence in U.S. cities.  
2 
 
A common method city officials in Baltimore, New York City, Philadelphia and 
Detroit use to manage vacant lots in the inner-city as well as in surrounding 
neighborhoods is to transform them into green space, called “urban greening” (Baltimore 
Greenspace, 2013; Greening of Detroit, 2014; Pennsylvania Horticultural Society, 2014; 
Urban Green Council, 2014). The primary objective of urban greening is to improve the 
conditions of vacant lots by taking them from disorderly to orderly, which should then 
promote quality of life. To achieve this objective, urban greening efforts across the 
country commonly rely on a set of traditional (or standard) techniques referred to as 
“traditional urban greening practices.” These traditional practices include planting grass 
and low-maintenance plants (e.g., trees) as well as focus on efforts to maintain the 
environment referred to as maintenance. Maintenance efforts involve removing debris 
and overgrowth, as well as other forms of physical disorder.  Physical disorder includes 
environmental features that reflect neighborhood dilapidation such as graffiti, trash and 
other debris (Skogan, 1990). It is viewed as an environmental cue signaling unsafe places 
(Doran & Lees, 2005; Hanyu, 1993), which in turn can increase fear of crime (Cinar & 
Cubukcu, 2012; Fisher & Nasar, 1992; Nasar & Jones, 1997; Warr, 1990). Despite urban 
greening’s attention to maintenance, thus far, no one has examined whether urban 
greening fully capitalizes on the opportunity to reduce fear of crime.  
Urban greening’s ability to transform disorderly environments through its focus 
on maintenance fits nicely within the framework of Crime Prevention though 
Environmental Design (CPTED), which aims to reduce crime and fear of crime (Pain, 
2000; Perkins, Wandersman, Rich, & Taylor, 1993; Schneider & Kitchen, 2007). CPTED 
attempts to achieve these reductions through the implementation of environmental design 
3 
 
standards while maintaining the underlying goal, like that of urban greening, of 
improving the quality of life of residents (Crowe, 1991; Perkins, Wandersman, Rich & 
Taylor, 1993; Schneider & Kitchen, 2007).  
Unlike traditional urban greening practices, CPTED provides instruction on how 
to simultaneously enforce numerous design standards in addition to efforts to maintain 
the environment. These include the four most dominant CPTED standards: territoriality, 
surveillance, access control and maintenance. Due to their complementary nature, these 
design standards are more commonly implemented together as a package than in isolation 
of one another (Bureau of Justice Assistance, 2013). In fact, the complementary nature of 
these design standards makes it extremely difficult to untangle the ability of each 
standard to impact crime and fear of crime (Schneider & Kitchen, 2002).  
In addition, these standards are able to collectively influence an individual’s 
perception of the environment which helps inform his or her level of fear of crime. 
Consequently, it is reasonable to believe that the collective enforcement of these design 
standards is able to affect an individual’s level of fear of crime beyond that which could 
be achieved by the enforcement of one design standard alone. Although urban greening 
fits nicely within the framework of CPTED, it has yet to explicitly incorporate CPTED 
standards into traditional urban greening practices. 
Incorporating CPTED standards into traditional urban greening practices is a 
worthwhile and relatively low-cost alteration to investigate. CPTED has the ability to 
transform vacant lots beyond the rudimentary maintenance efforts that are the focus of 
traditional urban greening practices. For example, urban greening practices have the 
tendency to create unstructured spaces that are not formally organized. Unstructured 
4 
 
spaces may include, for instance, trees that are planted too closely together which create 
hiding places for offenders and impair individuals from quickly and easily escaping the 
advances of a potential offender. Such qualities may lead individuals to perceive 
themselves to be vulnerable to victimization. This perception ultimately mitigates urban 
greenings’ ability to reduce fear of crime beyond that which is achieved through its focus 
on maintenance alone.  CPTED, when fully incorporated into the greening process as a 
package, is expected to result in greater reductions of fear of crime by ensuring that space 
is not only maintained but also actively upholds additional design standards which 
collectively work together to guard against these, and other, qualities of unstructured 
spaces. By not collectively enforcing CPTED standards, it is likely that many cities in the 
U.S. are failing to capitalize on the opportunity to reduce fear of crime within 
communities in their conversion of vacant lots.  
The present study will provide some preliminary commentary on whether 
greening incorporating CPTED standards into traditional urban greening practices might 
pay important and substantive dividends above and beyond traditional practices. To 
explore this uncultivated research area, this study will randomly assign subjects to one of 
three conditions. Participants will either receive an image of a disorderly lot (i.e., a vacant 
lot that is left in a disorderly state) a greened lot (i.e., a vacant lot that is greened 
according to the traditional practices previously mentioned) or a CPTED lot (i.e., a vacant 
lot that incorporates CPTED standards into the greening process). Using these images, 
this study will test the assumption that incorporating CPTED standards into traditional 
urban greening practices of vacant lots may provide added value with regard to fear of 
crime above and beyond the traditional endeavor. CPTED standards will be evaluated as 
5 
 
a package rather than in isolation of one another. In the end, this study will provide an 
assessment on whether incorporating CPTED standards into the traditional urban 
greening practices described is a worthwhile effort to pursue in regards to achieving 
greater reductions of fear of crime. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6 
 
 
Chapter 2: Literature Review  
Crime Prevention through Environmental Design (CPTED): General Background  
CPTED uses a framework of environmental design standards with the goals of 
achieving a reduction in the incidence and fear of crime as well as an improvement in 
residents’ quality of life (Crowe, 2000). In particular, CPTED relies on its ability to 
influence individuals’ perceptions of the environment which are facilitated by the 
enforcement of its environmental design standards: territorially, surveillance, access 
control and maintenance. This study discusses these design standards, which represent 
the most dominant standards in CPTED literature, in detail in a later section.  These 
environmental design standards work together to influence offenders’ and non-offenders’ 
(i.e., potential victims’) perceptions of the environment differently. From an offender 
standpoint, these perceptions affect decisions to commit criminal acts. Alternatively, 
from a non-offender’s or potential victim’s standpoint, these perceptions affect fear of 
crime. The present section provides an overview of CPTED’s historical development 
which highlights the mechanisms and environmental conditions (or cues) that influence 
criminal behavior and the formulation of CPTED standards. Lastly, this section begins 
building a rudimentary foundation to explore the relationship between the environment 
and fear of crime which is the focus of a later section and of this study. 
CPTED was first introduced in Jeffery’s (1971) publication of Crime Prevention 
through Environmental Design. This work was largely conceptual in nature. In it, Jeffery 
(1971) emphasized the role of the physical and social environment to produce positive 
(pleasurable) and negative (painful) experiences for offenders, a concept borrowed from 
Skinner’s (1938) stimulus-response model. Under Skinner’s (1938) perspective, 
7 
 
offenders make choices based on their expectations of the positive and negative outcomes 
that could result from their decisions. Following this logic, Jeffery (1971) believed that it 
was possible to alter the environment to enhance the likelihood of conforming behavior 
while simultaneously mitigating criminal behavior. In particular, he advocated altering 
the physical environment as a means to decrease the positive expectations associated with 
decisions to offend and in doing so deter criminal behavior.  In this way, Jeffery’s (1971) 
work mirrored earlier analyses conducted by Jane Jacobs (1960), Elizabeth Wood (1961) 
and Scholomo Angel (1968) which suggested that alterations to certain physical design 
features of urban environments could deter potential offenders from committing crime. 
Later, Jeffery (1977) emphasized that the environment never influenced behavior 
directly. Rather, environmental cues trigger individuals’ to formulate perceptions of their 
surroundings that are used in cognitive processes that influence behavior.  For example, 
an environmental cue may signal to offenders that criminal acts will likely go undetected 
and unpunished. In this circumstance, offenders’ positive expectations of their decisions 
to offend (e.g., monetary reward) would likely outweigh their expectations of any 
negative outcomes (e.g., punishment). Therefore, for offenders, this perception would 
likely motivate criminal behavior. Unfortunately, Jeffery (1971, 1977) does not elaborate 
on the underlying mechanism, fear of crime, which motivates the behaviors of non-
offenders. Other scholars, such as Newman (1972), fill this void.   
Taking Jeffery’s (1971) concept, Newman (1972) was one of the first scholars to 
discuss CPTED in practical terms. Unlike Jeffery (1971) who focused primarily on 
developing a conceptual backdrop for CPTED, Newman (1972) elaborated on CPTED by 
providing actual standards for what the physical environment should resemble in order to 
8 
 
reduce not only crime but fear of crime and to promote quality of life.  Similar to Jeffery 
(1971) and his predecessors, Newman (1972) argued that it was possible to design 
physical environments to decrease opportunities for crime as well as fear of crime by 
affecting the behavior of offenders and non-offenders. In particular, he suggested the 
following design standards were crucial and critical to this objective: territoriality, 
surveillance, image and milieu (Newman, 1972). Newman (1972) believed that a good 
design encourages residents to defend their communities and, in doing so, creates safer 
areas by signaling to offenders that criminal acts will not go undetected or unpunished. 
To this end, he advocated applying these standards in the creation of physical layouts of 
residential areas.  For example, Newman (1972) emphasized the importance of design 
layouts that allowed residents to be able to regularly and effortlessly observe their 
surroundings. This ability deters crime and decreases fear of crime by allowing residents 
the opportunity to quickly and easily identify offenders which in turn increases offenders’ 
risk of detection and likelihood of punishment while allowing non-offenders the security 
and comfort of avoiding unanticipated attacks.   
Since the presentation of Newman’s (1972) standards, several other scholars have 
helped shape the construction of CPTED standards. The contributions of these scholars 
have resulted in the development of arguably the four most dominant environmental 
design standards: territoriality, surveillance, access control and maintenance. Appleton’s 
(1975) prospect-refuge theory and Wilson and Kelling’s (1982) broken windows thesis 
are among the most significant contributions of these scholars. These contributions 
expand on Newman’s (1972) assessment of CPTED by further describing certain 
9 
 
physical conditions of the environment (i.e., environmental cues) that influence not only 
criminal behavior but also fear of crime.   
In prospect-refuge theory, Appleton (1975) describes environments that facilitate 
feelings of safety and in doing so reduce fear of crime and the likelihood of criminal 
behavior. These environments include the following: 1) Environments that allow 
potential victims to have a clear view of their surroundings (prospect); 2) Environments 
that lack places for offenders to hide (refuge); and 3) Environments that allow 
opportunities for potential victims to easily and quickly escape the advances of an 
offender. However, in broken windows thesis, Wilson and Kelling (1982) describe an 
element that mitigates the effects of these environments. This element is physical 
disorder: environmental features that reflect neighborhood dilapidation such as graffiti, 
trash and other debris (Skogan, 1990). Physical disorder impairs prospect, provides 
refuge for offenders and mitigates the ability of individuals to quickly and easily escape 
the advances of potential offenders. In particular, physical disorder provides concealment 
for the perpetration of physical and sexual assaults and is also useful in providing 
opportunities for other attacks where surprising an unsuspecting victim benefits the 
attacker, such as in the case of robbery.  
 In addition, perceptions of physical disorder are a pivotal part of the broken 
windows process, a debilitating cycle that facilitates crime, fear of crime and the spread 
of disorder (Wilson & Kelling, 1982).  Physical disorder is more likely to be perceived 
(or recognized) in places where it is most prevalent, such as in vacant lots. Once 
perceived, physical disorder helps promote the belief that a community is free from the 
constraints of social control and that criminal acts will go unpunished.  This belief 
10 
 
facilitates increases in residents’ fear of crime and the spread of physical disorder as well 
as encourages the proliferation of crime and unwanted public behaviors that are 
threatening but may or may not be unlawful (Keizer, Lindenberg, & Steg, 2008; Wilson 
& Kelling, 1982). Included among these behaviors is harassment, defined as unwelcome 
verbal and nonverbal behavior that occurs in public by a stranger or group of strangers. A 
bystander is more likely to become the victim of harassment in areas with high levels of 
physical disorder because in such places physical disorder is more likely to be perceived 
by offenders and non-offenders alike as a breakdown of the regulation, both formally and 
informally, of individual and group behavior. In essence, in areas with high levels of 
physical disorder, such as in and around vacant lots, there are less social inhibitions to 
discourage unwanted behaviors like harassment.  Contrastingly, the beliefs that a 
community is free from the constraints of social control and that criminal acts will go 
unpunished are not likely to be enforced in places that have little physical disorder. 
Rather, such places foster the perception that they are under the surveillance and 
protection of the community and that unwanted behaviors will be detected and offenders 
punished. These perceptions help decrease residents’ fear of crime, stifle the spread of 
physical disorder and strengthen the barriers against crime as well as unwanted public 
behaviors. 
In light of Appleton’s (1975) and Wilson and Kelling’s (1982) contributions, 
CPTED has developed its standards to incorporate many of the principles of prospect-
refuge theory and to include efforts to minimize physical disorder. These efforts have 
resulted in territoriality, surveillance, access control and maintenance becoming dominant 
CPTED standards.  
11 
 
Today, CPTED is often incorporated into community-based crime prevention 
strategies. Foremost among these strategies is situational crime prevention. Like CPTED, 
situational crime prevention focuses on reducing criminal opportunities. With strong 
roots in rational choice theory, situational crime prevention seeks to diminish criminal 
opportunities by making crime more difficult and less rewarding and by increasing an 
offender’s risk of detection (Clarke, 1995; Cornish & Clarke, 1986). In doing so, 
situational crime prevention strategies often involve altering the physical environment by 
applying CPTED standards.  
However, situational crime prevention represents a broader approach than 
CPTED. Whereas CPTED focuses on the design of the physical environment, situational 
crime prevention may incorporate legal and management solutions into its prevention 
strategies. These legal and management solutions may include enforcing laws (e.g., legal 
solution) or rules or procedures (e.g., management solution). For example, motorcycle 
thefts have declined in places where motorcyclists by law are required to wear helmets 
(Mayhew, Clarke & Elliott, 1989; Mayhew, Clarke, Sturman & Hough, 1976). In such 
places, rational offenders are dissuaded from stealing motorcycles because doing so 
without wearing a helmet-something the typical offender does not carry with him or her- 
will increase their likelihood of detection and arrest. In addition, research shows that 
refund policies, a set of rules retail merchants establish to manage the return or exchange 
of unwanted goods, can result in a reduction of refund fraud at retail stores (Challinger, 
1997) 
Contrary to most CPTED interventions, situational crime prevention strategies are 
highly focused on preventing specific forms of crime and, consequently, often require 
12 
 
problem-solving by a variety of public and private agencies. Furthermore, unlike CPTED, 
situational crime prevention is not focused on improving quality of life within 
communities by reducing fear of crime. Despite these differences, CPTED is often 
creatively applied in many situational crime prevention strategies (Glazer, 1987; La 
Vigne, 1997; Painter & Farrington, 1997; Poyner & Webb, 1987). For example, Poyner 
and Webb (1987) studied the prevalence of robberies and purse-snatchings at Lisson 
Green, a London public-housing estate. They discovered that restricting access to and 
egress from buildings by removing walkways that linked them significantly reduced the 
number of robberies and purse-snatchings (Poyner & Webb, 1987). This situational crime 
prevention strategy takes advantage of CPTED’s access control standard.  
CPTED Dominant Standards: Territoriality, Surveillance Maintenance & Access Control 
In order to improve quality of life, which requires reducing fear of crime, CPTED 
adheres to certain environmental design standards. As previously mentioned, the most 
prominent of these standards are territoriality, surveillance, access control and 
maintenance (Cozens, Saville& Hillier, 2005; Crowe, 1991; Parnaby, 2006). Ideally, 
these standards “act in concert to use physical attributes to separate public, public-private 
and private space, to define ownership and define acceptable patterns of usage, in 
addition to promoting opportunities for surveillance” (Cozens, Saville & Hillier, 2005, p. 
331). Due to this complementary relationship, these design standards are commonly 
employed and evaluated together as a package. Table 1 provides each design standard’s 
underlying objective and recommended adjustments to the environment. This table is 
followed by a more detailed description of each CPTED standard.  
 
13 
 
------------------------------- 
             Insert Table 1 about here 
------------------------------ 
 “Territoriality” aims to promote a sense of ownership and facilitate proprietary 
concern (Cozens, Saville & Hillier, 2005; Crowe, 1991).  Personalization design 
elements, such as artwork and landscaping, are often used to achieve these goals. 
Features such as fences, landscape design and signage also help define the difference 
between private and public property and in doing so are thought to facilitate a sense of 
ownership (Hedayati, Abdullah, Razak & Maghsoodi, 2012a, 2012b; Sakip, Johari & 
Salleh, 2012; Sutton, Cherney & White, 2008). 
“Surveillance” promotes an individual’s ability to have clear views of his or her 
surroundings. Ultimately, this ability minimizes the availability of hiding places that 
offenders can use in wait of an innocent bystander. To promote surveillance, see-through 
fencing and lighting are often utilized; the placement of plants and walkways are also 
considered (Cozens, Saville & Hillier, 2005; Crowe, 1991). Moreover, of utmost 
importance to upholding CPTED’s surveillance standard is to ensure that physical 
disorder does not obstruct a user’s visibility.  
 “Access control” involves constructing a design that manages the safe usage and 
movement of individuals into and out of a space, with a focus on decreasing opportunities 
for illegitimate use (Paranaby, 2007). Places with defined routes and entrances and exits 
that provide safe and convenient movement are valued. Other design features of access 
control include low fencing or signage. These elements are classified as symbolic barriers 
because they do not pose real obstacles to access and egress but, through increasing 
perceptions of risk, may dissuade the misuse of a space (Perkins et al., 1993). 
14 
 
Lastly, “maintenance” refers to all efforts required to uphold a tidy and orderly 
appearance of a space. This may involve removing debris and graffiti as well as 
maintaining plants.  Maintenance is an especially important standard to uphold because 
without it territoriality, surveillance and access control are critically jeopardized. For 
example, maintenance promotes unobstructed sightlines required to ensure visibility. It 
also helps create an attractive public space that is perceived to be under the care of an 
individual or a group of individuals. In this way, maintenance helps facilitate the image 
of ownership related to CPTED’s territoriality concept.  Additionally, maintenance 
supports CPTED’s access control concept by allowing unhindered access and egress.  
The Impact of CPTED on Crime and Fear of Crime: An Empirical Review  
The primary argument of CPTED is that its alterations to the environment, guided 
by its design standards, are able to achieve reductions in crime as well as fear of crime 
within communities. In both situations, the argument is that CPTED affects individuals’ 
perceptions of the physical environment. From an offender’s perspective, these 
perceptions affect rational offenders’ decisions to commit criminal acts. Alternatively, 
from a non-offender’s or potential victim’s perspective, these perceptions affect fear of 
crime.  
 The impact of CPTED on crime, as opposed to fear of crime, in communities is 
the focus of most empirical examinations of CPTED (Armitage, 2000; Casteel & Peek-
Asa, 2000; Cozens, Saville & Hillier 2005; Hendricks, Landsittel, Amandus, Malcan, & 
Bell, 1999; Morrow & Hutton, 2000; Welsh & Farrington, 2004). However, many of 
these studies utilize (experimentally) weak designs which call into question the veracity 
of their findings (Bickman, Maltz & Lavrakas, 1977). For example, many CPTED studies 
15 
 
use a pre-posttest intervention design without a control group (Clifton, 1987; Erickson, 
1998; Figlio &Aurand, 1991; Hunter, 1990). In doing so, the impact CPTED has on 
crime cannot be evaluated separately from the influence of extraneous risk factors such as 
general crime trends (Casteel & Peek-Asa, 2000). In other words, this type of design does 
not test whether the difference in crime between the pre-intervention and post-
intervention periods would have occurred in the absence of the intervention (i.e., 
CPTED). Despite this shortcoming, many CPTED intervention tactics are able to provide 
preliminary evidence of CPTED’s ability to reduce crime (Casteel & Peek-Asa, 2000; 
Cozens, Saville & Hillier 2005; Hendricks et al., 1999; Welsh & Farrington, 2004). 
Unsurprisingly, CPTED has been found to be especially effective in reducing crimes in 
which offenders rely on the physical environment to provide concealment and offer a 
quick escape such as burglary, robbery, assault and vandalism (Casteel & Peek-Asa, 
2000; Cozens, Hillier & Prescott, 2001; Kushmuk & Whittermore, 1981; Ward, 1973).  
In addition, empirical assessments of the impact of CPTED on fear of crime – a 
main outcome that scholars invoke as “malleable” under a CPTED framework - are fairly 
mixed. Among this research, CPTED evaluations that focus on assessing the impact of 
one CPTED standard on fear of crime are generally more successful in finding a negative 
relationship than evaluations of CPTED interventions where multiple standards are 
enforced. In particular, negative relationships are most consistently observed in studies 
that analyze the effects of territoriality, surveillance or maintenance on fear of crime 
(Braga, 2008; Brower, 1980; Brower, Dockett & Taylor, 1983; Brown & Werner, 1985; 
Farrington & Welsh, 2002; Hinkle, 2005; Nasar & Fisher, 1992; Sampson & 
Raudenbush, 2004; Sloan, Fisher & Wilkins, 1995).   
16 
 
Some scholars question the validity of findings from studies that evaluate 
CPTED’s various standards individually (Schneider & Kitchen, 2002). For example, 
according to Schneider and Kitchen (2002), in many cases, “the physical, management 
and community organizational interventions… are woven together in complex ways that 
defy individual analysis” (p. 158). To further emphasis this point, Schneider and Kitchen 
(2002) comment that “it would be as difficult as untangling a spider’s web to evaluate the 
effectiveness of specific place-based crime prevention measures” (p.158).  In part, it is 
exceedingly difficult to evaluate the effectiveness of individual CPTED standards to 
reduce fear of crime because of their interconnectedness. In addition to this difficulty, the 
fact that many CPTED interventions do not enforce one design standard but multiple 
makes studying the effectiveness of individual design standards less relatable to real-
world applications of CPTED compared to evaluating the effectiveness of CPTED 
standards as a package. 
Furthermore, research on environmental cues, such as those introduced by 
Appleton’s (1975) prospect-refuge theory, offer indirect support of CPTED’s ability to 
reduce fear of crime (Appleton, 1975; Nasar & Fisher, 1992: Nasar & Jones, 1997).  For 
example, Appleton (1975) and others have linked the number of available hiding places 
for offenders, known as refuge, to elevated levels of fear of crime (Nasar & Fisher, 1992; 
Nasar & Jones, 1997). In this way, CPTED’s ability to reduce fear of crime is implied 
through its ability to reduce the number of available hiding places for offenders by 
implementing, most notably, its surveillance and maintenance design standards.    
Lastly, studies that investigate the extent to which CPTED standards are enforced 
by residents and compare these findings to residents’ levels of fear of crime have not 
17 
 
consistently found support for CPTED’s ability to reduce fear of crime (Abdullah, 
Marzbali & Tilaki, 2013; Hedayati et al., 2012a; Minnery & Lim, 2005). Rather, these 
studies provide evidence that CPTED may influence fear of crime through other 
channels. For example, finding no direct relationship between CPTED and fear of crime, 
Hedayati et al. (2012a) used structural equation analysis and discovered a negative 
indirect relationship through individuals’ prior victimization experiences.  
Although this mixed body of research may seem discouraging, the impact of 
CPTED on fear of crime is greatly affected by the ways in which both these concepts 
have been measured – an issue which has been greatly debated over the decades of 
research. This debate has resulted in the inconsistent measurement of CPTED and fear of 
crime, making across study comparisons difficult. For example, in their analyses on the 
impact of CPTED on fear of crime, both Minnery and Lim (2005) and Hedayati et al. 
(2012a, 2012b) identify and define CPTED standards and fear of crime differently.  
However, progress is being made toward developing a standard CPTED scale of 
measurement (Abdullah, Marzbali & Tilaki, 2013; Hedayati et al., 2012a, 2012b; 
Minnery & Lim, 2005). In part, this progress is possible because of researchers’ unifying 
understanding of CPTED’s goals. In other words, a common conceptual understanding of 
CPTED has helped facilitate a movement toward developing a standard CPTED scale of 
measurement. Unlike CPTED research, fear of crime research is not motivated by a 
unifying conceptual understanding of the construct of fear of crime; no comprehensive 
theoretical model of fear of crime exists (Alper & Chappell, 2002). 
 
 
18 
 
Understanding Fear of Crime 
As suggested, there is substantial heterogeneity in how researchers interpret and 
measure fear of crime. This heterogeneity is emphasized in Ferraro and LaGrange’s 
(1987) evaluation of fear of crime research. As part of this evaluation, Ferraro and 
LaGrange (1987) investigated the ways in which fear of crime was measured by creating 
a two-dimensional classification system.  The first dimension includes the level at which 
crime perceptions were evaluated: general (community-oriented) or personal (self-
oriented). The second dimension includes the type of crime perception: cognitive (e.g., 
focuses on the likelihood of an event occurring) or affective (e.g., focuses on the emotion 
associated with an event).  
Ferraro and LaGrange (1987) understood fear of crime as an affective (or 
psychological) state.  However, recent scholarship has started to recognize the 
multidimensional character of fear of crime. This recognition has led to advances in, 
among other things, how fear of crime is defined (Doran & Burgess, 2012; Ferraro & 
LaGrange, 1987; Jackson, 2004, 2006; Skogan, 1986; Vanderveen, 2006). In particular, 
fear of crime definitions are now beginning to include both cognitive and affective 
appraisals of crime (Vanderveen, 2006). However, the primary interest of this present 
study is the ways in which both of these appraisals are interpreted in the context of the 
physical environment. In relevant literature, cognitive appraisals of crime often include 
an individual’s perception of safety and his or her perceived risk of victimization 
(Hedayati et al., 2012a, 2012b; LaGrange, Ferraro & Supancic, 1992; Nasar & Jones, 
1997; Rountree & Land, 1996; Toet & van Schaik, 2012). In addition, this literature also 
19 
 
includes the more traditional affective appraisal of an individual’s fear of victimization 
(Park, 2008; Warr, 1990).  
Thus, the concept of fear of crime includes several elements, most notably 
perceptions of safety, risk of victimization and fear of victimization. These elements are 
related but are thought to be conceptually distinct (Dubow, McCabe & Kaplan, 1979; 
Ferraro & LaGrange, 1987; Rountree & Land, 1996; Vanderveen, 2006). Researchers 
may be better able to capture the concept of fear of crime if they use all these elements to 
measure it rather than only one.  
Additionally, a more valid and useful measurement of these fear of crime 
elements is how they relate to specific crime types, rather than to crime in general 
(Ferraro & LaGrange, 1987; Gabriel & Greve, 2003; Jackson, 2004,2006; Rountree & 
Land, 1996).  Relevant research on CPTED commonly examines its impact on the 
following crime types: robbery, physical assault, harassment and sexual assault. These 
crime types are common in CPTED research because they can occur outside and are 
therefore amenable to CPTED standards (Branas et al., 2011; Casteel & Peek-Asa, 2000; 
Kushmuk & Whittermore, 1981; Troy, Grove & O’Neil-Dunne, 2012).  Furthermore, as 
discussed in an earlier section, offenders are able to take advantage of their surrounding 
physical environments to perpetrate these crimes. The same features of the environment 
that aid offenders to commit these crimes also influence non-offenders’ levels of fear of 
crime.  
The literature on fear of crime suggests that it is related to many things that 
generally cluster into individual and situational-level factors. For instance, an individual’s 
sex and age have been consistently found to be related to one’s fear of crime response, 
20 
 
with higher levels of fear of crime found among women and the elderly despite being the 
least victimized groups (Baumer, 1985; Bennett & Flavin, 1994; Flynn, Slovic, & Mertz, 
1994; Hindelang, Gottfredson & Garofalo, 1978; Perkins & Taylor, 1996). In addition, an 
individual’s hometown has also been linked to one’s level of fear of crime. Research 
shows that those who primarily live in urban environments are more likely to have higher 
levels of fear of crime than those who primarily live in rural or suburban environments 
(Lab, 2014; Toseland, 2002; Will, 1995; Will & McGrath, 1995). This may be because 
crime is more prevalent in urban areas than it is in other places.  Lastly, many studies 
have reported that those who have the highest levels of fear of crime are minority 
individuals and those of low socioeconomic status (Bennett & Flavin, 1994; Parker & 
Ray, 1990; Skogan & Maxfield, 1981; Smith & Lab, 1991; Will & McGrath, 1995). For 
example, Will and McGrath’s (1995) analysis on neighborhood fear and the underclass 
found that non-whites had higher levels of neighborhood fear than whites.2 However, 
when differentiating participants by class status, the effect of a respondent’s race was no 
longer significant. Regardless of place, Will and McGrath (1995) found that the 
underclass were more fearful than the non-poor and that this effect was the strongest for 
women. 
The relationship between an individual’s victimization history and his or her level 
of fear of crime is less understood. Some research indicates that victims of crime have 
higher levels of fear of crime than non-victims (Garofalo, 1979; Gomme, 1986; Keane, 
1995; Lumb, Hunter & McLain, 1993; Skogan, 1987). However, fear of crime may not 
                                                          
2  Will and McGrath (1995) measured neighborhood fear as a binary variable. Participants were asked the 
following question:  “Is there any area right around here where you would be afraid to walk alone at 
night?” (p. 168) Affirmative responses were coded as 1 and negative responses as 0.  
21 
 
have such a clear relationship with one’s actual victimization experiences. Many other 
studies have failed to find any relationship between past victimization experiences and 
fear of crime (Ferraro, 1995, Garofalo, 1979; McGarrell, Giacomazzi & Thurman, 1999; 
Perkins & Taylor, 1996). Interestingly, research provides evidence that an individual’s 
indirect victimization experiences may play a leading role in affecting his or her level of 
fear of crime (Arnold, 1991; Hale, 1996; Skogan & Maxfield, 1981). For example, 
Arnold (1991) found that an individual’s indirect, rather than direct, victimization 
experiences played a more significant role in influencing his or her level of fear of crime.  
Whereas some individual characteristics equally pertain to all elements of fear of 
crime, such as those previously discussed, others do not. An individual’s perceived locus 
of control is a case of this phenomenon; it is more closely related to one’s perceived risk 
of victimization than to the other elements of fear of crime.  Rotter (1966) suggested that 
one’s locus of control, internal versus external, is related to the extent to which one 
believes he or she can control a dangerous situation. Those with internal loci of control 
may believe they can control a dangerous situation, whereas their counterparts likely 
view the same situation as beyond their control. Subsequent research has supported 
Rotter’s (1966) claim, finding perceptions of risk lower in individuals who believe 
themselves to be able to take control over a risk or hazard (Prociuk, Breen & Lussier, 
1976; Riechard & Peterson, 1998; Slovic, 2010).   
Individual characteristics play an important role in influencing fear of crime. 
However, of particular interest to this study are the situational-level factors linked to 
CPTED’s alterations to the physical environment that affect an individual’s level of fear 
of crime. Fear of crime has been demonstrated to be unevenly distributed over space and 
22 
 
time (Herzog & Bryce, 2007; Nasar & Jones, 1997; Spelman, 1993; Stamps, 2005; 
Wilson & Kelling, 1982). For example, some areas, such as vacant lots (e.g., Spelman, 
1993), and times of day, such as nighttime hours (e.g., Nasar & Jones, 1997), are 
associated with elevated fear of crime responses. These responses are thought to be 
influenced by a repository of mental images that individuals store of their surroundings, 
which are developed by ongoing experiences with the environment. Furthermore, these 
images are associated with perceptions about, among other things, safe and unsafe areas 
(Hanyu, 1993; Kaplan, 1979; Nasar & Jones, 1997).  In other words, these mental images 
allow individuals to quickly identify components of their surroundings that act as 
environmental cues signaling safe and unsafe areas. This process, in turn, influences 
individuals’ levels of fear of crime.  In theory, these mental images (linked to individuals’ 
levels of fear of crime) help individuals quickly formulate predictions of what will 
happen to them in a particular environment, evaluate its consequences and formulate a 
response (Kaplan, 1979; Liska, Sanchirico & Reed, 1988; Nasar & Jones, 1997; Warr, 
1990).  
As suggested, this present study is particularly interested in environmental cues 
that influence fear of crime assessments that not only are affected by CPTED’s 
alterations to the built environment but are also within this study’s control to manipulate. 
In this regard, research on perceptions of physical disorder, territorial features and the 
related concepts of prospect, concealment and entrapment are able to provide insight into 
the role environmental cues play in influencing fear of crime. Armed with this 
knowledge, and facilitated by the conceptual understanding of CPTED, the ability of 
CPTED standards to decrease fear of crime can be better understood.  
23 
 
Perceptions of Disorder 
Numerous studies have assessed the relationship between perceptions of disorder 
and fear of crime (Covington & Taylor, 1991; Hinkle & Weisburd, 2008; Markowitz, 
Bellair, Liska & Liu, 2001). Many of these studies support the argument that residents 
who perceive more disorder related cues (e.g., graffiti, trash and overgrown foliage) in 
their immediate surroundings perform worse on fear of crime assessments than their 
counterparts. Here, disorder acts as environmental cues which signal to residents that 
deviant behaviors are unable to be controlled.  (Wilson & Kelling, 1982). In this way, 
disorder serves as a signal to identify dangerous areas which subsequently affects fear of 
crime assessments. However, some research has found that individuals who are more 
readily exposed to disorderly environments can become desensitized to the presence of 
disorder (Harcourt, 2001; Taylor, Shumaker & Gottfredson, 1985). In such cases, 
perceptions of disorder are not shown to illicit the reactions previously described. Yet, in 
the aggregate, research is largely supportive of a positive relationship between 
perceptions of disorder and fear of crime. 
Studies that assess perceptions of disorder generally identify stronger positive 
relationships with fear of crime than studies that collect data on the level of disorder 
through systematic observations of social space (Perkins & Taylor, 1996; Sampson & 
Raudenbush, 2004; Taylor, 1996).  Although this dichotomy is curious, it does not 
undermine the utility of perception-based research. In fact, much research has found 
perceptions of disorder and objective measures of disorder to be highly correlated 
(Perkins, Meeks & Taylor, 1992; Sampson & Raudenbush, 2004; Taylor, Shumaker & 
Gottfredson, 1985). Furthermore, perceptions of disorder reflect a psychological 
24 
 
construct that is more relatable to the cognitive and affective elements of fear of crime.  
In regards to CPTED, recall that Jeffery (1977) emphasized the ability of environmental 
cues, such as physical disorder, to trigger individuals to formulate perceptions of their 
surroundings that are used in cognitive processes that affect behavior. Although not 
explicitly stated by Jeffery (1977), perceptual, as well as affective, elements of fear of 
crime are influenced by these cognitive processes. Therefore, it would seem that attention 
to perceptions of disorder rather than to objective measures of disorder more wholly 
reflects the mechanisms by which CPTED influences fear of crime.  
However, not all elements of disorder can be easily manipulated for purposes of 
experimentation.  Disorder consists of both physical and social elements. Physical 
disorder, previously introduced, includes environmental features that reflect 
neighborhood dilapidation such as graffiti and debris (Skogan, 1990). Moreover, social 
disorder consists of behavioral issues within communities such as panhandling and 
loitering (Skogan, 1990). Unlike social disorder, physical disorder can be more easily 
experimentally manipulated for research purposes. In addition, it is arguably less likely to 
be misinterpreted by respondents when viewed in a picture (the medium of this study). 
Similar to research that makes no distinction between the types of disorders (physical or 
social), research deciphering the effects of physical disorder from social disorder have 
fairly consistently found a negative relationship with fear of crime (Covington & Taylor, 
1991; Perkins & Taylor, 1996; Wikström, 1990).  
CPTED has the ability to affect perceptions of disorder through its maintenance 
standard. Recall that CPTED’s maintenance standard requires environments to be free of 
physical disorder. In this way, CPTED affects perceptions of physical disorder by 
25 
 
reducing the number of disorder-related cues in the environment.  In doing so, CPTED is 
able to reduce fear of crime by altering, for example, expectations linked to individuals’ 
perceptions of disorder. These expectations include how others will act in a given 
environment which aid individuals in defining safe and unsafe places. Traditional urban 
greening practices also focus on improving disorderly conditions by removing disorder-
related cues and, in this way, are thought to similarly affect perceptions of disorder and 
fear of crime.  
Territorial Features 
Territorial features, such as personalization items and clear demarcation between 
public and private spaces, have also been shown to influence fear of crime.  In particular, 
this body of research has shown that residents and non-residents alike use territorial 
markers as indicators on how others will receive them (Brower, 1980; Greenbaum & 
Greenbaum, 1981; Taylor & Harrell, 1996). Research has found individuals to have 
lower levels of fear of crime in areas that contain territorial features, suggesting that 
perceptions of ownership and proprietary concern may play a leading role in the 
interpretations individuals make about their surroundings (Abdullah, Hedayati, Haron, 
Bahauddin, & Maghsoodi, 2013; Brower, Dockett & Taylor, 1983; Brown & Werner, 
1985; Newman, 1972; Pollack & Patterson, 1980; Taylor, Shumaker& Gottfredson, 
1985). For example, according to CPTED, territorial features trigger the perception that a 
location is under the care of an individual or group of individuals and, consequently, that 
deviant acts will not go unpunished. Interpreted in this way, territorial features function 
as environmental cues that signify safe areas. Such cues, as previously explained, are 
associated with lower levels of fear of crime. Research on territorial features also 
26 
 
suggests that more threatening environments, such as those which are disorderly, require 
more territorial features to decrease fear of crime (Brower, Dockett & Taylor, 1983).  
CPTED’s territoriality standard promotes the use of territorial features and, 
consequently, takes advantage of the opportunity to optimize reductions in fear of crime. 
Contrastingly, traditional urban greening practices do not focus on improving the 
environment in this way. Rather, the enforcement of territoriality may be an unintended 
consequence. For example, the arrangement in which trees or bushes are planted may, 
inadvertently, aid in providing a distinction between public and private property, 
consequently facilitating the concepts of ownership and proprietary concern associated 
with territoriality. Although this arrangement may be intentional its consequence, 
facilitating CPTED’s territoriality concept, is not. The effect of urban greening’s 
unintended enforcement of territoriality is arguably not equivalent to its explicit 
enforcement by CPTED. CPTED includes more territorial features and in doing so 
capitalizes on the opportunity to reduce fear of crime.  
Prospect, Concealment and Entrapment  
Related to Appleton’s (1975) prospect-refuge theory, another set of physical cues 
that individuals use to infer what will happen to them in a particular environment are 
those associated with the interrelated concepts of prospect, concealment and entrapment.  
Prospect refers to an individual’s ability to see into a space. In places like vacant lots, this 
ability is impaired by physical disorder (e.g., overgrown foliage and debris). Physical 
disorder also provides concealment for offenders to hide in wait of a passerby who will 
not be able to anticipate their advances. In addition, physical disorder impairs an 
individual’s ability to quickly and easily escape an offender’s attack by obstructing his or 
27 
 
her access to and egress from a space. The inability to quickly and easily escape a 
potential offender is known as entrapment.   
Collectively, individuals interpret features of the environment that impair prospect 
(hereon referred to as blocked prospect), provide hiding places for offenders 
(concealment) and impede escape (entrapment) as cues signifying dangerous places. 
Unsurprisingly, individuals exposed to environmental conditions that facilitate blocked 
prospect, concealment and entrapment have, on average, heightened levels of fear of 
crime (Day, 1994; Fisher, Sloan & Wilkins, 1995; Nasar & Fisher, 1992; Nasar & Jones, 
1997).  
Vacant lots’ disorderly conditions readily provide opportunities for blocked 
prospect, concealment and entrapment. Traditional urban greening practices and CPTED 
both seek to rectify disorderly conditions by focusing on lot maintenance. In particular, 
traditional urban greening practices’ focus on maintenance alone will arguably not 
achieve the greatest returns in reducing opportunities for concealment and entrapment. 
Maintenance that is reinforced by CPTED’s surveillance and access control standards 
help optimize reductions in concealment and entrapment by promoting visibility (i.e., 
prospect) and easy access to and egress from a space, respectively. In doing so, the 
combined effect of enforcing maintenance, surveillance and access control is likely better 
able to reduce fear of crime than that which would likely be achieved by enforcing 
maintenance alone. This specified attention to surveillance and access control, in addition 
to maintenance, is what distinguishes CPTED from traditional urban greening practices.   
 
 
28 
 
CPTED & Fear of Crime: Greening Vacant Lots 
 Vacant lots are common features of cities and their surrounding neighborhoods 
that facilitate fear of crime (Spelman, 1993; Wilson & Kelling, 1982). Methods to 
alleviate fear of crime deserve our attention due to its association with debilitating 
outcomes that affect an individual’s social, mental and physical health and, on a larger 
scale, the vitality of his or her community (Green, Gilbertson & Grimsley, 2002; Morrall 
et al., 2010; Ross & Mirowsky, 2001; Wilson & Kelling, 1982). These debilitating 
outcomes negatively affect residents’ quality of life (Blakely & Snyder, 1997; Green, 
2002).  
One way in which community officials attempt to reduce fear of crime, and in 
doing so increase quality of life, is to transform vacant lots into green space (Baltimore 
Green Space, 2013; Greening of Detroit, 2014; Pennsylvania Horticultural Society, 2014; 
Urban Green, 2014). Although these efforts are admirable, traditional urban greening 
practices likely do not take advantage of opportunities to reduce fear of crime. Rather, it 
has been argued that the implementation of CPTED standards into greening practices will 
capitalize on this opportunity. CPTED’s territoriality, surveillance, access control and 
maintenance standards are specifically designed to reduce fear of crime by, among other 
things, reducing disorder-related cues, introducing territorial markers, improving 
visibility, eliminating hiding places for offenders and providing easy access to and egress 
from a space.  
In part, urban greening provides an ideal framework for the implementation of 
CPTED because it is already organized around one of CPTED’s staple standards: 
maintenance. Given the interconnectedness of CPTED standards, introducing 
29 
 
territoriality, access control and surveillance into traditional urban greening practices 
could easily be achieved. This effort is an important step towards discovering urban 
greening techniques that are able to achieve greater reductions of fear of crime than 
traditional approaches. These new techniques also have a greater potential to help 
alleviate residents from the debilitating social, mental and physical conditions that 
suppress their quality of life and the welfare of their communities. 
The present study will test the following hypotheses:   
Hypothesis 1: On average, participants who view either a greened lot or a 
CPTED lot will report lower levels of fear of crime than participants who view a 
disorderly lot.  
 
Hypothesis 2: On average, participants who view a CPTED lot will report lower 
levels of fear of crime than participants who view a greened lot relative to a 
disorderly lot. 
Due to the interconnectedness of CPTED standards, this study evaluates CPTED as a 
package. This study will test these hypotheses using a computer program to adjust an 
image of a vacant lot to meet three conditions: A disorderly lot; 2) A greened lot; and 3) 
A CPTED lot. These three images represent the three levels of the study’s independent 
variable and will be randomly distributed in the form of questionnaires to undergraduate 
students. Additionally, to facilitate fear of crime as a collective concept that includes both 
cognitive and affective elements, this questionnaire will assess perceptions of safety, risk 
of victimization and fear of victimization in relation to four previously identified crime 
types: robbery, physical assault, harassment and sexual assault. This study will then use 
Ordinary Least Squares Regression (OLS) to test its hypotheses.  
 
 
30 
 
Chapter 3: Methods  
Sample 
The present study collected its data from a convenience sample of undergraduate 
students enrolled at a large East Coast university. This sample consisted of students who 
were enrolled in one of eight criminal justice classes whose professors agreed to set aside 
class time for the purpose of this study. In particular, these classes included a range of 
introductory to advanced courses. There were two exclusion criteria for participation. 
First, students that had previously participated in this study were not allowed to 
participate a second time. This precaution was necessary given that this study required 
that students be surveyed from multiple classes. Lastly, this study also excluded students 
who were younger than 18 years of age and required the permission of a parent or 
guardian to participate.  A total of 551 students were surveyed. Due to missing data, 523 
participants were retained for analysis.3  
The purpose of sampling students from multiple classes was to help ensure 
sufficient sample sizes for each of the study’s three conditions which represent the three 
levels of the independent variable: a disorderly lot, a greened lot and a CPTED lot. Extant 
literature proposes many different methods for determining sample size including power 
analytic techniques and accuracy in parameter estimation (AIPE) (Cohen, 1988; Kelley & 
Maxwell, 2003; Maxwell, 2000).  Many of these methods require making assumptions 
about the sample such as its expected effect size as well as its corresponding population 
standard deviation and mean. By studying similar research, scholars are often able to 
                                                          
3 This study did not pursue imputing missing data (or other estimation techniques) due to the low number 
of missing cases.  
31 
 
make well-informed assumptions. However, the utility of these methods to determine 
sample size becomes suspect when researchers, like in this present case, are unable to 
turn to past research to make educated assumptions about the likely effect size, for 
example. For this reason this study did not attempt to calculate an appropriate sample 
size. Rather, it sought to optimize its sample size by drawing data from students enrolled 
in multiple different classes.  
This study’s sample is not without its limitations. A chief limitation of student 
samples is the fact that students represent a select group that is not representative of the 
general population. Therefore, student samples limit a study’s external validity. For 
example, participants in college-drawn student samples are more likely to come from 
affluent backgrounds (Sears, 1986; Wells, 1993). Consequently, these samples are less 
likely to capture data from minority and lower-class individuals who tend to be 
underrepresented in many U.S. colleges. Despite this limitation, a student sample is 
reasonable for this study’s hypotheses because it allows the opportunity to begin building 
a foundational understanding of an unexplored research area without the costs to time and 
financial resources needed to procure a wider demographic of participants. In other 
words, the primary objective of this study is not to strive to achieve external validity. 
Rather, its primary objective is to begin a discussion on the potential utility of improving 
traditional urban greening practices in order to optimize reductions in fear of crime and, 
in doing so, determine whether a large-scale study utilizing a sampling technique that 
captures participants that are representative of a future undetermined target population is 
worthwhile.  
 
32 
 
Procedure  
The primary investigator asked students during their lecture hour to participate in 
a study on the impact of certain environmental features on fear of crime by completing a 
nine-page questionnaire approved by the university’s Institutional Review Board (IRB). 
At this time, the primary investigator read a script to the potential participants that 
contained the purpose of the study and a brief overview of the questionnaire (see 
Appendix A).   
Each questionnaire had a number located in the top right-hand corner of every 
page. This number identified each questionnaire as containing either an image of a 
disorderly lot (identified as 1), a greened lot (identified as 2) or a CPTED lot (identified 
as 3).  This was the only difference across questionnaires. This study used these three 
numbers (1, 2 and 3) to simulate random assignment. To do so, this study used a 
computer program (Research Randomizer) to generate a random sequence of these three 
numbers. The primary investigator arranged the questionnaires in accordance to this 
random number sequence. This procedure helped ensure that all three numbers appeared 
approximately the same number of times and that roughly equal numbers of participants 
were in each condition.  
Moreover, prior to distributing the questionnaires and consent forms, the primary 
investigator instructed participants to not sign either document in order to protect their 
anonymity and to read the consent form before completing the questionnaire. After 
providing this instruction, the primary investigator (and her research volunteers) then 
distributed the questionnaires along with consent forms. Completing the questionnaire 
indicated that participants did not already participate in the study, were at least 18 years 
33 
 
of age, had read the consent form and agreed to participate in this research study. A total 
of 551 out of 826 individuals agreed to participate, making this study’s refusal rate 
approximately 33.29%.4 
 Environmental Simulation: A Justification for using Images  
As mentioned, this study utilized a computer program to adjust a photograph 
taken of a vacant lot (see Figure 1) to meet the following conditions: 1) A disorderly lot; 
2) A greened lot; and 3) A CPTED lot. A disorderly lot is one that has physical 
manifestations of neighbor dilapidation (e.g., trash and overgrown foliage), while a 
greened lot is one that has been transformed according to traditional urban greening 
standards. Lastly, a CPTED lot is one that incorporates CPTED standards into the 
traditional greening process. 
Using pictures to represent these three conditions is not only cost-effective, but 
also reduces potential confounders. This study was able to minimize (if not eliminate) 
confounding variables by creating a template from which adjustments were made to 
simulate each condition (i.e., a disorderly lot, a greened lot and a CPTED lot). Because 
all images used the same basic template, one can be more confident that any observed 
differences across conditions in respondents’ levels of fear of crime are reflective of 
actual treatment effects rather than the result of unaccounted for differences across 
images. This confidence would not be so easily obtained in a “real-world” application of 
                                                          
4 This study’s refusal rate is not atypical for student samples (Jans & Roman, 2007).   
That being said, this study suspects that its high refusal rate was partially due to the time during classes that 
questionnaires were distributed. Questionnaires were distributed either at the end of class or at the 
beginning of class. The time at which questionnaires were distributed was dictated by professors’ 
preferences. Students were more likely to refuse to participate in this study when questionnaires were 
distributed at the end of class than at the beginning of class. In particular, the refusal rate was 43.8053% for 
the six classes where questionnaires were distributed at the end of class and 20.8556% for the remaining 
two classes where questionnaires were distributed at the beginning of class. 
34 
 
this study. For example, in a real-world scenario each lot would only be able to represent 
one condition: a disorderly lot, a greened lot or a CPTED lot. It is likely that each lot 
would have unique features, such as its distance to schools, parks and businesses, which 
would affect individuals’ levels of fear of crime (Branas et al., 2011; Heckert & Mennis, 
2012). These features are likely to be ignored (or improperly controlled), obscuring the 
identification of a (valid) treatment effect.   
The high level of control offered by images is something that real-world 
environments simply cannot offer. For this reason, many studies have utilized images to 
assess the impact of environmental conditions on fear of crime (Andrews & Gatersleben, 
2010; Austin & Sanders, 2007; Cinar & Cubukcu, 2012; Herzog & Flynn-Smith, 2001; 
Jorgensen, Ellis & Ruddell, 2013; Stamps & Smith, 2002; Wang & Taylor, 2007). For 
example, Austin and Sanders (2007) exposed participants to photographs taken of graffiti 
to assess whether viewing these images affected perceptions of safety.  However, 
researchers need not use images to gain this high degree of control. This ability is a 
defining feature of what is referred to as “artificial digital environments.” Artificial 
digital environments are environments that are created by using technology (e.g., cameras 
or computer software) to simulate conditions similar to that of the real-world. Artificial 
digital environments include static media (e.g., photographs and 360° panoramic images) 
as well as dynamic media (e.g., virtual environments).  
Of course, a primary benefit of real-world environments is their ability to offer 
“ecological validity.” Ecological validity refers to the extent to which the methods, 
materials and setting of the study approximate real-world conditions (Schmuckler, 2001; 
35 
 
Shadish, Cook & Campbell, 2002). Improving ecological validity can improve external 
validity which is concerned with the generalizability of a study’s results.  
In their analysis on the effect of disorder on fear of crime in real and virtual 
environments, Toet and van Schaik (2012) explain that for an artificial digital 
environment (e.g., photographs and virtual environments) to be considered ecologically 
valid it must evoke “cognitive, emotional and affective user responses that are similar to 
those that would be evoked by a real equivalent” (p. 260). Toet and van Schaik (2012) 
then explain that studies that utilize artificial digital environments do not need to fulfill 
all aspects of ecological validity (hereon referred to as “full ecological validity”) in order 
to assess the influence of physical interventions in the built environment, such as 
environmental cues and urban design, on human behavior (i.e, cognition, emotion and 
affect). Said differently, artificial digital environments do not have to evoke similar 
responses for all human behaviors to those thatwould be evoked by a real-world 
equivalent. Rather, it is permissible, in most circumstances, for artificial digital 
environments to affect only some human behaviors, such as fear of crime, in a similar 
way as their real-world equivalents and in doing so achieve “partial ecological validity” 
(van Hagen, 2011).   
Meta-analyses comparing static media to evaluations gathered from their 
equivalent real-world environments offer considerable support of the ability of responses 
obtained by viewing images to achieve partial ecological validity (Palmer & Hoffman, 
2001; Stamps, 1990; Stamps, 2010). In particular, using the key words “visual simulation 
validity,” Stamp (2010) identified 33 articles in Science Citation Index and Google Schol-
ar which together included more than 3,511 participants and 451 (both real and digital) 
36 
 
environments.  From these articles, Stamp identified a strong correlation (r=.86) between 
responses obtained from viewing (color) images and responses from similar real-world 
environments.  
Despite this support, there are some noteworthy limitations of using photographs. 
For example, unlike real or virtual environments, images are highly dependent on 
viewpoint (Palmer & Hoffman, 2001). Images only allow individuals to view one angle 
of a location.  A common solution to this issue of limited viewpoint is computer 
simulated 360° panoramic images.  Research that links aspects of the physical 
environment to passengers’ levels of fear of crime in railway stations commonly use 
these images (Cozens, Neale & Hillier, 2003; Cozens, Neal, Hillier & Whitaker, 2004; 
Cozens, Neal, Whitaker & Hillier, 2003). 
For this study, a 360° panoramic image would require systematically stitching 
together separate images using a computer program. This process alone is very tedious; 
images must be precisely stitched together to create the impression of seamlessness. This 
impression is critical to building an artificial digital environment that is realistic or, in 
other words, similar to its real-world equivalent.  This task is made even more difficult 
when, in this process of stitching together images, one must account for the large amount 
of editing that must be performed for this 360° panoramic image to meet the three lot 
conditions of this study. This editing process would likely jeopardize attaining seamless 
360° panoramic images. In addition, 360° panoramic images have not yet been compared 
against responses obtained from equivalent real-world environments.  Therefore, no 
empirically-based argument can be made for using 360° panoramic images over regular 
37 
 
photographs. For these reasons this study did not pursue constructing 360° panoramic 
images.   
In addition, images, including 360° panoramic images, are often criticized as not 
fully capturing the dynamics of a real-world environment. In particular, common 
criticisms of images include lack of ambient sounds and the inability to walk through and 
interact with an environment. These criticisms are often addressed through developing 
virtual environments. Recent research has used virtual environments to study the impact 
of environmental cues on fear of crime (Park, 2008; Park et al., 2011, 2012). Studies that 
compare the responses received from a simulated environment to their real-world 
counterpart have garnered some support for the ability of virtual environments to affect 
human behavior, such as fear of crime, in a similar way (Bishop & Rohrmann, 2003; 
Houtkamp & Junger, 2010; Schwebel, Gaines & Severson, 2008; Toet & van Schaik, 
2012). Unfortunately, creating a virtual environment is beyond the financial capability 
and time-constraint of this study.5   
Measures 
Dependent Variables  
This study uses three fear of crime elements to assess the fear of crime concept: 
perceptions of safety, risk of victimization and fear of victimization. As previously 
discussed, a more valid and useful measurement of these fear of crime elements is how 
they relate to specific crime types, rather than to crime in general (Ferraro & LaGrange, 
1987; Gabriel & Greve, 2003; Jackson, 2004, 2006; Rountree & Land, 1996).  The 
present study chose to focus on the crimes of robbery, physical assault, harassment and 
                                                          
5 The cost of creating a virtual environment was estimated to be around $10,000 and would have taken 
approximately one year to design.   
38 
 
sexual assault based upon their ability to be conducted outside and their prevalence in 
relevant research on CPTED (Branas et al., 2011; Casteel & Peek-Asa, 2000; Kushmuk 
& Whittermore, 1981; Troy, Morgan & O’Neil-Dunne, 2012). Furthermore, in earlier 
sections it was discussed how offenders and non-offenders are influenced by features 
present in the environment. This discussion supports this study’s focus on these crime 
types.  Recall that the same features of the environment that influence and aid offenders 
to commit robbery, physical assault, harassment and sexual assault also affect non-
offenders’ levels of fear of crime. For example, physical disorder provides offenders 
hiding places where physical and sexual assaults are able to be conducted undetected.  
Non-offenders interpret locations where physical disorder is prevalent as dangerous 
places where these and other crimes likely occur. The questionnaire provides participants 
with definitions of each crime type which were adapted from the Uniform Crime Report 
(UCR) as well as from additional non-governmental sources. 
In order to develop measurement scales for these fear of crime elements, this 
study identified common themes in how these elements were previously measured. In 
part, developing scales in this way allowed this study to remain as consistent as possible 
with past operationalizations and measurements of fear of crime. This study then adapted 
each scale to fit the format of its questionnaire. Table 2 displays the descriptive statistics 
for these three measures for the full sample as well as for each level of the independent 
variable.  
------------------------------- 
             Insert Table 2 about here 
------------------------------ 
 
39 
 
Perception of Safety. Perceptions of safety are often measured by asking 
individuals to indicate on a Likert scale how safe they would feel walking alone at night 
often within the context of a particular location, such as a block or a neighborhood (e.g., 
National Crime Victimization Survey (NCVS), Crime Survey (CS) (formerly referred to 
as the British Crime Survey) and General Social Survey (GSS)). Such measures have 
been criticized for being limited to nighttime, not directly mentioning crime, referring to 
vague geographical areas (i.e. the neighborhood) and only crudely measuring intensity 
(Ferraro, 1995). These measures have also been criticized for asking about an activity 
(walking alone at night) that people rarely or never do in their daily lives (Ferraro, 1995). 
In an attempt to address some of these criticisms, this study asked participants to indicate 
on a seven-point Likert scale the extent to which they would feel safe from becoming a 
victim of  each crime type (robbery, physical assault, harassment or sexual assault) while 
walking alone by or in a specific location (the pictured lot).  Responses ranged from one 
to seven and included the following: “very safe”(1), “safe”(2), “somewhat safe”(3), 
“neutral”(4) “somewhat unsafe”(5), “unsafe”(6) and “very unsafe”(7). 
This new perception of safety measure addresses five general criticisms. First, it is 
not limited to nighttime. It also explicitly mentions crime by asking respondents to report 
their perceived levels of safety in regards to robbery, physical assault, harassment and 
sexual assault.  In addition, this measure refers to a concrete geographic area: the pictured 
location an individual randomly received. Furthermore, this study utilized a seven-point 
Likert scale in order to create an instrument that is better able to measure intensity. 
Lastly, this measure is purposefully ambiguous about the time of day an individual walks 
40 
 
alone in order to better facilitate responses based on activities that are relevant to 
participants’ daily lives.  
Risk of Victimization. Risk of victimization is commonly measured by asking 
individuals to indicate on a Likert scale their perceived likelihood of falling victim to a 
specific crime (Hale, 1996; Jackson, 2006, 2011; Warr, 1985, 1987). To remain 
consistent with past measurements, this study also used a similar measure. This study 
required participants to indicate on a seven-point Likert scale how likely- at the location 
(picture) shown- they thought it would be that they would become a victim of each crime 
type (robbery, physical assault, harassment or sexual assault).  Responses ranged from 
one to seven and included the following: “very unlikely” (1), “unlikely”(2), “somewhat 
unlikely”(3),  “neutral”(4), “somewhat likely”(5), “likely”(6) and “very likely” (7).     
Fear of Victimization. Fear of victimization is commonly measured by asking 
individuals to indicate on a Likert scale how afraid they are that they will become a 
victim of a specific crime (Ferraro, 1996; Warr, 1987, 1990 1994; Warr & Stafford, 
1983).  Again, to remain consistent with past measures, this study also included a similar 
item. This study required participants to indicate on a seven-point Likert scale how 
afraid- at the location (picture) shown- that they would become a victim of each crime 
type (robbery, physical assault, harassment or sexual assault).  Responses ranged from 
one to seven and included the following: “not at all afraid”(1), “not afraid”(2), 
“somewhat not afraid”(3), “neutral”(4), “somewhat afraid”(5), “afraid”(6) and “very 
afraid”(7).   
This study used all 12 of its fear of crime items in a factor analysis. In essence, 
factor analysis is used to identify relationships among a set of items and from these 
41 
 
relationships determine a reduced set of common factors (Grice, 2001, 2007). These 
common factors are unobserved latent constructs that hold theoretical significance (Grice, 
2001, 2007; Porter & Fabrigar, 2007).  This method is commonly used in fear of crime 
research to understand how fear of crime items are related to one another (Ferraro, 1995; 
Vanderveen, 2006).  Using the above fear of crime items, a factor analysis will allow this 
study to identify any existing factors and to create measures that reflect underlying latent 
constructs using factor scores. 
There are two main types of factor analysis: Confirmatory Factor Analysis (CFA) 
and Exploratory Factor Analysis (EFA).  CFA is used to test hypotheses on the 
relationships between observable items (or variables) and existing underlying latent 
constructs. Here, researchers use theory and empirical research to guide them on this 
relationship a priori. These hypotheses are then statistically tested.  Unlike CFA, EFA 
explores the underlying structure of interrelated items without imposing any restrictions 
on the outcome (Child, 1990). In other words, researchers conducting EFA set no 
expectations on the nature of the items and the number of underlying latent constructs. 
EFA is often used when there are many competing theoretically-related reasons for 
determining a measurement model. Recall that there is substantial heterogeneity in how 
researchers interpret and measure fear of crime (Farrall, Bannister, Ditton & Gilchrist, 
1997; Ferraro & LaGrange, 1987). As a consequence, no comprehensive theoretical 
understanding of fear of crime exists (Alper & Chappell, 2002). This heterogeneity lends 
credence to the use of EFA in fear of crime research. For this reason, this study used EFA 
to identify factors related to fear of crime. 
42 
 
Prior to extracting factors, it is prudent to assess whether the data are suitable for 
factor analysis.  Factor analysis requires large samples to ensure that the correlations 
among variables are reliable estimates. Relatedly, it is also preferable that the ratio of 
subjects-to-variables is large. Although there is no hard and fast rule for these criteria, a 
sample size of 150 or larger and subject-to-variable ratio of 10 to 1 is generally accepted 
(Beavers et al., 2013). This study’s sample size (N=523) and subject-to-variable ratio 
(≈43 to 1) meet these criteria. Furthermore, factor analysis requires that there are no 
outliers and that the relationship among pairs of variables is linear. Using scatterplots of 
the fear of crime items, this study validated that these conditions were upheld. Lastly, 
factor analysis does not require multivariate normality, however it is desirable because it 
improves statistical inference. Tests for normality reveal that this study’s fear of crime 
items do not have multivariate normal distributions.  
There are several other more rigorous methods for determining whether data are 
suitable for factor analysis. These methods include examining a correlation matrix and 
conducting the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy test.  The 
Pearson correlation matrix is perhaps the most commonly used matrix to assess the 
interrelationships between sets of variables. This matrix assumes that variables are 
measured on an interval or ratio scale and that there are linear relationships between 
variables. The Pearson correlation matrix is not appropriate for the present analysis given 
the ordinal nature of its fear of crime items. In fact, the Pearson correlation matrix often 
underestimates the relationship between individual ordinal items and can lead to biased 
factor loadings (Bernstein & Teng, 1989). For this reason, this study used a polychoric 
correlation matrix to assess the relationships between individual items. This technique is 
43 
 
more amenable to ordinal data. It estimates the relationship between two bivariate 
normally distributed continuous variables measured on ordinal scales (Olsson, 1979). The 
assumption of bivariate normality may seem limiting, however it is only necessary 
that some monotonic transformation of the responses under examination exist such that 
their transformed versions have a bivariate normal distribution (Uebersax, 2015). Like the 
Pearson correlation matrix, the polychoric correlation matrix provides a statistic ranging 
from -1 to 1. Scholars suggest that factor analysis should be reconsidered if few 
correlation coefficients exceed ±0.30 (Costello & Osborne, 2005). Table 3 displays the 
polychoric correlation matrix. The matrix reveals high to moderate correlation 
coefficients among all paired items. This finding suggests not only that factor analysis is 
appropriate but also that these items may be united by one underlying common factor.  
------------------------------- 
            Insert Table 3 about here 
------------------------------ 
This study then conducted KMO measure of sampling adequacy test. The KMO 
measure of sampling adequacy test provides a statistic ranging from 0 to 1 that reflects 
the proportion of variance among the variables that is shared or common. Higher values 
are more suggestive of the existence of an underlying common factor(s). The present 
study’s KMO test yielded a statistic of 0.886 which suggests that items share a large 
proportion of variance and are suitable for factor analysis.6  
After verifying that the data were suitable for factor analysis, this study then 
selected an appropriate extraction method to determine the number of underlying latent 
                                                          
6 Bartlett’s test of sphericity is another commonly used test to determine whether factor analysis is 
appropriate. However, this study did not conduct this test due to its sensitivity to samples that depart from 
multivariate normality. 
 
44 
 
factors.  EFA extraction methods are iterative processes that rely on matrix algebra to 
create linear combinations of items that explain the maximum amount of variance 
between items (Beavers et al., 2013). The first extraction in this process is based off the 
assumption that each linear combination is independent (Beavers et al., 2013).  These 
linear combinations represent factors. This iterative process continues until all of the 
sample’s variance is accounted for (Suhr, 2006).  
The two most commonly used extraction methods for EFA include Maximum 
Likelihood (ML) and Principal Axis Factoring (PAF). The ML extraction method 
assumes that the data follow a multivariate normal distribution and that variables are 
continuous (Pett, Lackey & Sullivan, 2003). For these reasons, ML was not an 
appropriate extraction method for this study’s data.  Unlike ML, PAF does not make 
strong assumptions about normality (Fabrigar et al., 1999). Consequently, it is less likely 
than ML to produce improper solutions. However, a downside of PAF is that it does not 
allow researchers to analyze confidence intervals or significance tests. It also only 
provides a restricted range of goodness-of-fit indexes. Despite these restrictions, this 
study selected PAF as an extraction method due to its relaxed assumption regarding 
normality. 
Recall that factor analysis interprets inter-correlations between variables and with 
this knowledge identifies common factors. Consequently, the core of factor analysis can 
be said to be its correlation matrix. The Pearson correlation matrix is the default in most 
statistical software packages for factor analysis such as STATA. This study used a 
polychoric matrix to conduct its EFA because the Pearson correlation matrix requires that 
items are measured on an interval or ratio scales. Using a polychoric correlation matrix 
45 
 
and PAF as an extraction method, this study applied EFA to its 12 fear of crime items. 
There are several criteria for identifying the appropriate number of latent factors. For 
EFA, these include identifying differences in eigenvalues, the cumulative percent of 
variance extracted by each factor and parallel analysis. It is highly recommended that 
multiple approaches are used to determine the number of appropriate factors.  
Table 4 displays the eigenvalues and differences between eigenvalues for each 
factor. Eigenvalues represent the maximum amount of variance that has not been 
accounted for by previous factors (Suhr, 2006). They are produced by the PAF extraction 
method. Since PAF uses an iterative process to determine eigenvalues, the first factor’s 
eigenvalue represents the greatest amount of variance among items.  As seen in Table 4, 
some eigenvalues are negative. In essence, negative eigenvalues occur because the matrix 
used to calculate them is not of full rank meaning that the rows and columns of the matrix 
are not linearly independent. In other words, the corresponding factors of these negative 
eigenvalues do not explain any additional common variance among items that is not 
already accounted for by previous factors. Consequently, the calculations of these 
eigenvalues overlap with other eigenvalues.  Some authors suggest that negative 
eigenvalues indicate multicollinearity and that redundant variables should be removed. 
However, multicollinearity is not a major problem in factor analysis since its goal is to 
consolidate redundant items into factors.  
------------------------------- 
             Insert Table 4 about here 
------------------------------ 
High eigenvalues should be considered for retention because they represent 
factors that capture substantial proportions of variance (Maroof, 2012). Contrastingly, 
small eigenvalues do not capture adequate proportions of variance and likely reflect error 
46 
 
or the influence of a few number of items. The Kaiser-criterion is a commonly used 
factor extraction method that states that factors that have eigenvalues greater than one 
should be retained. However, this criteria is highly debated and not appropriate for EFA 
because it is based on the assumption that factors that have eigenvalues of greater than 
one consist of items that are able to contribute one unit of variance (Beavers et al., 2013). 
Since EFA only considers common variance, the variance included for each item must be 
less than one (Beavers et al., 2013). In fact, the Kaiser criteria has been shown to 
overestimate the number of factors (Ruscio & Roche, 2012).  
Kaiser-criterion aside, eigenvalues and the proportion of the variance their 
corresponding factors explain are still useful to gain a rudimentary understanding of the 
number of factors that should be extracted. The difference between the first and second 
eigenvalues displayed in Table 4 suggests that this study only has one factor.  A plot of 
the factors and their corresponding eigenvalues, known as a scree plot, better displays the 
dramatic difference between the first and second eigenvalues (see Figure 6). Generally, 
the factors that are most eligible for retention are the ones that appear before the steep 
drop. By this criteria, the scree plot also seems to suggest a one factor solution. In 
addition, the first factor alone explains a substantial proportion of the sample’s variance: 
77%.  Together, these findings mirror the earlier analysis of the polychoric correlation 
matrix which suggested a one factor solution. A one factor solution makes substantive 
and theoretical sense since all items included in this factor analysis have been used in 
research in some form to assess fear of crime.  
Parallel analysis (PA) is yet another way to determine the number of factors for 
retention.  PA generates random parallel datasets. The mean eigenvalues of each factor 
47 
 
generated from these datasets are then compared to the sample’s eigenvalues. It is 
suggested that the factors whose eigenvalues are greater than the eigenvalues generated 
randomly should be considered for retention (Maroof, 2012).  A comparison of EFA and 
PA eigenvalues seems to suggest a 6 factor solution. However, the differences between 
many of these comparisons are substantively negligible.  Setting aside these negligible 
differences, a one or two factor solution gains the most credibility. A one factor solution 
is appropriate because all the item loadings are considerably higher on the first factor 
than the second.7  Moreover, there is no logical substantive or theoretical reason to 
explain what the second factor represents. 
Next, this study created factor scores from a one factor solution using a least 
squares regression approach.  Each participant received a factor score that represented his 
or her standing on the latent concept of fear of crime. Factor scores ranged from a 
minimum of 1.2157 to a maximum of 8.5098.8 These factor scores represent this study’s 
fear of crime outcome. For simplicity, this study will refer to this outcome as a “fear of 
crime index.”  Scores at the higher end of this index represent a greater standing on the 
latent concept of fear of crime. Table 5 displays the descriptive statistics for the fear of 
                                                          
7 This study makes the argument that the difference between the first factor’s eigenvalues observed in the 
PA is more dramatic than all other eigenvalue differences. In addition, the items’ factor loadings -which 
represent the correlation of an item with a factor- on the second factor are not nearly as strong as their 
loadings on the first factor. Depending on the criteria, many of the items could be considered cross-loaded 
between the first and second factor. To be considered cross-loaded, an item’s factor loadings must be 
considered high on more than one factor. The criteria for “high” is subjective. Factor loadings as low as 
±0.30 are often interpreted as high loadings (Costello & Osborne, 2005).  However, cross-loading only 
becomes an issue when more than two factors are eligible for retention. Consequently, cross-loading is not 
an issue for the present study and does not affect the calculations of factor scores attained from a one factor 
solution.   
8 The least squares regression approach to calculating factor scores does not produce factor scores that are 
standardized to a mean of zero as indicated by a positive minimum and maximum factor score. This is 
because this study conducts its EFA using a polychoric correlation matrix. Nonetheless, this study’s 
approach still produces a continuous variable that is approximately normally distributed.  
48 
 
crime index for the whole sample and for each level of the independent variable.  This 
variable is approximately normally distributed and continuous.  
------------------------------- 
           Insert Table 5 about here 
------------------------------ 
Independent Variable  
As mentioned, this study used a computer program to adjust a photograph taken 
of a vacant lot to reflect a disorderly lot, a greened lot and a CPTED lot (i.e., the three 
levels of the study’s independent variable). This study commissioned a graphic designer 
to make these adjustments. Prior to these adjustments, this study provided its graphic 
designer with instructions to create a template image. The goal of creating a template 
image was to develop a blank canvas from which prior adjustments could be easily made 
to simulate each level of the independent variable.  The following items were removed 
from the original photograph to create this template: the telephone pole and telephone 
lines; the wooden, brick and metal fences; the cracks in the street pavement and sidewalk; 
the evergreen tree; and the cars parked in the street (see Figure 1 and Figure 2). After 
these adjustments were made, the graphic designer used this template to construct a 
disorderly, greened and CPTED lot. 
To create a disorderly lot, this study instructed its graphic designer to adjust the 
template to include physical disorder. As previously defined, physical disorder includes 
physical elements of the environment that act as signals of neighborhood dilapidation 
(Skogan, 1990). Moreover, while constructing this disorderly environment the graphic 
designer was instructed to pay particular attention to creating environmental conditions 
that would limit the ability of non-offenders to see into the lot, allow offenders the 
opportunity to hide undetected and include barriers preventing victims from easily 
49 
 
escaping the advances of potential offenders. These three qualities of the environment 
(i.e., blocked prospect, concealment and entrapment) are associated with high levels of 
fear of crime and are staple features of vacant lots (Day, 1994; Fisher, Sloan & Wilkins, 
1995; Nasar & Fisher, 1992; Nasar & Jones, 1997). In particular, the graphic designer 
added the following items to the template image to simulate a disorderly environment: 
overgrown grass, bushes and trees; a pile of tires; piles of cement; trash bags; an 
abandoned car; and miscellaneous debris (see Figure 3)  
This study determined the requirements to create a greened lot by comparing the 
techniques used by major urban greening organizations across the country to transform 
vacant lots (Baltimore Greenspace, 2013; Greening of Detroit, 2014; Pennsylvania 
Horticultural Society, 2014; Urban Green Council, 2014).  These requirements have been 
referred to as traditional urban greening practices and represent the most basic and 
commonly employed urban greening procedures. A unifying and foundational quality of 
all traditional urban greening practices is their transformation of vacant lots from 
disorderly to orderly conditions. In addition, low-maintenance plants, such as trees and 
grass, are planted in an effort to beautify the lot. In particular, trees are often planted in a 
systematic pattern. Many of these features, such as grass and no (physical) disorder, are 
already reflected in the template image. Consequently, the graphic designer was only 
required to include (young) trees arranged in some systematic fashion in the template 
image (see Figure 4). 
 This study used a two-pronged approach to develop the requirements to create a 
CPTED lot. First, an analysis of relevant literature revealed the most foundational and 
agreed upon standards of CPTED (Atlas, 2008; Bell, 1991, 1992; Brantingham & 
50 
 
Brantingham, 1981; Clarke & Mayhew, 1980; Cozens, Saville & Hillier, 2005; Crowe, 
1991; Fennelly & Crowe, 2013; Jeffery, 1971, 1977; Moffat, 1983; Newman, 1972; 
Wallis & Ford, 1981). As previously introduced, these standards include territoriality, 
surveillance, access control and maintenance. Next, this study investigated the ways in 
which these standards are commonly applied to the physical environment (Hedayati et al., 
2012a, 2012b; Minnery & Lim, 2005; Paranaby, 2007; Perkins et al., 1993; Sakip, Johari 
& Salleh, 2012; Sutton, Cherney & White, 2008). From this analysis, this study created a 
table of the underlying objectives of each CPTED standard and the most commonly used 
adjustments to the environment to achieve them (see Table 1). The graphic designer was 
instructed to add these adjustments to the template to simulate a CPTED lot (see Figure 
5). Table 6 outlines the adjustments required to create the template image, a disorderly 
lot, a greened lot and a CPTED lot. This study provided its graphic designer with a copy 
of this table.  
------------------------------- 
          Insert Table 6 about here 
------------------------------ 
As previously mentioned, each participant was randomly assigned one of the three 
images. Random assignment resulted in N=172 participants assigned the disorderly lot 
image, N=172 participants assigned the greened lot image and N=179 participants 
assigned the CPTED lot image.  
Individual-level Factors 
 The present study collected data on several previously discussed individual-level 
factors that are known to influence fear of crime.  In addition to providing insight into the 
characteristics of the sample, this data makes it possible to directly test for equivalence 
among the three experimental conditions.  
51 
 
 Sex. Participants identified their sex as either male or female.  This study 
measured sex as a binary variable that indicates whether participants identified as male or 
female (0=female, 1=male). 
 Race. Participants selected one racial or ethnic category that they identified with 
the most. Responses included “White,” “Asian/Pacific Islander,” “Black/African 
American,” “American Indian or Alaskan Native” or “Some Other Race.” This study 
created four binary variables using these categories: “White,” “Asian/Pacific Islander,” 
“Black/African American” and “Other Race.” The “Other Race” variable includes 
participants who identified as either “American Indian or Alaskan Native” or as “Some 
Other Race.”  
Hispanic. Participant’s identified their ethnic origin as either Hispanic, Latino or 
Spanish or not Hispanic, Latino or Spanish. This study measured Hispanic, Latino or 
Spanish ethnic origin as a binary variable, “Hispanic” (0= not of Hispanic, Latino or 
Spanish origin, 1= Hispanic, Latino or Spanish origin). 
Age. Participants provided their age in years. 
 Hometown. Participant’s selected the description that best represented their 
hometowns. This study defined hometown as the place where an individual spent the 
majority of his or her life.  Responses included “Rural” (a small town or village in the 
country), “Urban” (a large central city) or “Suburban” (a smaller town within a 
metropolitan region but not the core city).  To facilitate responses, this study provided 
examples of places that fit the definitions of each of these categories. This study created 
three binary variables from participants’ responses: “Rural,” “Urban” and “Suburban.”  
52 
 
 Socioeconomic Status. Past research has commonly used receipt of a Federal Pell 
Grant as a proxy for students’ socioeconomic status in lieu of more traditional measures, 
such as the income-brackets of students’ parents/guardians, which are likely to be 
inaccurately reported (Heller, 2007). In light of this research, this study asked participants 
to indicate whether they have ever been the recipient of a Federal Pell Grant. Responses 
included either “Yes” or “No” (0=No, 1=Yes). To facilitate responses, this study also 
provided students with a formal definition of a Federal Pell Grant (a form of student 
financial aid awarded by the government and requires no repayment). 
Direct Victimization Experience.  Participants indicated every crime type that they 
had ever been a victim of within the last 12 months.  These crime types included robbery, 
physical assault, harassment, sexual assault, motor vehicle theft, burglary and other (not 
listed) crime. This study also provided participants with definitions of each crime type 
which were adapted from the Uniform Crime Report (UCR) and from non-governmental 
sources. Furthermore, this study selected the reference period of the last 12 months based 
upon its use by multiple reputable sources to measure victimization experience (e.g., CS, 
GSS, Gallup Organization and NORC).  Due to the scarcity of direct victimization 
experiences,  this study measured direct victimization experience as the binary variable 
“Direct Victimization” which indicates whether participants were the victim of any of the 
previously identified crime types within the last 12 months  (0=no direct victimization 
experience, 1=direct victimization experience) 
Indirect Victimization Experience. Participants indicated every crime type that a 
family member or friend had ever been a victim of in the last 12 months.  To remain 
consistent, the same crime types and reference period that were used to measure direct 
53 
 
victimization experiences were used to measure indirect victimization experiences.  This 
study also provided participants definitions of each crime type. Similar to the direct 
victimization measure, indirect victimization experiences were scarce. For this reason, 
this study measured indirect victimization experience as the binary variable “Indirect 
Victimization” which indicates whether participants were aware of any friends or family 
members who had been the victims of any of the previously identified crime types within 
the last 12 months (0=no indirect victimization experience, 1= indirect victimization 
experience) 
Locus of Control. As previously discussed, it is prudent to identify participants’ 
perceived loci of control when risk of victimization is included in the construction of a 
fear of crime measure. To identify one’s locus of control, Rotter (1966) developed a 29-
item, forced-choice scale. Due to its length, Rotter’s (1966) scale is not often 
implemented when time is at a premium, such as in the case of distributing questionnaires 
in a classroom setting. Because of its impracticality in a timed setting, many scholars 
have sought to validate an abbreviated version of Rotter’s (1966) scale. Lumpkin (1985) 
is one such scholar. 
 Lumpkin (1985) identified six items from Rotter’s (1966) scale. To rate these 
items, Lumpkin (1985) used a five-point Likert scale ranging from “strongly disagree” to 
“strongly agree” (see Table 7).9 The Cronbach’s alpha coefficient for Lumpkin’s (1985) 
scale is .68, making it comparable to coefficients reported by scales developed by Rotter 
(1966) and other scholars (Bugaighis & Schumm, 1983).  Moreover, to establish validity, 
Lumpkin conducted correlations between scores on the abbreviated version of the locus 
                                                          
9 Lumpkin (1985) only indicates what the extreme levels (1 and 5) of the 5-point Likert scale signify. 
Consequently, this present questionnaire does not provide a description for the remaining levels.  
54 
 
of control scale with several other constructs which have been shown to be related to 
locus of control through prior research. Among these constructs, Lumpkin (1985) found a 
negative correlation between internal locus of control and perceived risk, mirroring the 
findings of past research (Rudnick & Deni, 1980).  
------------------------------- 
           Insert Table 7 about here 
------------------------------ 
In light of these demonstrations of internal consistency and validity, Lumpkin’s 
(1985) abbreviated locus of control scale appears to be a worthy substitute for Rotter’s 
(1966) lengthier locus of control scale.  For this reason, this study used Lumkpkin’s 
(1985) abbreviated locus of control scale. Participants’ responses were reverse-coded so 
that larger scores on the five-point scale indicated a more internal locus of control. 
Participants’ scores were then summed and recorded as counts. This study used these 
counts to create the binary variable “Internal” which indicated whether participants had 
internal or external loci of control (0=external locus of control, 1=internal locus of 
control). To create this measure, this study used the cutoff of score of 17, the mean and 
median of the distribution. This cutoff was selected to remain consistent with past 
measurements of this variable (Caliendo, Cobb-Clark & Uhlendorf, 2010).  Participants 
whose total summed scores were greater than 17 were classified as having internal loci of 
control whereas all other participants were classified as having external loci of control.  
Table 8 displays the descriptive statistics of the individual-level factors for the 
whole sample and for each level of the independent variable. Overall, respondents are, on 
average, female (52%), white (64%), non-Hispanic (87%), 20 years of age, from the 
suburbs (75%) and not the recipients of Federal Pell-Grants (74%). In addition, 
participants have, on average, external loci of control (52%) and never been victims, or 
55 
 
have friends or family members who have been victims, of any of the identified crime 
types within the last 12 months (75% and 63%, respectively).10 
------------------------------- 
           Insert Table 8 about here 
------------------------------ 
This study will use these individual-level factors to evaluate equivalence across 
the three levels of the independent variable. In essence, the purpose of evaluating 
equivalence is to assess whether the differences between two or more groups on some 
outcome are small enough to be considered negligible or similar (Rusticus & Lovato, 
2011). Random assignment dictates that individual-level factors should be approximately 
equally distributed across groups. Assessing equivalence helps determine the extent to 
which this is true.  
This study will use a two-proportion z-test for binary outcomes and a two-sample 
t-test from the continuous outcome age to assess equivalence across the three levels of the 
independent variable. The two-proportion z-test tests the null hypothesis that there is no 
statistically significant difference between two population proportions. Alternatively, the 
two-sample t-test tests the null hypothesis that there is no statistically significant 
difference between two population means. These tests must be conducted multiple times 
for equivalence testing involving more than two groups. The present study will compare 
differences in proportions/means between the greened and disorder lot conditions, the 
CPTED and greened lot conditions, and the CPTED and disorder lot conditions for each 
                                                          
10  Demographic statistics on the undergraduate student body were obtained from the large East Coast 
university from which this study’s sample was drawn. The undergraduate student body is predominantly 
male (52.97%), 20 to 21 years of age, white (55.3%) and non-Hispanic (92.1%). Furthermore, 12.1% of the 
undergraduate student body identified as Black or African American and 14.9% as Asian. With the 
exception of sex, there are some substantive similarities between the demographic characteristics of the 
undergraduate student body and the sample of students who participated in this study. 
56 
 
individual-level factor, totaling 42 comparisons. A benefit of this approach is that it 
offers a more comprehensive account of the nature of the differences across groups.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
57 
 
Chapter 4: Results  
 
Using the individual-level factor variables, this study first checked for balance 
across the three levels of the independent variable using a two-proportion z-test for 
binary outcomes and a t-test for the continuous outcome Age. Table 9 displays the results 
of these tests. This study used a cut-off value of p< 0.05 to identify statistically 
significant differences in proportions and means. Of these comparisons, the only 
statistically significant differences across levels was whether participants were the 
recipients of a Federal Pell Grant and whether they had been victimized in the last 12 
months. More specifically, Federal Pell Grant recipients were more likely to have been 
assigned a greened or CPTED lot than a disorderly lot. Furthermore, participants who had 
been victimized in the last 12 months were more likely to have been assigned a CPTED 
lot than a disorderly lot. Due to their imbalance across the levels of the independent 
variable, both the Federal Pell Grant and Direct Victimization measures were included in 
subsequent OLS regression models as controls.11   
------------------------------- 
           Insert Table 9 about here 
------------------------------ 
This study conducted OLS regression models to test its hypotheses. This study 
tested its first hypothesis by analyzing the slope coefficients of the Greened and CPTED 
dummy variables against the Disorder reference category. To test its second hypothesis, 
this study used a post-estimation t-test to compare participants’ relative standings on the 
latent concept of fear of crime who viewed a CPTED lot to those who viewed a greened 
                                                          
11  It may be the case that this imbalance is the result of random chance due to the number of tests that were 
run and do not identify real differences across the three levels of the independent variable. To account for 
this possibility, Table 11 depicts the OLS regression model without the controls Federal Pell Grant and 
Direct Victimization. The results of this model are substantively similar to the results of the model in which 
these measures were included as controls.   
58 
 
lot. Here, the Greened and CPTED dummy variables are compared relative to the 
Disorder reference category. This comparison reflects the fact that disorderly vacant lots 
are the starting place of many urban-greening interventions.  To reflect this reality, this 
study is concerned with potential reductions in fear of crime that occur from the 
transformation of disorderly vacant lots into CPTED lots over greened lots and not from 
the transformation of greened lots into CPTED lots.12    
 Table 10 displays the OLS regression results.13 Holding all else constant, 
participants who viewed either a greened or CPTED lot had significantly lower standings 
on the latent concept of fear of crime than participants who viewed a disorderly lot. In 
particular, participants who viewed a CPTED lot scored, on average, 1.6164 points lower 
on the fear of crime index than participants who viewed a disorderly lot, whereas 
participants who viewed a greened lot scored, on average, 1.2933 points lower. This 
finding is consistent with this study’s first hypothesis.  
------------------------------- 
            Insert Table 10 about here 
------------------------------ 
Furthermore, a post-estimation t-test identified a statistically significant difference 
in participants’ standings on the latent concept of fear of crime between those who 
viewed a greened lot and those who viewed a CPTED lot. The magnitude and the sign of 
the slopes of these two variables suggest that, on average, participants who viewed a 
                                                          
12  For the reason previously described, this study is not interested in the direct comparison of CPTED and 
greened lots. Nonetheless, this study conducted a separate analysis in which disorderly lot cases were 
dropped so that CPTED lots could be directly compared to greened lots, the reference category, using OLS 
regression (𝛽=-0.31313, p=0.053). Both the Federal Pell Grant and Direct Victimization measures were 
included in this model as controls.  
13 This study also tested its hypotheses using factor scores produced using a Pearson correlation matrix. The 
results of these hypothesis tests were substantively similar to those found using factor scores produced from 
a polychoric correlation matrix.  
59 
 
CPTED lot had lower standings on the latent concept of fear of crime than participants 
who viewed a greened lot. This finding lends support to this study’s second hypothesis.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
60 
 
Chapter 5: Discussion & Conclusion  
 The present study is the first of its kind to assess the effect of incorporating 
CPTED standards into traditional urban greening practices of vacant lots on fear of crime. 
Supporting its first hypothesis, this study found that the traditional and CPTED 
approaches to greening vacant lots were associated with lower levels of fear of crime 
compared to an untransformed, disorderly vacant lot. This finding is not surprising given 
past research on the effects of environmental features, most notably physical disorder, on 
fear of crime. Moreover, this finding suggests that traditional urban greening practices 
and the CPTED adjustment are both viable options to reduce fear of crime. More 
interestingly, this study found that incorporating CPTED into traditional urban greening 
practices was associated with lower levels of fear of crime than the traditional approach.  
This comparison was made relative to a disorderly vacant lot in order to reflect the reality 
of when CPTED would be implemented in the greening process. This finding, which 
supports this study’s primary hypothesis, suggests that incorporating CPTED standards 
into traditional urban greening practices may help optimize reductions in fear of crime in 
neighborhoods.  
The reduction in fear of crime achieved by incorporating CPTED standards into 
traditional urban greening practices may have additional related benefits for 
communities. For example, if individuals feel safer they will be more likely to use public 
spaces and engage with other community members (Pitner, Yu & Brown, 2012). These 
activities facilitate collective efficacy and informal social control which make 
neighborhoods safer (and more desirable) places to live (Gibson et al., 2006).  In 
particular, research has shown that neighborhoods with high levels of collective efficacy 
61 
 
and informal social control are more intolerant to crime and more active to abate it 
(Gibson et al., 2002; Silver & Miller, 2004).  
This study has several theoretical and policy implications. This study’s 
construction of its fear of crime measure contributes to a major theoretical debate 
underlying fear of crime research. This debate involves how to define and subsequently 
measure fear of crime. In particular, this study found that perceptions of safety, risk of 
victimization and fear of victimization all measure the same underlying latent concept. In 
other words, this study did not find evidence to suggest that these fear of crime elements 
are conceptually distinct and that using one fear of crime measure over another provides 
added value in regards to capturing one’s level of fear of crime. This finding supports 
past research which questions fear of crime’s multi-dimensional character and the unique 
contributions of its cognitive and affective elements advocated by some scholars (Ferraro 
& LaGrange, 1987; Marzbali et al., 2012).    
This study’s findings also have theoretical import in regards to CPTED’s ability 
to impact fear of crime. Recall that CPTED relies on its ability to influence individuals’ 
perceptions of the environment which in turn inform their levels of fear of crime. This 
study’s findings suggest that CPTED is able to influence individuals’ perceptions of the 
environment that impact their levels of fear of crime in artificial (or computer-generated) 
environments.  More generally, individuals are able to identify and interpret cues present 
in artificial environments that affect their fear of crime responses. This study’s findings 
support the use of artificial environments to identify the impact of CPTED on fear of 
crime as well as to investigate the underlying mechanisms by which this relationship 
operates.  
62 
 
Furthermore, this study is able to provide some direction for future policy 
regarding urban greening despite the restricted generalizability of its findings.  Insofar as 
fear of crime is linked to a host of negative outcomes that debilitate communities, this 
study’s findings suggest that city officials need to more deliberately consider what can be 
done to optimize reductions in fear of crime in the transformation of vacant lots. To this 
point, city officials should investigate whether incorporating CPTED standards into 
traditional urban greening practices is a worthwhile endeavor to pursue in order to 
optimize reductions in fear of crime in their communities. City officials should conduct 
some preliminary investigations on this topic that are low-cost and easy to implement 
before action is taken to incorporate CPTED standards into the greening process.14   
For example, one low-cost, easy way in which city officials could assess the 
impact of incorporating CPTED into the greening process in regards to fear of crime is to 
host discussions on this topic with community members at community centers. To 
facilitate discussion, city officials could present images of vacant lots that have been 
greened according to traditional and CPTED practices. This approach would allow city 
officials to gain a nuanced understanding of fear of crime in communities and insight into 
whether incorporating CPTED standards into traditional greening practices would 
provide added benefits, such as increases in quality of life, beyond the traditional 
endeavor.  From here, city officials could decide whether it is worthwhile to develop 
                                                          
14  It is important to point out that what is considered to be a low-cost, easy evaluation likely varies across 
communities; what may be a low-cost evaluation for one community may be an expensive evaluation for 
another. In general, communities in which vacant lots are the most pervasive have limited financial 
resources.  In such places, it is not feasible to invest a substantial amount of money in preliminary 
investigations to assess whether incorporating CPTED into traditional urban greening practices is a 
worthwhile endeavor to pursue. In the discussion that follows, this study focuses on what low-cost, easy 
evaluations may resemble for communities with the most restricted budgets.  
63 
 
more costly and complex methods to evaluate whether incorporating CPTED standards 
into the greening process optimizes reductions in fear of crime over traditional practices.  
In particular, it is important that city officials consider the financial and social 
costs and benefits of one greening approach over another. The simplistic methods of the 
traditional urban greening approach likely means that it will be less financially costly to 
implement and maintain than the more intricate CPTED adjustment. However, the higher 
financial cost of greening vacant lots according to CPTED standards and maintaining 
these lots may produce greater social and financial benefits related to optimizing 
reductions in fear of crime over the less costly traditional greening alternative. These 
social and financial benefits may include greater improvements in quality of life and 
decreases in crime and the cost of crime in communities.15  Less-costly evaluations, such 
as the one previously described, are able to provide some insight into the extent of these 
benefits. However, more costly and statistically rigorous community evaluations on this 
topic are needed in order to gain a more comprehensive understanding of the nature and 
type of benefits that can be achieved by incorporating CPTED into traditional greening 
practices.16  
                                                          
15  Cost of crime research evaluates several direct and indirect economic consequences of crime related to, 
for example, criminal justice system costs, victim costs and intangible costs.  Direct consequences of crime 
may include the cost of neighborhood policing and property damage/loss. Indirect consequences, which are 
more difficult to measure in dollars, may include costs associated with victims’ residual pain from a 
criminal event and their loss of quality of life (Miller, Cohen & Wiersema, 1996).  
16 What may be a high-cost evaluation to one community may be a low-cost evaluation to another. For 
example, mail surveys are typically considered a low-cost evaluation strategy. However, they may be 
beyond the reach of some city officials to implement.  For those that have the financial means to do so, city 
officials could organize mail surveys to be distributed- that resemble those of the present study- to a 
random sample of households. This approach would allow city officials to identify a treatment effect of 
greening vacant lots incorporating CPTED standards on fear of crime over traditional practices. 
Furthermore, city officials could also implement a pre-posttest quasi-experimental design in which they 
select segments of a community to transform vacant lots by traditional greening practices or by the CPTED 
adjustment. They could then compare the levels of fear of crime and crime as well as perceptions of quality 
of life around these lots before and after these interventions. This more costly and time consuming 
approach would allow city officials to better understand the related benefits of optimizing reductions of fear 
64 
 
This study offers several directions for future research. First, future research 
should attempt to dissect the effect of each CPTED standard on fear of crime. As 
previously explained, the interconnectedness of CPTED standards makes it very difficult 
to separately distinguish their effects on fear of crime. Consequently, the present study 
evaluates CPTED as a package. However, it may be the case that only some standards are 
necessary to reduce fear of crime and that implementing CPTED as a package is not a 
very cost efficient option for greening vacant lots. To speak to this point, communities in 
which vacant lots are the most prevalent often have limited budgets devoted to 
transforming vacant lots through urban greening. In these communities, greening 
practices that are able to optimize reductions in fear of crime are arguably needed the 
most.  For this reason, it would be helpful to evaluate the effect of each CPTED standard 
on fear of crime in order to determine whether this greening method can be made more 
affordable for communities that need it the most.  
Second, future research should investigate the mechanisms underlying this study’s 
treatment effect. This study did not attempt to understand how disorderly, greened and 
CPTED lots affected participants’ fear of crime responses. Rather, like most experiments, 
it sought to identify a causal treatment effect. Open-ended survey questions and 
qualitative interviews both allow researchers to identify underlying mechanisms. In doing 
so, these techniques may provide some leverage over conducting evaluations of the effect 
of each CPTED standard on fear of crime.  For example, researchers could add an open-
ended question to this study’s questionnaires that asks participants to describe what it is 
about a particular lot that makes them more or less afraid of falling victim to a crime. The 
                                                          
of crime and whether it is worthwhile to incorporate CPTED standards into traditional greening practices 
on a larger scale.  
65 
 
results from this open-ended question could be compared to what is known about the 
connection between the physical environment and fear of crime from theory. However, it 
should be noted that open-ended survey questions will increase the amount of time 
subjects need to complete questionnaires. For this reason, open-ended survey questions 
may be especially problematic when surveys are administered in classroom settings 
where time is limited.  Participants are also more likely to skip open-ended questions due 
to the effort needed to complete them (Fowler, 2014). When participants do respond, 
their response may range from very detailed to of minimal use (Fowler, 2014).  
Researchers could also conduct in-person qualitative interviews to gain insight 
into underlying mechanisms. For example, researchers could show participants all three 
lot conditions and ask them to compare the images and rate how safe they would feel at 
each and why. Researchers will need to devout a considerable amount of time and 
resources to implement this approach. However, a major benefit is that they will be able 
to gain a deeper, more descriptive understanding of what it is about each lot that affects 
participants’ levels of fear of crime. In doing so, researchers are more likely to develop a 
more comprehensive understanding of the underlying mechanisms influencing 
participants’ levels of fear of crime which they can compare to relevant theory. 
Third, there are several other types of evaluations that future research should 
conduct to better understand the effect of disorderly, greened and CPTED lots on fear of 
crime. In particular, research shows that fear of crime is higher at night than during the 
day (Nasar & Jones, 1997). This study’s findings are specific to daytime conditions. It 
remains to be seen whether the gains in fear of crime reduction achieved by CPTED over 
traditional practices hold for nighttime conditions. Furthermore, future research should 
66 
 
also investigate demographic differences in fear of crime across disorderly, greened and 
CPTED lots. Investigating these differences will help researchers gain a more nuanced 
understanding of fear of crime and, in particular, which groups of individuals benefit the 
most-in regards to reductions in fear of crime-from a particular greening approach.   
Last, this study’s findings suggest that it is worthwhile to purse a large-scale 
investigation of the effect of incorporating CPTED into traditional urban greening 
practices on fear of crime that utilizes more sophisticated computer software to create lot 
conditions and a sampling technique that captures participants that are representative of a 
future undetermined target population. As previously described, there are several more 
sophisticated alternatives to computer generated images, such as virtual environments 
and panoramic imaging, which were beyond the scope and means of this study. Future 
research should investigate these alternatives in light of their contributions to ecological 
validity. In addition, this study’s student sample, although appropriate for its research 
purpose, greatly restricts the generalizability of its findings. To achieve greater external 
validity, future research should adopt sampling techniques that are better able to capture 
representative estimates of a population. Ideally, these sampling techniques should focus 
on capturing those individuals who are most likely to be exposed to disorderly vacant lot 
conditions in their communities.  
 It may also be worthwhile for researchers to study the effect of traditional and 
CPTED greening approaches on fear of crime using a sample of individuals who are not 
residents of the neighborhoods under evaluation but who frequent these neighborhoods 
for commercial purposes or who are prospective homeowners. Communities that contain 
disorderly vacant lots lack the vibrancy needed to attract outsiders to invest in their 
67 
 
welfare. To elaborate, environmental cues associated with vacant lots (e.g., trash, graffiti 
and over grown vegetation) promote fear of crime and dissuade outsiders from 
frequenting local businesses and purchasing homes in these communities. Lack of 
support from outsiders can severely stifle local economies. In order to help revitalize 
communities, it is not only important to consider the impact traditional and CPTED 
greening approaches have on residents’ fear of crime, but also the impact they have on 
the fear of crime of those outsiders who can help transform local economies.  
In conclusion, this study has provided a starting place from which future research 
can begin investigating the impact of greening vacant lots incorporating CPTED 
standards on fear of crime. This study has demonstrated that greening vacant lots 
incorporating CPTED standards is capable of reducing fear of crime beyond traditional 
practices. Greening incorporating CPTED standards not only may optimize reductions in 
fear of crime in (blighted) communities but may also lead to substantial improvements in 
other outcomes affected by fear of crime, such as quality of life, beyond that which could 
be achieved by traditional greening practices alone. Future research should heed this 
study’s recommendations for developing this research area.   
 
 
 
 
 
 
 
68 
 
Appendix A 
Introduction Script 
Hello, my name is Alaina De Biasi. I am a graduate student in Criminology and I 
am conducting a study for my thesis on the impact of the environment on fear of crime. I 
would greatly appreciate your help by completing a questionnaire. You must be at least 
18 years of age to participate in this study. Also, if you have already completed this 
questionnaire you cannot participate in this study a second time. I am now going to 
distribute questionnaires along with consent forms. Please do not do anything with these 
documents until I tell you to.  
 
The questionnaire is nine-pages in length and should take approximately 10 to 15 
minutes to complete. The first section asks you to report some demographic information 
such as your age and sex. It also asks if you or your friends or family members have been 
a victim of any of the following crime types: robbery, physical assault, harassment, 
sexual assault, motor vehicle theft, burglary or some other crime. The questionnaire 
provides definitions of these crime types. The second section provides you with a picture 
that you will use as a reference to answer questions related to fear of crime. Finally, the 
last section asks you to consider the extent to which you personally agree or disagree 
with a series of statements.  
 
Please read the consent form first. You can keep the consent form for your 
records. After reading the consent form, if you choose to participate in this study please 
complete the questionnaire. In order to protect your privacy, do not sign your name on 
the questionnaire. By not signing your name on the questionnaire your responses will 
remain completely anonymous.  
 
Please raise your hand when you would like to turn in your questionnaire and 
someone will come collect it from you. If you choose not to participate in this study, you 
must still turn in your questionnaire. Once you have turned in your questionnaire, please 
wait patiently while other students complete their questionnaires. After 15 minutes have 
elapsed, I must collect all remaining questionnaires.  
 
Please feel free to contact me if you have any questions regarding this study or 
would like to receive a summary of its results. My contact information is included on the 
consent form. You may now begin reviewing the consent form and if you so choose 
move forward with completing the questionnaire. 
 
 
 
 
 
 
 
 
 
 
69 
 
 
Figure 1. Original image of the vacant lot  
 
Figure 2. Template image  
 
70 
 
Figure 3. Disorderly lot 
Figure 4. Greened lot 
 
71 
 
Figure 5. CPTED Lot 
 
Figure 6. Scree Plot 
 
 
 
72 
 
Table 1.  
CPTED Standards  
Territoriality Surveillance Access Control  Maintenance  
Objective Aims to promote a 
sense of ownership 
and facilitate 
proprietary 
concern 
Promotes an 
individual’s ability 
to have clear views 
of his or her 
surroundings 
Involves constructing 
a design that manages 
the safe usage and 
movement of 
individuals into and 
out of a space, with a 
focus on decreasing 
opportunities for 
illegitimate use 
Refers to all 
efforts required 
to uphold a tidy 
and orderly 
appearance of a 
space 
Adjustments  Fencing 
 Signage 
 Artwork 
 Flowers & low 
lying bushes 
 Lighting  Clearly defined 
entrance and exist 
 Defined route 
 Fencing 
 Removal of 
debris 
 Groomed 
foliage 
 
Table 2. 
 Fear of Crime Descriptive Statistics  
 Full Sample Disorder Greened CPTED 
  N=523 N= 172 N= 172 N= 179 
Variable ?̅? SD ?̅? SD ?̅? SD ?̅? SD 
Safety         
Robbery 4.2906 1.6341 5.2326 1.2582 3.9651 1.4981 3.6983 1.6855 
Physical Assault 4.3098 1.7896 5.2209 1.3456 3.9942 1.7219 3.7374 1.8912 
Harassment 4.2753 1.6332 5.1221 1.3901 3.9884 1.5967 3.7374 1.5663 
Sexual Assault 4.3881 1.9963 5.0698 1.7888 4.1802 1.9845 3.9330 2.0323 
Afraid         
Robbery 3.8757 1.6786 4.7209 1.4481 3.6395 1.5553 3.2905 1.6842 
Physical Assault 3.8872 1.7378 4.6802 1.4896 3.6047 1.6134 3.3966 1.8157 
Harassment 3.7400 1.6966 4.5698 1.6511 3.4128 1.5962 3.2570 1.5401 
Sexual Assault 3.8987 1.9405 4.6047 1.8275 3.6512 1.8904 3.4581 1.9145 
Likely         
Robbery 3.5679 1.5139 4.3547 1.4170 3.4302 1.3937 2.9441 1.3850 
Physical Assault 3.3270 1.4418 4.1512 1.4430 3.1163 1.2970 2.7374 1.1961 
Harassment 3.6176 1.4981 4.3372 1.5530 3.3430 1.3610 3.1899 1.3144 
Sexual Assault 3.2027 1.5662 3.9360 1.6898 3.0465 1.4421 2.6480 1.2650 
 
73 
 
Table 3.  
Polychoric Correlation Matrix  
 
S
a
f
e
-
R
o
b
b
e
r
y
S
a
f
e
-
P
h
y
s
i
c
a
l
 
A
s
s
a
u
l
t
S
a
f
e
-
H
a
r
a
s
s
m
e
n
t
S
a
f
e
-
S
e
x
u
a
l
 
A
s
s
u
a
l
t
L
i
k
e
l
y
-
R
o
b
b
e
r
y
L
i
k
e
l
y
-
P
h
y
s
i
c
a
l
 
A
s
s
a
u
l
t
L
i
k
e
l
y
-
H
a
r
a
s
s
m
e
n
t
L
i
k
e
l
y
-
S
e
x
u
a
l
 
A
s
s
a
u
l
t
A
f
r
a
i
d
-
R
o
b
b
e
r
y
A
f
r
a
i
d
-
P
h
y
s
i
c
a
l
 
A
s
s
a
u
l
t
A
f
r
a
i
d
-
H
a
r
a
s
s
m
e
n
t
A
f
r
a
i
d
-
S
e
x
u
a
l
 
A
s
s
a
u
l
t
S
a
f
e
-
R
o
b
b
e
r
y
1
.
0
0
0
0
S
a
f
e
-
P
h
y
s
i
c
a
l
 
A
s
s
a
u
l
t
0
.
8
8
9
9
1
.
0
0
0
0
S
a
f
e
-
H
a
r
a
s
s
m
e
n
t
0
.
8
0
6
3
0
.
8
1
6
6
1
.
0
0
0
0
S
a
f
e
-
S
e
x
u
a
l
 
A
s
s
a
u
l
t
0
.
7
9
3
7
0
.
8
8
2
0
0
.
7
9
7
7
1
.
0
0
0
0
L
i
k
e
l
y
-
R
o
b
b
e
r
y
0
.
5
0
2
3
0
.
4
5
8
0
0
.
4
7
6
1
0
.
4
6
3
5
1
.
0
0
0
0
L
i
k
e
l
y
-
P
h
y
s
i
c
a
l
 
A
s
s
a
u
l
t
0
.
5
0
6
3
0
.
5
0
4
6
0
.
5
3
0
1
0
.
4
8
7
4
0
.
8
4
1
7
1
.
0
0
0
0
L
i
k
e
l
y
-
H
a
r
a
s
s
m
e
n
t
0
.
4
2
3
9
0
.
4
3
2
5
0
.
5
6
6
9
0
.
4
7
5
1
0
.
7
2
1
5
0
.
7
7
2
0
1
.
0
0
0
0
L
i
k
e
-
S
e
x
u
a
l
 
A
s
s
a
u
l
t
0
.
4
3
8
9
0
.
4
3
4
9
0
.
5
3
5
3
0
.
5
6
5
9
0
.
6
6
6
1
0
.
7
6
5
2
0
.
7
0
9
6
1
.
0
0
0
0
A
f
r
a
i
d
-
R
o
b
b
e
r
y
0
.
6
2
0
5
0
.
5
7
7
2
0
.
5
7
3
0
0
.
5
6
3
8
0
.
6
3
7
4
0
.
5
9
1
7
0
.
5
3
4
6
0
.
5
2
2
0
1
.
0
0
0
0
A
f
r
a
i
d
 
P
h
y
s
i
c
a
l
 
A
s
s
a
u
l
t
0
.
6
0
2
5
0
.
6
4
5
7
0
.
6
2
1
4
0
.
6
3
1
0
0
.
5
7
2
7
0
.
6
1
3
3
0
.
5
5
6
4
0
.
5
6
2
8
0
.
9
1
6
2
1
.
0
0
0
0
A
f
r
a
i
d
-
H
a
r
a
s
s
m
e
n
t
0
.
5
4
4
2
0
.
5
4
4
9
0
.
6
7
7
4
0
.
5
8
3
0
0
.
5
3
5
8
0
.
5
5
8
2
0
.
6
4
0
1
0
.
6
0
4
9
0
.
8
1
0
3
0
.
8
4
5
0
1
.
0
0
0
0
A
f
r
a
i
d
-
S
e
x
u
a
l
 
A
s
s
a
u
l
t
0
.
5
5
6
8
0
.
5
8
2
8
0
.
5
9
3
0
0
.
7
0
2
4
0
.
5
0
6
3
0
.
5
4
4
1
0
.
5
6
1
2
0
.
6
8
1
6
0
.
7
9
9
3
0
.
8
7
8
0
0
.
8
3
9
1
1
74 
 
Table 4.  
Eigenvalues 
Factor Eigenvalue Difference 
Factor 1 7.7332  
Factor 2 1.2599 6.4733 
Factor 3 0.8569 0.4030 
Factor 4 0.3537 0.5033 
Factor 5 0.1889 0.1648 
Factor 6 0.0442 0.1447 
Factor 7 0.0123 0.0319 
Factor 8 -0.0272 0.0395 
Factor 9 -0.0515 0.0243 
Factor 10 -0.0649 0.0134 
Factor 11 -0.0768 0.0119 
Factor 12 -0.0787 0.0018 
 
Table 5.  
Fear of Crime Index Descriptive Statistics  
 
 
 
 
 
 
 
 
 
Variable SD SD SD SD
Fear of Crime Index 4.717 1.6433 5.6933 1.4509 4.4095 1.4964 4.0744 1.5225
Full Sample Disorder Greened CPTED
N=523 N= 172 N= 172 N=179
75 
 
Table 6.  
 
Adjustments  
Template Disorderly Traditional 
Practices 
CPTED 
Adjustments The following 
features will be 
removed:  
 Telephone 
pole and 
telephone 
lines  
 Evergreen  
 Wooden, brick 
and metal 
fences  
 Cracks in the 
street 
pavement and 
sidewalk  
  Evergreen 
 Cars parked in 
the street 
The following 
features will be 
added: 
 Overgrown 
grass, bushes 
and trees  
 A pile of tires  
 Piles of 
cement 
 Trash bags  
 Abandoned 
car 
 Miscellaneous 
debris  
The following 
features will be 
added: 
 Trees planted 
in a systematic 
fashion 
The following 
features will be 
added: 
 Fencing  
 Signage 
 Artwork 
 Flowers and 
low-lying bushes 
 Clearly defined 
exit and entrance 
 Clearly defined 
route 
 Lighting 
 
 
 
 
Table 7.  
 
Lumpkin’s Abbreviated Locus of Control Scale 
 
Internals   Externals  
When I make plans, I am almost certain that I 
can make them work. 
  
Many of the unhappy things in people's lives are 
partly due to bad luck. 
 
Getting people to do the right things depends 
upon ability; luck has nothing to do with it.  
  
Getting a good job depends mainly on being in 
the right place at the right time. 
 
What happens to me is my own doing. 
  Many times I feel that I have little influence over 
the things that happen to me. 
76 
 
Table 8.  
Individual-level Characteristics Descriptive Statistics 
 
 
 
 
 
 
 
77 
 
Table 9. 
Assessment of Equivalence   
Variable Greened & Disorder CPTED & Greened CPTED & Disorder 
Sex -0.6472 -0.136 -0.7895 
Age 0.3279 -1.0833 -0.6505 
Pell Grant 2.3751* -0.3581 2.0426* 
Direct Victimization 1.1752 1.4335 2.6046** 
Indirect Victimization 0.0000 0.6985 0.6985 
Suburban -1.3644 0.7083 -0.6715 
Rural 0.4628 -0.1558 0.3116 
Urban 1.376 -0.7957 0.5996 
White -0.5646 -0.0473 -0.6173 
Black 1.5298 -0.3107 1.2358 
Hispanic -0.3229 0.181 -0.1451 
Asian -0.1599 0.3234 0.162 
Other Race -0.8159 0.1168 -0.7089 
Internal 0.1081 0.9152 1.0243 
Note. The values presented in this table for all variables other than Age are z-statistics. The values presented for 
the variable Age are t-statistics. 
 *p<0.05, **p<0.05 
    
 
Table 10.  
OLS Regression 
Variable 𝛽 Coefficient Standard Error  
Greened -1.2933** 0.1618  
CPTED -1.6164A** 0.1608 
Direct Victimization -0.1369 0.1516735  
Pell Grant 0.1507 0.1497  
N=523, *p<0.05, **p<0.01   
A = This value is significantly different from the greened value  
 
 
 
 
78 
 
Table 11. 
OLS Regression with No Controls   
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Variable 𝛽 Coefficient Standard Error  
Greened -1.2838** 0.1608    
CPTED -1.6190 ** 0.1591 
N=523, *p<0.05, **p<0.01   
A = This value is significantly different from the greened value  
79 
 
References 
Abdullah, A., Hedayati, M. M. & Maghsoodi, T. M. J. (2013). Predicting the influence of  
CPTED on perceived neighbourhood cohesion: considering differences across  
age. Journal of Environmental Psychology, 36(3), 54-64. 
 
Abdullah, A., Hedayati, M. M., Haron, S. H., Bahauddin, A. & Maghsoodi, T. M. J.  
(2013). A cross-cultural validation of the territorial functioning construct in 
residential neighbourhoods: A multigroup invariance analysis. Review of 
European Studies, 5(5), 30-39. 
 
Alper M., & Chappell A.T. (2012). Untangling fear of crime: A Multi-theoretical  
Approach to examining the causes of crime-specific fear. Sociological 
Spectrum, 32(4), 346-363.  
 
Andrews, M. & Gatersleben, B. (2010). Variations in perceptions of danger, fear and  
preferences in a simulated natural environment. Journal of Environmental 
Psychology, 30, 473-481. 
 
Angel, Schlomo. (1968). Discouraging crime through city planning. (Paper No. 75).  
Berkeley, CA: Center for Planning and Development Research, University of 
California at Berkeley. 
 
Appleton, J. (1975). The Experience of Landscape. London: Wiley. 
 
Armitage, R. (2002). To cctv or not to cctv? a review of current research into the  
effectiveness of cctv systems in reducing crime. London: Nacro, Crime and Social 
Policy Section. 
 
Arnold, H. (1991). Fear of crime and its relationship to directly and indirectly  
experienced victimization: A binational comparison of models. In K. Sessar and 
H. J. Kerner,(Eds.), Developments in crime and crime control research: German 
studies on victims, offenders, and the public, Springer-Verlag; New York. 
 
Atlas,  R (1986). Crime prevention through building codes. Journal of Security 
Administration, 9(2), 3-12. 
 
Austin, D. M., & Sanders, C. (2007). Graffiti and perceptions of safety: A pilot study 
using photographs and survey data. Journal of Criminal Justice and Popular 
Culture, 14(4), 292-316.  
 
Baltimore Greenspace. (2013). Retrieved from http://baltimoregreenspace.org/ 
 
Baumer, T.L. (1985). Testing a general model of fear of crime: Data from a national 
sample. Journal of Research in Crime and Delinquency, 22(3), 239-255. 
 
80 
 
Beavers, Amy S., Lounsbury, John W., Richards, Jennifer K., Huck, Schuyler W., 
Skolits, Gary J. & Esquivel, Shelley L. (2013). Practical considerations for using 
exploratory factor analysis in educational research. Practical Assessment, 
Research & Evaluation, 18(6). 
 
Bell, W. (1991). The role of urban design in crime prevention. Australian Planner, 29(4)  
206-210. 
 
Bell, W. (1992). Crime prevention: A planning and urban design approach (Research 
report). Adelaide: Crime Prevention Unit, SA Attorney General's Department. 
 
Bennett, R. R., & Flavin, J. M. (1994). Determinants of fear of crime: The effect of  
cultural setting. Justice Quarterly, 11, 357-381. 
 
Bernstein, I. H., & Teng, G. (1989). Factoring items and factoring scales are different: 
Spurious evidence for multidimensionality due to item categorization. 
Psychological Bulletin, 105(3), 467–477.  
 
Bickman, L., Maltz, M.D. & Lavrakas, P.J. (1977). The evaluation of Crime prevention 
through  
environmental design programs. USA: Westinghouse Electric Corporation.  
 
Bishop, I. D. & Rohrmann, B. (2003). Subjective responses to simulated and real 
environments: A comparison. Landscape and Urban Planning, 65(4), 261-277. 
 
Blakely, E. J. & Snyder, M. G. (1997). Fortress America. Washington, DC: Brookings  
Institution. 
 
Bowman, A.O. & Pagano, M.A. (2000). Vacant land in cities: an urban resource. The 
Bookings Institution: Center of Urban on Urban and Metropolitan Policy, 1-9. 
 
Braga, A. A. (2008). Problem-oriented policing and crime prevention. Second edition. 
Boulder, CO: Lynne Rienner Publishers. 
 
Branas, C. C., Cheney, R. A., MacDonald, J. M., Tam, V. W., Jackson, T. D., & Ten, H. 
T. R. (January 01, 2011). A difference-in-differences analysis of health, safety, 
and greening vacant urban space. American Journal of Epidemiology, 174(11), 
1296-306. 
 
Brantingham, P. J., & Brantingham, P. L. (1981). Environmental criminology. Beverly 
Hills, CA: Sage Publications. 
 
Brower, S. (1980). Territory in urban settings. In A.A. Rapoport, & J. F. Wohlwill  
(Eds.), Culture and environment. human behavior and environment, 4. New York:  
Plenum. 
 
81 
 
Brower, S., Dockett, K. & Taylor, R. B. (1983). Residents' perceptions of territorial 
features and perceived local threat. Environment and Behavior, 15(4), 419-437.  
 
Brown, B. B., & Werner, C. M. (1985). Social cohesiveness, territoriality and holiday  
decorations: the influence of cul-de-sacs. Environment and Behavior, 17, 539-
565. 
 
Bugaighis, M. & Walter R. (1983). Alternative measures of perceived locus of control.  
Psychological Reports, 52, 819-823. 
 
Bureau of Justice Assistance, U.S. Department of Justice. (2015). Center for Program  
Evaluation and Performance Measurement. Retrieved from 
https://www.bja.gov/evaluation/program-crime-prevention/cpted1.htm. 
 
Bureau of Justice Statistics. (2013). National crime victimization survey. Retrieved from  
http://www.bjs.gov/index.cfm?ty=dcdetail&iid=245 
 
Burkholder, S. (2012). The new ecology of vacancy: rethinking land use in shrinking  
cities. Sustainability, 4(6), 1154-1172. 
 
Caliendo, M., Cobb-Clark, D. & Uhlendorff, (2010) A. Locus of control and job 
strategies. (Discussion Paper No. 4750). 
 
Casteel, C. & Peek-Asa, C. (2000). Effectiveness of crime prevention through 
environmental  
design (CPTED) in reducing robberies. American Journal of Preventive 
Medicine, 18(4), 99-115. 
 
Challinger, Dennis. (1997). Refund fraud in retail stores. In R. V. Clarke (Ed.),  
Situational crime prevention: Successful case studies (Second Edition). Albany, 
NY: Harrow and Heston. 
 
Cinar, E. A., & Cubukcu, E. (2012). The influence of micro scale environmental  
characteristics on crime and fear. Procedia - Social and 
BehavioralvSciences, 35(2), 83-88. 
 
Clarke, R. V. (1995). Situational crime prevention. In M. Tonry & D. Farrington  
(Eds.), Crime and justice: building a safer society (pp. 91-150). Chicago: Chicago 
University Press. 
 
Clarke, R. V. & Derek B. C. (1985). Modeling offenders' decisions: a framework for 
research and policy. In M. Tonry and N. Morris (Eds.), Crime and Justice: an 
annual review of research (pp. 147-185). Chicago: University of Chicago Press. 
 
Clarke, R.V. & Mayhew P. (Eds.). (1980). Designing out crime. London: Home Office 
Research and Planning Unit (HMSO). 
82 
 
 
Clifton W. (1987). Convenience store robberies in Gainesville, Florida: An intervention 
strategy by the Gainesville Police Department. Gainesville, FL: Gainesville Police 
Department. 
 
Cohen, J. (1970). Approximate power and sample size determination for common one-
sample and two sample hypothesis tests. Educational and Psychological 
Measurement, 30, 811–831. 
 
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd 
ed.). Hillsdale, NJ: Lawrence Erlbaum. 
 
Costello, Anna B. & Jason Osborne (2005). Best practices in exploratory factor analysis:  
four recommendations for getting the most from your analysis. Practical 
Assessment Research & Evaluation, 10(7).  
 
Covington, J., & Taylor, R. B. (1991). Fear of crime in urban residential neighborhoods:  
Implications of between- and within-neighborhood sources for current  
models. Sociological Quarterly, 32(2), 231-249. 
 
Cozens, P., Hillier, D. & Prescott, G. (2001). Crime and the design of residential property 
exploring the perceptions of planning professionals, burglars and other users: Part 
2. Property Management, 19(4), 222-248. 
 
Cozens, P., Neale, R., & Hillier, D. (2003). Managing crime and the fear of crime at 
railway stations: A case study in South Wales (UK). International Journal of 
Transport Management, 1(3), 121-132. 
 
Cozens, P., Neal, R., Hillier, D. & Whitaker, J. (2004). Tackling crime and fear of crime 
while waiting at Britain’s railway stations. Journal of Public Transportation, 7(3), 
23-42. 
 
Cozens, P., Neal, R., Whitaker, J. & Hillier, D. (2003). Investigating personal safety at 
railway stations using “virtual reality” technology. Facilities, 21(7/8), 188-194. 
 
Cozens, P., Saville, G. & Hillier D. (2005). Crime prevention through environmental 
design (CPTED): a review and modern bibliography. Property 
Management, 23(5), 328-356. 
 
Crowe, T. D., & National Crime Prevention Institute (University of Louisville). 
(1991).Crime prevention through environmental design: applications of 
architectural design and space management concepts. Boston: Butterworth-
Heinemann. 
 
Data Driven Detroit. (2015). Motor City Mapping. Retrieved from 
http://datadrivendetroit.org/data-mapping/ 
83 
 
 
Day, K. (1994). Conceptualizing women’s fear of sexual assault on campus: A review of 
causes and recommendations for change. Environment and Behavior, 26(6), 742–
765. 
 
Doran, B. J. & Burgess, M. B. (2012). Putting fear of crime on the map: investigating 
perceptions of crime using geographic information systems. New York, NY: 
Springer. 
 
Doran, B. J. & Lees, B. G. (2005). Investigating the spatiotemporal links between 
disorder, crime, and the fear of crime. The Professional Geographer, 57(1), 1-12. 
 
Dubow, F., McCabe F. & Kaplan, G. (1979). Reactions to crime: a critical review of the  
literature. Washington, DC: U.S.  
 
Erickson, R.J. (1998). Convenience store security at the millennium. Alexandria, VA: 
National Association of Convenience Stores. 
 
Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating 
the use of exploratory factor analysis in psychological research. Psychological 
Methods, 4(3), 272–299.  
 
Farrall, S., Bannister, J., Ditton, J., & Gilchrist, E. (1997). Questioning the measurement 
of the 'fear of crime': Findings from a major methodological study. British 
Journal of Criminology, 37(4), 658-679. 
 
Farrington, D.P. & Welsh, B. (2002). Improved street lighting and crime 
prevention. Justice Quarterly, 19(2), 313-342. 
 
Federal Bureau of Investigation. (2014). Uniform Crime Report. Retrieved from  
http://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/2011/crime-in-the-u.s.-
2011/about-cius. 
 
Fennelly, L., & Crowe, T. (2013). Crime Prevention Through Environmental Design.  
Burlington: Elsevier Science. 
 
Ferraro, K.F. (1995) Fear of crime: Interpreting victimization risk. Albany: State 
University of New York Press.  
 
Ferraro, K. F. (1996). Women's fear of victimization: shadow of sexual assault?. Social 
Forces, 75(2), 667-690. 
 
Ferraro, K. F. & LaGrange, R. (1987). The Measurement of Fear of Crime. Sociological 
Inquiry, 57(1), 70-97. 
 
84 
 
Figlio, R. & Aurand, S. (1991). Closed circuit interactive television and crime. In, 
Convenience store security: complete text reports with summary. Alexandria, VA: 
National Association of Convenience Stores. 
 
Fisher, B.S. & Nasar, J.L. (1992). Fear of crime in relation to three exterior site features: 
prospect, refuge, and escape. Environment and Behavior, 24(1), 35-65.  
 
Fisher, B.S., Sloan, J. J. & Wilkins, D. L. (1995). Fear of crime and perceived risk of 
victimization in an urban university setting. In B. S. Fisher & J. J. Sloan (Eds.), 
Campus crime: Legal, social and policy perspectives. Springfield, IL: Charles 
Thomas 
 
Flynn, J., Slovic, P. and Mertz, C.K. (1994). Gender, race, and perception of 
environmental health risks, Risk Analysis, 14(6), 1101-1108. 
 
Fox, K., Nobles, M. & Piquero, A. (2009). Gender, crime victimization and fear of crime. 
Security Journal, 22(1), 24-39. 
 
Fowler, F. J. (2014). Survey research methods (5th ed.). Newbury Park: Sage 
Publications.  
 
Gabriel, U., & Greve, W. (2003). The psychology of fear of crime: conceptual and 
methodological perspectives. The British Journal of Criminology, 43(3), 600-614. 
 
Garofalo, J. (1979). Victimization and the Fear of Crime. Journal of Research in Crime 
and Delinquency, 16(1), 80-97. 
 
Glazer, N. (1979). On subway graffiti in New York. Public Interest, 54, 3-11. 
 
Gibson, C., Zhao, J., Lovrich, N., & Gaffney, M. (2002). Social integration, individual 
perceptions of collective efficacy, and fear of crime in three cities. Justice 
Quarterly, 19 (3), 537-564. 
 
Gomme, I.M. (1986). Fear of crime amongst Canadians: A multi-variate analysis.  
Journal of Criminal Justice, 14, 249-258. 
 
Green, G. (2002). Fear of crime and health in residential tower blocks: A case study in 
Liverpool, UK. The European Journal of Public Health, 12(1), 10-15. 
 
Green, G., Gilbertson, J. M., & Grimsley, M. F. (2002). Fear of crime and health in 
residential tower blocks: a case study in Liverpool, UK. European Journal of 
Public Health, 12(1), 10-5. 
 
Greenbaum, P. & Greenbaum, S. (1981). Territorial personalization. Environment and  
Behavior, 13(5), 574-589. 
 
85 
 
The Greening of Detroit. (2014). Retrieved from http://www.greeningofdetroit.com/. 
 
Grice, J. W. (2001). Computing and evaluating factor scores. Psychological Methods,  
6(4), 430-450. 
 
Grice, J.W. (2007). Factor scores. In N. Salkind & K. Rasmussen (Eds.), Encyclopedia of  
measurement and statistics. Thousand Oaks, CA: Sage Publications. 
 
Hale, C. (1996). Fear of crime: A review of the literature. International Review of  
Victimology, 4, 79-150. 
 
Hanyu, K. (1993). The affective meaning of Tokyo: Verbal and nonverbal approaches.  
Journal of Environmental Psychology, 13, 161-172.  
 
Harcourt, B. E. (2001). Illusion of order: The false promise of broken windows policing.  
Cambridge, MA: Harvard University Press. 
 
Heckert M. & Mennis J. (2012).  The economic impact of greening urban vacant land: A  
spatial difference-in-differences analysis. Environment and Planning, 44(12), 
3010 – 3027 
 
Hedayati, M.M., Abdullah, A., Razak, N. A. & Maghsoodi, T. M. J. (2012a). The  
influence of crime prevention through environmental design on victimisation and 
fear of crime. Journal of Environmental Psychology, 3(2), 79-88. 
 
Hedayati, M.M., Abdullah, A., Razak, N. A. & Maghsoodi, T. M. J. (2012b). Validating 
crime prevention through environmental design construct through checklist using 
structural equation modelling. International Journal of Law, Crime and 
Justice, 40(2), 82-99. 
 
Heckert, M. & Mennis, J. (2012). The economic impact of greening urban vacant land: a  
spatial difference-in-differences approach. Environment and Planning, 44(12), 
3010-3027. 
 
Heller, D.E. (2007).Pell Grant recipients in selective colleges and universities. The 
Century Foundation.  
 
Hendricks, D., Landsittel, H., Amandus, J. & Malcan, J. Bell. (1999). A matched case-
control study of convenience store robbery risk factors. Journal of Occupational 
Environmental Medicine, 41(11), 995–1004 
 
Herzog, T. & Bryce, A. (2007). Mystery and preference in within-forest 
settings. Environment and Behavior, 39(6), 779-796. 
 
86 
 
Herzog, T. R. & Flynn-Smith, J. A. (2001). Preference and perceived danger as a 
function of the perceived curvature, length, and width of urban alleys. 
Environment and Behavior, 33(5), 653-666. 
 
Hindelang, M. J., Gottfredson, M. R. & Garofalo, J. (1978). Victims of personal crime: 
An empirical foundation for a theory of personal victimization. Cambridge, MA: 
Ballinger Publishing Company. 
 
Hinkle, J. C. (2005). The impact of disorder on fear of crime: A test of the first link of 
broken windows. College Park, MD: University of Maryland. 
 
Hinkle, J. C. & Weisburd, D. (2008). The irony of broken windows policing: A  
micro-place study of the relationship between disorder, focused police 
crackdowns and fear of crime. Journal of Criminal Justice, 36(6), 503-512. 
 
Houtkamp, J. M., & Junger, M. L. A. (2010). Affective qualities of an urban environment 
on a desktop computer. In E. Banissi (Ed.), Proceedings of the 14th International 
Conference information Visualisation (IV 2010) (pp. 597-603). Los 
Alamitos, CA, USA: IEEE Computer Society. 
 
Hunter, R.D. (1990). Convenience store robbery in Tallahassee: a reassessment. Journal  
of Security Administration, 13(1-2), 3–18. 
 
Jackson, J. (2004). Experience and expression social and cultural significance in the fear  
of crime. The British Journal of Criminology, 44(6), 946-966. 
 
Jackson, J. (2006). Introducing fear of crime to risk research. Risk Analysis, 26(1), 253-
264. 
 
Jackson, J. (2011). Revisiting risk sensitivity in the fear of crime. Journal of Research in  
Crime and Delinquency, 48(4), 513-537. 
 
Jacobs, J. (1961). The Death and Life of Great American Cities. New York: Random  
House. 
 
Jans, M. & Roman, A. M. (2007). National response rates for surveys of college students: 
Institutional, regional, & design factors. American Association for Public Opinion 
Research Annual Conference, Anaheim. Published in the American Statistical 
Association Proceedings. 
Jeffery, C. R. (1971). Crime Prevention Through Environmental Design. Beverly Hills,  
CA:Sage Publications. 
 
Jeffery, C. R. (1977). Crime Prevention Through Environmental Design. London, UK:  
Sage Publications. 
 
Jorgensen, L. J., Ellis, G. D., & Ruddell, E. (2013). Fear perceptions in public parks:  
87 
 
Interactions of environmental concealment, the presence of people recreating, and 
gender. Environment and Behavior, 45(7), 803-820. 
 
Kaplan, S. (1979). Perception and landscape: conceptions and misconceptions. In Elsner,  
G. H. & and Smardon, R.C. (Eds.), Proceedings of our national landscape: a 
conference on applied techniques for analysis and management of the visual 
resource, Berkeley, CA: U.S. Department of Agriculture.  
 
Keane, Carl. (1995). Victimization and fear: Assessing the role of the offender and the  
offence. Canadian Journal of Criminology, 37, 431-455. 
 
Kelley, K. & Maxwell, S. E. (2003). Sample size for multiple regression: Obtaining  
regression coefficients that are accurate, not simply significant. Psychological 
Methods, 8, 305-321. 
 
Kushmuk, J. & Whittermore, S. (1981). A reevaluation of the CPTED program in  
Portland. Washington DC: NIJ. 
 
La Vigne, N.G. (1997). Visibility and vigilance: Metro's situational approach to  
preventing subway crime. Washington, DC: U.S. Dept. of Justice, Office of 
Justice Programs, National Institute of Justice. 
 
Lab, S. P. (2014). Crime prevention: Approaches, practices, and evaluations. Boston,  
MA: Anderson Publishing/Elsevier. 
 
LaGrange, R.L., Ferraro, K.F. & Supancic, M. (1992). Perceived risk and fear of crime:  
role of social and physical incivilities. Journal of Research in Crime and 
Delinquency, 29(3), 311-334. 
 
Lamb, R.C., Hunter, R.D. & McLain, D.J. (1993). Fear reduction in the Charlotte 
Housing Authority. In Zahm, D. & P. Cromwell (Eds.), Proceedings of the 
international seminar on the environmental criminology and crime analysis. 
Coral Gables, FL: Florida  
Criminal Justice Executive Institute. 
 
Liska, A. E., Sanchirico, A. & Reed, M. D. (1988). Fear of crime and constrained  
behavior specifying and estimating a reciprocal effects model. Social 
Forces, 66(3), 827-837. 
 
Lumpkin, J. R (1985). Validity of a brief locus of control scale for survey research.  
Psychological Reports, 57, 655-659.  
 
Markowitz, F., Bellair, P., Liska, A. & Liu, J. (2001). Extending social disorganization 
theory: Modeling the relationships between cohesion, disorder, and 
fear. Criminology, 39(2), 293-319. 
 
88 
 
Maroof, D. (2012). Statistical methods in neuropsychology: Common procedures made  
comprehensible. New York: Springer. 
 
Maxwell, S. E. (2000). Sample size and multiple regression. Psychological Methods, 5,  
434-458. 
 
Mayhew, P., Clarke, R.V. & Elliott, D. (1989). Motorcycle theft, helmet legislation and  
displacement. The Howard Journal of Criminal Justice, 28(1), 1-8. 
 
Mayhew, P., Clarke, R.V., Sturman, A. & Hough, J.M. (1976). Crime as opportunity.  
Home office research study. London: H.M. Stationery Office. 
 
McGarrell, E., Gizcomazzi, A. L. & Thurman, Q. C. Reducing disorder, fear, and  
crime in public housing: A case study of place-specific crime prevention. Justice 
Research and Policy. 1(2), 61-87. 
 
Minnery, J. R. & Lim, B. (2005). Measuring crime prevention through environmental  
design. Journal of Architectural and Planning Research, 22(4), 330-341. 
Moffat, R. (1983). Crime prevention through environmental design - a management 
perspective. Canadian Journal of Criminology, 25(4), 19–31. 
 
Morrall, P., Marshall, P., Pattison, S. & Macdonald, G. (2010). Crime and health: a 
preliminary study into the effects of crime on the mental health of UK university 
students. Journal of Psychiatric and Mental Health Nursing, 17(9), 821-828. 
 
Morrow, E. N., & Hutton, S. A. (2000). The Chicago alley lighting project: final  
evaluation report. Chicago, IL: Illinois Criminal Justice Information Authority. 
 
Nasar, J.L. & Fisher, B.S. (1992). Design for vulnerability: cues and reactions to fear of  
crime. Sociology and Social Research, 76(2), 48-58. 
 
Nasar, J.L. & Jones, K. (1997). Landscapes of fear and stress. Environment and  
Behavior, 29(3), 291-323. 
 
National Opinion Research Center. (2012). General Social Survey. Retrieved from  
http://www.norc.org/Research/Projects/Pages/general-social-survey.aspx. 
 
Newman, O. (1972). Defensible Space: Crime Prevention through Urban Design. New  
York: Macmillan. 
 
Office for International Statistics. (2014). Crime survey. Retrieved from 
http://www.crimesurvey.co.uk/previous-research.html. 
 
Olsson, U. (1979). Maximum likelihood estimation of the polychoric correlation  
coefficient. Psychometrika, 44(4), 443–460. 
89 
 
 
Pain, R. (2000). Place, social relations and the fear of crime: a review. Progress in  
Human Geography, 24(3), 365-387. 
 
Painter, K. & Farrington, D.P. (1997). The crime reducing effect of improved street  
lighting: The dudley project.  In R.V. Clarke (Ed.), Situational crime prevention: 
successful case studies, 2nd edition, New York, NY: Harrow and Heston. 
 
Palmer, J. F. & Hoffman, R. E. (2001). Rating reliability and representation validity in 
scenic landscape assessments. Landscape and Urban Planning, 54(1-4), 149-161. 
 
Park, A.J. (2008). Modeling the role of fear of crime in pedestrian navigation. Burnaby, 
BC: Simon Fraser University. 
 
Park, A. J., Clare, J., Spicer, V., Brantingham, P. L., Calvert, T. & Jenion, G. (2012). 
Examining context-specific perceptions of risk: Exploring the utility of “human-
in-the-loop” simulation models for criminology. Journal of Experimental 
Criminology, 8(1), 29-47. 
 
Park, A. J., Hwang, E., Spicer, V., Cheng, C., Brantingham, P. L. & Sixsmith, A. (2011). 
Testing elderly people’s fear of crime using a virtual environment. In Proceedings 
of the 2011 European Conference on Intelligence and Security Informatics 
(EISIC) (pp. 63-69). IEEE. 
 
Parker, K. & Ray, M.C. (1990). Fear of crime: An assessment of related factors. 
Sociological Spectrum, 10(1), 29-40 
 
Parnaby, P. F. (2006). Crime prevention through environmental design: discourses of 
risk, social control, and a neo-liberal context. Canadian Journal of Criminology 
and Criminal Justice, 48(1), 1-30. 
 
The Pennsylvania Horticultural Society. (2014). Retrieved from http://phsonline.org/ 
 
Perkins, D.D. & Taylor, R. (1996). Ecological assessments of community disorder: Their  
relationship to fear of crime and theoretical implications. American Journal of 
Community Psychology, 24(1), 63-107. 
 
Perkins, D.D., Meeks, J.W. & Taylor, R.B. (1992). The physical environment of street 
blocks and resident perceptions of crime and disorder: Implications for theory and 
measurement. Journal of Environmental Psychology, 12, 21-34. 
 
Perkins, D.D., Wandersman, A., Rich, R.C., & Taylor, R.B. (1993). The physical 
environment of street crime: defensible space, territoriality and incivilities. 
Journal of Environmental Psychology, 13, 29-49.  
 
Petherick, N. (2000). Environmental design and fear: The prospect-refuge model and the  
90 
 
university college of the cariboo campus. Western Geography, 10/11, 88-112.  
 
Pett, M., Lackey, N. & Sullivan, J. (2003). Making sense of factor analysis. Thousand  
Oaks: Sage Publications, Inc. 
 
Pitner, R., Yu, M., & Brown, E. (2012). Making neighborhoods safer: Examining 
predictors of residents' concerns about neighborhood safety. Journal of 
Environmental Psychology, 32 (1), 43-49. 
 
Pollack, L. M. & Patterson, A. H. (1980). Territoriality and fear of crime in elderly and  
nonelderly homeowners. The Journal of Social Psychology, 111, 119-129. 
 
Porter, R. & Fabrigar, L. (2007). Factor Analysis. In N. Salkind & K. Rasmussen (Eds.),  
Encyclopedia of measurement and statistics (pp. 342-246). Thousand Oaks, CA: 
SAGE Publications.  
 
Poyner, B. and Webb, B. (1987) Successful crime prevention case studies. Document No.  
2T 563. London: Tavistock Institute of Human Relations. 
 
Prociuk, T. J., Breen, L. J. & Lussier, R. J. (1976). Hopelessness, internal-external  
locus of control, and depression. Journal of Clinical Psychology, 32(2), 299-300. 
 
Riechard, D. E. & Peterson, S. J. (1998). Perception of environmental risk related to  
gender, community socioeconomic setting, age, and locus of control. Journal of 
Environmental Education, 30(1), 11-19. 
 
Ross, C. E. & Mirowsky, J. (2001). Health and illness - neighborhood disadvantage,  
disorder, and health. Journal of Health & Social Behavior, 42(3), 258-276. 
 
Rossi, J. S. (1990). Statistical power of psychological research: What have we gained in  
20 years? Journal of Consulting and Clinical Psychology, 58, 646–656. 
 
Rotter, J. B. (1966). Generalized expectancies for internal versus external control of  
reinforcement. Psychological Monographs, 80(1), 1-28. 
 
Rountree, P. W. & Land, K. C. (1996). Perceived risk versus fear of crime: Empirical  
evidence of conceptually distinct reactions in survey data. Social Forces, 74, 
1353-1376. 
 
Ruscio, J., & Roche, B. (2012). Determining the number of factors to retain in an 
exploratory factor analysis using comparison data of known factorial structure. 
Psychological Assessment, 24(2), 282–92.  
 
Rusticus, Shayna A. & Chris Y. Lovato (2011). Applying Tests of Equivalence for  
Multiple Group Comparisons: Demonstration of the Confidence Interval 
Approach. Practical Assessment, Research & Evaluation, 16(7). 
91 
 
 
Sakip, S. R. M., Johari, N. & Salleh, M. N. M. (2012). The Relationship between crime  
prevention through environmental design and fear of crime. Procedia - Social and 
Behavioral Sciences, 68(4), 628-636. 
 
Sampson, R. & Raudenbush, S. (2004). Seeing disorder: neighborhood stigma and the  
social construction of “broken windows”. Social Psychology Quarterly, 67(4), 
319-342. 
 
Schmuckler, M.A. (2001). What is ecological validity? A dimensional analysis.  
International Society on Infant Studies, 2(4), 419-436. 
 
Schneider, R.& Kitchen, T. (2002). Planning For Crime Prevention: A Transatlantic  
Perspective, Routledge, London. 
 
Schneider, R. H. & Kitchen, T. (2007). Crime prevention in the built environment. New  
York, NY: Routledge, Taylor & Francis Group. 
 
Schwebel, D. C., Gaines, J. & Severson, J. (2008). Validation of virtual reality as a tool 
to understand and prevent child pedestrian injury. Accident Analysis & 
Prevention, 40(4), 1394-1400. 
 
Shadish, W. R., Cook, T. D. & Campbell, D. T. (2002). Experimental and quasi- 
experimental designs for generalized causal inference. Boston, MA: Houghton 
Mifflin. 
 
Silver, E. & Miller, L. (2004). Sources of informal social control in 
Chicago. Criminology, 42 (3), 551-584. 
 
Skinner, B. F. (1938). The Behavior of Organisms: An Experimental Analysis. New York,  
London: D. Appleton-Century Company, Incorporated. 
 
Skogan, W. (1986). Fear of Crime and Neighborhood Change. Crime and Justice, 8, 203- 
229. 
 
Skogan, W. (1987). The impact of victimization on fear of crime. Crime and  
Delinquency, 33, 135-154.  
 
Skogan, W. (1990). Disorder and decline: Crime and the spiral of decay in American  
neighborhoods. New York: Free Press. 
 
Skogan, W. & Maxfield, M. G. (1981). Coping with crime: Individual and neighborhood  
reactions. Beverly Hills: Sage Publications. 
 
Sloan, J.J., Fisher, B.S. & Wilkins, D.L. (1996). Reducing perceived risk and fear of  
victimization on campus: a panel study of faculty members, staff, and  
92 
 
students. Journal of Contemporary Criminal Justice, 12(1), 81-107. 
 
Slovic, P., & International Institute for Environment and Development. (2010). The  
 feeling of risk: New perspectives on risk perception. London: Earthscan. 
 
Smith, G. & Lab, S. (1991).  Urban and rural attitudes toward participating in an auxiliary  
policing crime prevention program. Criminal Justice and Behavior, 18(2), 202-
216. 
 
Spelman, W. (1993). Abandoned houses: magnets for crime? Journal of Criminal Justice,  
21, 481-495. 
 
Stamps, A.E. (1990). Use of photographs to stimulate environments: A meta-analysis.  
Perceptual and Motor Skills, 71, 907-913.  
 
Stamps, A.E. (2010). Use of static and dynamic media to simulate environments: A meta- 
analysis. Perceptual and Motor Skills, 111(2), 355-364. 
 
Stamps, A.E. (2005). Visual permeability, locomotive permeability, safety, and  
enclosure. Environment and Behavior, 37(5), 587-619. 
 
Stamps, A. E., & Smith, S. (2002). Environmental enclosure in urban settings.  
Environment and Behavior, 34(6), 781-794. 
 
Suhr, D. (2006). Exploratory or Confirmatory Factor Analysis. SAS Users Group  
International Conference (pp. 1 - 17). Cary: SAS Institute, Inc. 
 
Sutton, A., Cherney, A., & White, R. (2008). Crime prevention: principles, perspectives  
and practices. Cambridge, UK: Cambridge University Press.  
 
Taylor, R. B. (1996). Neighborhood responses to disorder and local attachments: The  
systemic model of attachment, social disorganization, and neighborhood use 
value. Sociological Forum, 11(1), 41. 
 
Taylor, R. B. & Harrell, A. (1996) Physical environment and crime. Washington:  
National Institute of Justice.  
 
Taylor, R. B., Shumaker, S. & Gottfredson, S. (1985). Neighborhood-level links between  
physical features and local sentiments. Journal of Architectural Planning and 
Research, 2, 261-275. 
 
Toet, A., & van Schaik, M.G. (2012). Effects of signals of disorder on fear of crime in r 
real and virtual environments. Journal of Environmental Psychology, 32(3), 260-
276. 
 
Toseland, R.W. (2002). Fear of crime: Who is most vulnerable?. Journal of Criminal  
93 
 
Justice,10(3), 199-209. 
 
Troy, A., Morgan, G. J. & O’Neil-Dunne, J. (2012). The relationship between tree 
canopy and crime rates across an urban–rural gradient in the greater Baltimore 
region. Landscape and Urban Planning, 106 (3), 262-270. 
 
Uebersax, J. (2015). Introduction to the tetrachoric and polychoric correlation  
coefficients. Retrieved from http://john-uebersax.com/stat/tetra.htm 
 
United States Environmental Protection Agency. (2013). What are vacant lots? Retrieved  
from http://www.epa.gov/region1/eco/uep/vacantlots.html  
 
Urban Green Council. (2014). Retrieved from http://urbangreencouncil.org/ 
van Hagen, M. (2011). Waiting experience at train stations. Delft, Netherlands: 
Eburon. 
 
Vanderveen, G. (2006). Interpreting fear, crime, risk, and unsafety: conceptualisation 
and measurement. Den Haag: Boom Juridische Uitgevers. 
 
Wallis, A. & Ford, D. (1981). Crime prevention through environmental design: An  
operational handbook. Washington, D.C: U.S. Department of Justice, National 
Institute of Justice. 
 
Wang, K. & Taylor, R. B. (2006). Simulated walks through dangerous alleys: Impacts of  
features and progress on fear. Journal of Environmental Psychology, 26, 269- 
283. 
 
Ward, C. (Ed.). (1973). Vandalism. London: Architectural Press. 
 
Warr, M. (1985). Fear of rape among urban women. Social Problems, 32(3), 238-250.  
 
Warr, M. (1987). Fear of victimisation and sensitivity to risk. Journal of Quantitative  
Criminology, 3(1), 29-46. 
 
Warr, M. (1990). Dangerous situations: social context and fear of victimization. Social  
Forces, 68(3), 891-907. 
 
Warr, M. (1994). Fear of victimization: Why are women and the elderly more afraid?.  
Social Science Quarterly, 65(3), 681-702.  
 
Warr, M., & Stafford, M. (1983). Fear of victimization: A look at the proximate 
causes. Social Forces, 61 (4), 1033-1043. 
Welsh, B.C. & Farrington, D. P. (2004). Surveillance for crime prevention in public 
space: results and policy choices in Britain and America. Criminology & Public 
Policy, 3(3), 497-526. 
94 
 
Wikström, P. (1990). Delinquency and the urban structure. In P. Wikström (Ed.), Crime 
and measures against crime in the city, Stockholm, National Council for Crime 
Prevention.  
 
Will, J.A. (1995). Crime, neighborhood perceptions, and the underclass the relationship  
between fear of crime and class position. Journal of Criminal Justice, 23(2), 163-
176. 
 
Will, J.A. & McGrath, J.H. (1995). Crime, neighborhood perceptions, and the underclass:  
The relationship between fear of crime and class position. Journal of Criminal 
Justice, 23(2), 163-176. 
 
Wilson, J.Q. & Kelling, G.L. (1982). Broken windows: the police and neighborhood  
safety. Atlantic Monthly, 249(3), 29-38.  
 
Wilcox, R. P., & Land, K. C. (1996). Burglary victimization, perceptions of crime risk, 
and routine activities: a multilevel analysis across Seattle neighborhoods and 
census tracts. The Journal of Research in Crime and Delinquency, 33(2), 147-
180. 
Wood, E. (1961). Housing Design: A Social Theory. New York, NY: Citizens' Housing 
and Planning Counsel of New York.